1.1 The Stock Record
1.2 The Value of Record Accuracy
1.3 How Stock Record Inaccuracy invalidates Planning
1.4 Defining Accuracy
1.4a Percentage Piece Variance
1.4b Financial Percentage Piece Variance
1.4c Good Count / Bad Count
1.5 The Accuracy Goal
1.6 Tolerances
1.7 Assigning Tolerances
1.8 The Starting Accuracy Audit
2.1 Controlled Stores/Warehouse Access
Backflushing
Retail
2.2 The Value of Variable (Random) Locations
2.3 Equipment for Weighing and Measuring
2.4 On-Line Data Processing
2.5 Bar Coding and RFID Tags
2.6 The Working Environment
3.1 The Total System
3.2 The Data Processing Procedures
Typical Transaction Types
3.3 Procedures and Responsibilities for Raising Data
3.4 Other Responsibilities
3.4a Goods In
Incoterms
3.4b 'Put-aways'
The 'Magic Number' Method
3.4c Picking
3.4d Other Procedures
4.1 Introduction to Transaction (Audit) Trails
4.1a Negative Stock
4.1b Four Reasons for Investigation
4.2 Count Variance Investigation & Control Groups
4.3 Work-in-Progress
4.3a WIP and a Transaction Trail
4.3a Backflushing
4.3c Cycle Counting of WIP
4.4 Reconciliation of a Stock Count
4.4a Late Transactions
4.4b Premature Transactions
4.4c Solution to the Two Reconciliation Problems
Solution to Late Transactions
Solution to Premature Transactions
Reconciliation Software
5.1 100% Physical Inventories (The Annual Stocktake)
5.2 Definition of Cycle Counting
5.3 Two Purposes of Cycle Counting
5.3a Identifying Errors
5.3b The Correction of Errors
5.4 The Conduct of Cycle Counting
5.4a Blind Counting v. Verification Counting
5.4b Exception Reporting
5.4c Storage of Variances in a Special Data Field
5.5 Cycle Lengths
5.6 The Classic Approach to Cycle Counting
5.6a Introduction
5.6b ABC and Annual Value
5.6c ABC and Transaction Frequency
5.6d Alternative ABC Methods
5.7 Calculating the Count Load
Spreadsheets
5.8 Low Stock Counting
5.8a Computer Identification of Low Stock Items
5.8b Identification of Low Stock Items by 'Patrol' (walking around)
5.8c Ensuring All Items are Counted
5.9 Other Item Selection Methods
5.10 Location Auditing
5.11 Cycle Counting Variable (Random) Location Items
5.11a Conventional Cycle Counting
SKUs
5.11b Zero Stock Identification
5.11c Periodic Checks to ensure all locations have been Verified
5.11d Variable Location Cycle Counting - Reporting & Auditing
5.12 Reconciling Variances
5.13 Performing the Cycle Count, Documentation & Final Audit
5.14 The Valuation of Stock
6.1 Personal Accountability
6.1a The Stores/Warehouse Manager
6.1b Stores and Warehouse Staff
Recruitment
Selection
6.1c The Cycle Counter
6.1d The Purchasing Manager
6.2 Training
6.3 Determining the Limits of Performance of the Current System
6.4 Bringing the System under Control (by Removing Special Causes of Error)
6.5 Reducing the Effect of Common Causes of Error - Management Action
7.1 The Role of Management
7.2 The Data Processing System
7.3 Implementation of the IT System and Selecting Software
7.4 Education
7.5 Selecting Software
1. By quantifying the value of accuracy ... or alternatively, the cost of inaccuracy ... we will obtain valuable insight into what objectives we should aim for;
2. By identifying where inaccuracy is having the greatest positive or negative effect, we will see where best to concentrate our attention;
3. The value of records accuracy tells us how important it is to get the data right, and how justified we are in going to what may seem as relatively extreme lengths to achieve very high standards of exactness.
One value of accuracy is the avoidance of costs associated with inaccuracy. These include (a) split batches and short production runs (because there turned out to be insufficient material to start a full production job); (b) failures to meet deliveries (having promised the delivery on the basis of the computer record, the material was short when it came to picking it physically); (c) spoilage and obsolescence (unknown material eventually stumbled upon); (d) double handling (a location supposed to be empty according to the record turns out to be occupied); (e) excess stock holding (other staff may not know if you are holding a little too much of a product, but everyone in the company knows if you run out !).
A second value of accuracy is the elimination of quarantine areas. A typical quarantine area is a kitting area, where stock intended for a specific works order is physically quarantined by the production manager responsible so that he knows the material will be available when the time comes. Unfortunately, late component deliveries to kitting areas often find their way to the wrong kits. And even though they are under lock and key, kitting areas are robbed. Cycle counts frequently miss kitting areas. A better solution by far is to eliminate such physical areas altogether, and instead "reserve" the required material as being allocated, on the stock record itself. The planning system ensures the stock allocated to kits remains sacrosanct until it is issued to the specified job. This procedure clearly works only if the stock record is 100% accurate. Thus suppose the stock record read 100 units present, of which 40 are allocated and therefore 60 free. The 60 free units are then issued to a new job. However, now suppose that the number of units of actual physical stock was 90 units, not 100. After the issue of 60 'free' units, there would in fact remain only 30 units for the special job, not the full 40 units.
A third value of records accuracy is to maintain financial control. There are two reasons why the accountant must verify on an annual basis the value of all stock held. The first is the place of stock on the balance sheet. The balance sheet is struck at the end of the company's financial year, and states its assets and liabilities. Assets are split between fixed assets (machinery, land and buildings) and 'current' assets (cash, debtors and stock). Stock is regarded as an asset since it is acquired (ie made) speculatively, in the hope and expectation of a future sale - ie it is in effect a corporate investment. The second reason relates to the financial nature of stock. As just stated, stock is regarded as an investment made in the expectation of a future sale. Consequently, expenditure on the manufacture of stock is not regarded from the viewpoint of accountancy as an expense in the normal, financial sense of the word. It is (speculative) expenditure, which becomes a financial expense in the profit and loss account only when the sale is realised. Consequently, it is highly convenient from the accounting viewpoint to keep a close watch on the status and quantity of stock, since these inform the accountant how much money has been spent as stock investment, as opposed to having been spent on other "true" expenses such as rent and tax which are able to be charged to the accounts and set against revenue to derive profit.
1.3 How Records Inaccuracy invalidates Planning
The fourth value of accurate stock records is for production planning. Accurate data for planning is essential if the plans produced are to be valid and believable. It is only through accurate stock records that the advantages of such planning systems as MRPII will be achieved. It is often difficult to get the attention of senior management with regard to stock accuracy, but never difficult to attract attention to planning and logistics systems. Yet without stock accuracy, planning and logistics systems are worthless. Consider the plans to manufacture Product A and support the plans for A with plans for Product B immediately below it in the Bill of Materials. The plans for A and B must clearly be based on the starting stock balances of the two products as read on the products' computer records. Suppose however that the actual stocks of A and B are different from what is recorded. The real needs of product A are consequently different from what was previously calculated. If the needs for product A are different, then the needs for product B must also be different.
Stock records inaccuracy creates over-production, over stock and above all, uncertainty throughout the chain of manufacture. The reaction of most manufacturing staff to uncertainty and possible shortages is to over-produce ... no-one ever got promoted for creating shortages. An expensive MRPII system can be destroyed in weeks when the output therefrom loses credibility with the staff required to support it.
1.4 Defining Accuracy
There are a number of ways by which the accuracy of a group of records might be defined, and the choice of a preferred method depends on the purposes to which the accuracy figure is to be put. Strictly, the method which best expresses accuracy of the records is the mathematical measurement of "variance". It may be thought, however, that the refinement obtained by calculating the variance is obtained at too high a price in terms of complexity. Instead, the three methods below may be preferred - they provide a reasonable picture of accuracy and have the merit of being straightforward - simpler to calculate and simpler to comprehend.
1.4a The Percentage Piece Variance
The percentage piece variance is also known as the absolute piece variance, and compares the sum of the absolute variances of a group of actual product stock counts with the products' records. For example, if the stock counts of two products A and B were 98 units and 105 units, and the records of the stock figures were 100 each, the piece variances between the two sets of figures are - 2 and + 5. The absolute variances ignore the negative sign, however, and so that these are simply 2 and 5, total 7. The percentage absolute piece variance of the group of two products compares the total absolute variance of 7 to the total of the recorded figures (200 ... ie 100 + 100). Thus the absolute accuracy is 7 / 200, or 3.5%. This is expressed as 96.5%.
1.4b The Financial Percentage Piece Variance
The financial percentage piece variance , or absolute financial percentage piece variance, is calculated in a similar way to the percentage piece variance above, except that variances are expressed in terms of the monetary value of the stock. Thus if the stock value of products P and Q were £97 and £108, and the records of the stock figures were £110 and £102, the variances in these financial terms are - £13 and + £6. The absolute variances are thus £13 + £6, total £19. Consequently, the percentage financial piece variance is £19 / £212, or 8.9%. Thus the financial percentage absolute variance is 91.1%. Note that if the absolute figures were not used, and positives and negatives were therefore allowed to cancel, the records accuracy would appear much better - in this case, - £7, giving an accuracy of 96.7%. It is common for the accountant to express records accuracy in this way, since what is primarily of interest to him is the total value of the stock, not the accuracy of the records per se.
1.4c Good Count / Bad Count
This third definition is by far the most common in logistics management and will be used throughout the remainder of this on-line Course. It relies on the simple designation of a stock record, when compared to the count of actual stock, as either being correct ("good") or incorrect ("bad"). Thus if there are two products X and Y with actual stock of 96 units and 100 units, and their records show stock of 100 units each, then the record for X is incorrect and that for Y is correct. The definition of accuracy is Total number of good counts / Number of Counts in Total. In the case of X and Y, the accuracy of the records is therefore 1 / 2 x 100% = 50%.
The Good Count / Bad Count definition introduced above is the most common and widely understood measurement of accuracy when applied to the records of any file, whether held on a card index or computer, and is given by the expression below. Again, as implied, it is a particularly useful and obvious yardstick for inventory records, readily understood by all.
The number of records measured which are correct x 100%
The total number measured
A number of suggestions have been made by various gurus in industry as to the accuracy goal to be aimed for. For example, the late Oliver Wight suggested that to support the MRPII planning system, 95.0% was the minimum figure. Other logistics experts would nowadays however put forward 97.0% as the minimum. In fact, as we shall see below, because of the notion of tolerances, the precision of the number is meaningless, although 97.0% is a reasonable target to be accepted for now.
The accuracy actually attained will not be a constant figure from week to week. In any "system", there are a myriad of activities and disturbances affecting its operation which can cause deviations from perfect results. The reasons for these disturbances are termed special causes of error and common causes of error. Special causes of error in the stock record are unusual and untoward events each of which is directly traceable to some specific "special" cause, a cause which can be prevented from re-occurring in the future. (An example might be theft.) Common causes, by contrast, are natural to the working of the system itself; they cannot be eliminated as such, although the impact of them can be reduced by changing the system. (An example of a common cause might be carelessness in counting stock.) When the stock recording system has settled down, therefore - ie when it has attained a stable state - common causes of error will still be present. As a result of their presence, the stock accuracy figure will not be 100%; instead it will vary around an average and lie between an upper and and lower extreme. For example, if we attain an average figure of 84.0%, what in fact we will achieve is a preponderance of reading about 84.0%, but also other readings from a high of (say) 85.5% to a low of (say) 82.5%. To prove that the system is stable and that only common causes of error are at work, the percentage accuracy can be plotted from week to week (see Section 6.3 below), and the deviations around the "line of stability" (ie average) observed. It is likely to take some six to nine months to attain a stable system with a high level of accuracy . The improvements to be made to the system along the way consist of (1) the elimination, one by one, of special causes of error, and (2) the institution of revised systems procedures, including perhaps a new IT system, to reduce the impact of common causes of error. Progress along the way will be manifest in many small stepwise increases in the percentage accuracy recorded, not a smooth upwards progression. Each tiny step corresponds to the introduction of some new measure to improve the Stock Recording System.
1.6 Tolerances
The question that the subject of tolerances addresses is whether the computer record must match the physical stock quantity exactly for the record to be declared "correct". The degree of difference allowed between the recorded figure and the actual amount of stock present, such that the record is nevertheless deemed to be correct for the purpose of calculating good count / bad count, is termed its tolerance.
There are two interrelated constituents of the tolerance assigned to a particular product: (1) The percentage tolerance, and (2) the absolute tolerance. Both of them must be assigned to every stock record on an individual basis and together define the amount by which the record can differ from the physical stock present and still be considered "correct" for the purposes of arriving at a stock accuracy figure.
As an example, if the percentage tolerance is t1% and the physical count is n, then, if we were using this basis alone, the record would be considered to be correct if it was between (n - t1 x n) and (n + t1 x n). For example, if the percentage tolerance was 5.0% and the physical count was 1000, the record would be considered correct if it was between 950 and 1050. If the absolute tolerance is t2 and the physical count is n, the record would be correct on that basis alone if it were between (n - t2) and (n + t2). For example, if the absolute tolerance if 30 units and the physical count is 1000 units, the record would be considered correct if it were between 970 units and 1030 units. To use the two tolerance figures in conjunction with each other, the rule is that if t1 x n (ie the calculated percentage figure) is greater than t2 (the absolute figure) then it is the percentage tolerance which determines whether the record is correct. However, if t2 is greater than t1 x n it is the absolute tolerance which determines whether the record is correct.
For example, suppose the percentage and absolute tolerances on a product 'Size 4 brads' held in a certain store are 5% and 30. Then (1) on one occasion the physical count of brads in the store is 6,000 so that t1 x n is 300 (ie 6000 × 5%). Since this is greater than t2 (ie 30), the percentage tolerance applies. That is, the computer record is considered to be correct if it lies between 5,700 and 6,300. (2) On a second occasion the physical count is 400 so that t1 x n is 20 (400 × 5%). Since the absolute tolerance of 30 is now greater than the percentage tolerance, the record is now deemed to be correct if it lies between 370 and 430.
Many companies do not adopt the sophistication of both a percentage and absolute tolerance. Instead they rely only on a percentage figure. However, the reason for maintaining the two, interrelated, tolerances here becomes clear when comparatively low stocks are considered, especially those of common products that are replenished in large lots. One simple rule that might be applied is to set the absolute tolerance such that it 'takes over' when the stock quantity falls below, say, 2% of the delivery quantity. For example, by this rule, a product X replenished in lots of 100,000 would have an absolute tolerance of 2000. The necessity for an absolute figure can be illustrated with Size 4 Brads above: when the quantity in stock is reduced to 400 with the percentage tolerance of 5% alone still applying, the record would need to lie between 380 and 420 to be correct. This degree of exactitude may not seem reasonable in this case and would in any case jeopardise the viability of cycle counting systems within the stores based on low stock routes, described below. In the remaining text of this on-line course, the general term 'tolerance' is used and implies both the percentage and absolute. When a product/record tolerance is set, the two elements must be considered in conjunction with each other.
The rule for setting the absolute tolerance equal to the actual percentage tolerance at x% of the replenishment quantity may well be useful for warehouse products subject to steady customer demand and stores items which are consumables. Stores items planned in large lots by MRP or a similar planning system may have comparatively lower absolute tolerances.
Before dealing with the issue of the determination of tolerances themselves, the point should be made that they are best set and maintained for substantial periods of time, certainly being unchanged over the development phase of any stores' accuracy project. If they are tightened halfway through, resulting in a sudden drop in accuracy seemingly for no good reason, staff are likely to become demotivated. It does not follow, however, that tolerances should never be decreased. When very high levels of accuracy are being obtained, it may be possible to do so as part of a stores or warehouse operational improvement programme, even leaving the overall records accuracy unchanged.
Finally, let it be firmly stated that no attempt should be made to make the accuracy of a stores' or warehouse's records appear more accurate than they are by manipulating the tolerances that have been set on the products. For one thing, this reprehensible trick simply does not work! It would be necessary to enlarge the tolerance gaps to quite unacceptable margins to make any appreciable difference to the overall accuracy. The idea of tolerances is to give the company a more reasonable view of the records accuracy than a strict adherence to the requiring of absolutely exact correspondence between the record and the actual. Indeed, for products which are weighed or reckoned by length or volume, rather than simply counted in 'eaches', absolute exactness is impossible to measure even with the most sensitive measuring instruments (see 1.7 below). See also Section 5.12.
1.7 Assigning Tolerances
Before considering the actual assignment of tolerances to products in the stores or warehouse, it is first necessary to deal with some theory relating to measurement.
First, we can divide the data resulting from measuring stock into continuous data and count data.
C o n t i n u o u s D a t a (stock measured by weight, volume, length etc.)
There is inevitably a difference between the measured value M and the 'true' value of what is measured T. That is:
Measurement Error = M - T
The difference (M - T) can be reexpressed as (M + µ1) + (µ2 - T). The first term (M + µ1) merely states that there is a difference from the true value in the measurement taken, due to random variation in the measuring process. The second term (µ2 - T) expresses the bias in the measuring process due to such systematic errors as mis-calibration of equipment. The calibration of instruments and apparatus is a vital matter in stock accuracy and will be touched on below. Presence of the random element in measurement is inevitable. To prevent it obscuring the simple matter of determining whether a record is or is not accurate, a tolerance is set to eliminate its effect. A simple and usually effective way of setting tolerances on measurements of continuous data is first to rank the items in order of unit financial value. The 2% of the items at the top of the list may take a tolerance of 0.1 % and the remainder 5%. (Those at the top are managerially more sensitive, regardless of the measurement process).
C o u n t D a t a
A simple way of dealing with measurements having to do with counting is to divide the items into those which are dispensed on a day to day basis by hand counting, and those which are dispensed by weigh or scale counting. Hand counted items are assigned 0% tolerance and weigh/scale counted items may be assigned 5%.
In summary, the assignment of tolerances can be swiftly accomplished even where very large numbers of items are involved. As a quick rule of thumb, most continuous materials and all weigh counted materials can be assigned a 5% tolerance. Expensive continuous items and hand counted items will have a 0% tolerance.
Excessive time and effort spent on assigning tolerances is to be avoided. As stated above, it is shown by experience that it is not possible to increase the % accuracy of the records by manipulating tolerances. Imprecision of measurement is a very small factor contributing to inaccuracy.
There are two cases where there may be difficulty in making exact measurements, but where the use of tolerances may not be best or may not be acceptable.
If stock is of high value or is of similar critical importance, and measurement is difficult because of its physical nature and/or the type of apparatus needed to count it, it is necessary to take several measurements and average them. To reduce uncertainty by 50% in the average value calculated, it is necessary to take and average 3 readings. To reduce uncertainty to 25%, 10 measurements are necessary. (The theory from which these figures are obtained relates to the mathematical distribution of the random errors µ1 referred to above.) The greatest danger to accuracy in these circumstances, however, will remain the perennial enemy of systematic errors in the measuring process. These are undetectable errors caused by off-centre instrumentation, poorly calibrated apparatus and so on. Systematic errors must be reduced to negligible proportions by careful examination of the measurement process, both with regard to procedure and instruments.
One of the more sobering statements made on a regular basis by the UK Atomic Energy Authority relates to the loss each year of weapons- and reactor-grade plutonium. In the past 25 years, more than 80 kilograms have 'gone missing' in the Sellafield nuclear facility alone. A spokesman for the UK Atomic Energy Authority said the annual amount reported as lost "were mainly caused by margins of error in the counting process. Plutonium going through reprocessing plants is not always in a form in which it can be physically measured, so its presence has to be calculated. At each stage of the process as it is chopped up, dissolved, or whatever, calculations are made of how much plutonium there is. We cannot calculate it totally accurately. As well, outright losses have happened - sometimes a fuel rod gets lodged somewhere and avoids the processing line". [Sunday Telegraph, 7.12.97].
Where the value of stock is very low, but the nature of the material makes counting and control expensive, it may be worthwhile to give up counting altogether and revert to a simple two-bin system. The two-bin system ... the procedure we all use to control the replenishmemt of coffee in our kitchens ... is purely a replenishment system and does not rely on measurement as such.
1.8 The Starting Accuracy Audit
To begin the journey to high stock records accuracy, it is recommended that an "accuracy audit" be conducted to find the starting position. In fact, such audits should be performed from time to time on an on-going basis. Accuracy audits are not the same as cycle counts, since their purpose is different. The purposes of cycle counting are error correction and investigation. Here, the purpose is merely that of audit.
Although the final result of accuracy auditing may be a single figure relating to the records in total, it is usually desirable in practice to determine the accuracy of distinct separate groups of items, especially if it is suspected that the accuracy of product records in the groups identified will be different because of different factors affecting stock. In a stores, obvious examples of groups are raw materials; intermediate components; and finished goods. Other groups relating to a particular store's environment might be: fast moving items; items stored off-site; items which are very numerous; and material which is physically difficult to count. In order to find the accuracy of the records for a particular group, it is suggested that a randomly selected sample of 50 counts be taken from the group and the accuracy calculated in the way explained above. In order, later, to combine the individual group results and calculate the stock records accuracy of the stores overall, the percentage accuracies of the individual groups should be weighted according to the number of items in each of the groups. An example of doing so is given as follows:
Group accuracy of Raw Materials (from sample) 72.0%
Number of Raw Materials in Store 1100
Number of inaccurate Raw Material Records 1100 × 28% = 308
Group accuracy of Components (from sample) 84.0%
Number of Components in Store 900
Number of inaccurate Component Records 900 × 16% = 144
Group accuracy of Finished Goods (from sample) 94%
Number of Finished Goods in Store 500
Number of inaccurate Finished Goods Records 500 × 6% = 30
Overall records accuracy (2500 - 482) / 2500 = 80.7%
Care should be taken with combining accuracy figures calculated by financial value, to avoid positives and negatives cancelling, resulting in an unrealistically good figure.
The first assault on eliminating the special causes of variation in accuracy levels, on the journey to high, stable stock records accuracy, relates to the matter of controlled access to the stock itself.
Dealing first with a stores, if the stores supervisor and his staff are to be accountable for the accuracy of the stock records and the integrity of the material in their charge, it is essential that they should maintain full control and supervision, at all times, of stock issues and receipts. In short, the stores must be physically secured and manned at all times. One of the problems in achieving this, particularly in a manufacturing environment, is clearly the need - and the cost - of manning the entry to the facility, although this may not be as expensive as it first sounds if the current (informal) time is subtracted of supervisors and other shop floor operatives searching for parts themselves and counting out material. Although the principle of a fully-closed store must not be violated, there are a number of ways to mitigate its cost and possible inconvenience. Examples are (1) Instituting pre-planned issues. Although the normal manning of the stores may be (say) three people, it may be possible to reduce this number for the 2nd or 3rd shift by instituting pre-planned issues to the shop floor in the 1st shift, or by pre-kitting. (2) Bulk Issues to the Shop Floor of high usage, low cost items. (3) Making periodic supplies to the shop floor. Provided the system is rigidly controlled and supported by a good shop floor data collection system, each work centre might be issued with (say) 2 days of materials.
In general, physically securing the stores is a matter where it is necessary to bite the bullet. A storesman must man the gate to the store and insist on (1) authorisation, and (2) documentation from everyone before any issue is made or a receipt accepted. By eliminating this special cause of variation in records accuracy, the imposition of physical security will typically result in increased accuracy of some 20% in a very short time ... say, up from 60% to 80% in weeks. Note that since the shop floor must be capable of receiving service throughout the whole time that staff are working, it follows that the stores must be manned during that time. Managers in industry who take the view that shop floor workers (or one nominated shop floor worker entrusted with a key) are capable of maintaining high stock records accuracy are deluding themselves. They are undermining the responsibility of stores staff; their view has been disproved over the years, again and again.
Dealing next with a warehouse environment, the problems of access are different in nature, since the very job of a large number of the warehouse staff is the storing and removal of material. The rule in warehouse sites is that if a person's job does not of itself require him to handle stock, he should not have access to the place where such stock is stored.
Overcoming the practical problems of imposing limited access.
(1) The Creation of a Secondary Store/Warehouse
One means of overcoming the problem of limited access is to create a secondary store, or mini store, as a small version of the main storage area itself. Access to the main stores is strictly limited, as described above, but access is allowed to the mini stores (though not freely). The records in the mini stores will go wrong quickly. However, it is possible to make a 100% stock count of the mini stores in a very short time, and this is performed every two or three days to correct errors.
(2) Line-side stocks and the use of backflushing
Manufacture involving assembly lines, continuous production and cell production usually relies on "backflushing" to determine the amounts of components used - that is, to determine the amounts of material in stock on the shop floor. For example, suppose that a bill of material showed that Product A consisted of 2 units of Component B, 1 unit of Component C and 1 unit of Component D. Suppose further that 20 units of Product A are manufactured. (Careful account must be kept of spoiled manufacture and production waste.) If so, the stock quantity on the record of Product B should be reduced, or backflushed, by 40 units, the stock of Product C should be reduced by 20 units, and that of Product D by 20 units as well. If the starting quantities are 160 units of B, 80 of C and 80 of D, then the remaining amounts are 120 B, 60 C and 60 D. There are three circumstances where backflushing might be used to advantage in a manufacturing environment: (a) Where "point-of-use" stock is stored - ie "line-side" stock, always kept on the production line at a machine location; (b) Where bulk materials are used in production, especially since withdrawing exactly the correct amount of such material is sometimes difficult or impossible, and (c) Where scrap in production is very variable, so that production runs must continue until enough good material has been produced. Backflushing is, however, an error-prone method, since it assumes a perfect world - it cannot deal with non-standard actions such as shrinkage direct from the assembly line, components removed for a quality tests, or ones borrowed and never returned, or with variations in the usages of the bill of materials. It is consequently necessary to 100% count the stock on a frequent, perhaps daily basis. Minimum floor stock should be issued to minimise this counting task. See below for further information relating to backflushing.
Note that the principles of maintaining stock records accuracy expounded in this course apply to industrial stores and warehouse environments, not retail. The speed of removal (and placement) of stock in a retail stores/ warehouse make the continuous correct raising and processing of transactions impractical. If stock records are maintained for use in a retail operation (including mail order and Internet shopping), the physical stock must be continually checked by verification and the records adjusted where necessary. The situation is somewhat similar to the 'secondary stores' described above. As well, in the fast moving retail opereration, use can be made of the 'low stock' ideas described in Section 5.
Return to Agenda Return to top of Agenda Go to Top of Course Go to end Return To Glossary
2.2 The Value of Variable (Random) Locations
If a variable, or random, location system is adopted, the location associated with each separate occurrence of stock is clearly every bit as important as the quantity. The immense advantage of a random location system from the viewpoint of stock records is that locations in the store or warehouse are continually becoming empty. When they do so, the location is liable to be filled by another product as dictated by the "put- away" procedures of the computer program governing the operation of the system. However, in addition, when a location becomes empty, the fact that it has so become empty can be considered to constitute a free stock count (of zero stock). To make use of these free stock counts, the storesman or warehouseman must report back to the system whenever a location is emptied following a stock withdrawal.
A simple example of a stock location becoming empty is as follows: Suppose Product A's stock record showed the following data at the three locations where it is stocked: (1) Location X, 500 units (the most recent stock) (2) Location Y, 485 units (3) Location Z, 65 units (the oldest stock). Suppose it was now required to pick 80 units of stock of Product A. There are three alternative scenarios. First, there really are 65 units at Location Z, as the record states. Thus the picker successfully picks 65 units of stock from Location Z, picking on a FIFO basis (and picks the balance of 15 units from Location Y.) Location Z is empty - ie the physical stock "count" is zero. The stock record is also 0 (that is, 65 - 65 = 0), and so there is a stock reconciliation. Second, there are only 50 units of stock present at Location Z. The picker picks all 50 units present, leaving Location Z empty, (and picks the balance of 30 units to make up what he needs from Location Y). When the zero stock count is notified to the system, there is a stock adjustment of 15 units (that is, 65 - 50 = -15, against an actual count of 0). Third, there are 120 units of Product A present at Location Z. The picker is able to pick all 80 units from the location. Although in this last case there is not a count of zero, the system can clearly detect an error in the record at Location Z, which can now be separely investigated ( ie the record shows that there is insufficient stock, even though the picker has found 80 units.)
This subject is dealt with again under the Cycle Counting of Randon Locations.
2.3 Equipment for Weighing & Measuring
To help eliminate further special causes of records variation, and allied to the question of tolerance above, consideration must be given to the inherent accuracy and the calibration of the devices available in the facility for weighing and measuring. As everyone knows, there are a large number of machines on the market capable of counting and/or measuring, ranging from basic weighing machines to sophisticated computer-controlled electronic sensors.
W e i g h i n g M a c h i n e s
Accurate "sample scales" are used for small quantities to find the unit weight of products. Large electronic scales weigh the main stock, with direct links possible to the computer. Care must be taken to account correctly for the container weight (the 'tare'), and to ensure the items' weight is not distorted by oil, wetness etc.
S t a n d a r d O v e r - P a c k a g i n g
This is the simplest of all counting methods, obtained through the provision of packages which can contain only exact multiples of items (ie the egg box, holding 6 eggs).
T e c h n i q u e s S p e c i a l t o t h e P r o c e s s I n d u s t r i e s
Oil and chemical companies are required to measure bulk liquids, often choosing to do so by flowmeters or even simple dip tapes. Many of the methods used are acknowledged as being problematical, with comparatively wide tolerances arising inherent in the techniques themselves. Problems are often compounded by the need to take account of the temperature of the liquid, and then the further need to assume that the temperature is uniform throughout the material's bulk. The records accuracy of such companies' stock is frequently largely dependent on the measuring devices used. It is interesting to note that when a tanker discharges oil at a refinery, an 'arbitrator' appointed by the two parties (ie the refinery manager and the tanker captain) is on hand to decide the quantity of oil deemed to have been discharged for payment purposes.
Return to Agenda Return to top of Agenda Go to Top of Course Go to End Return To Glossary
2.4 On-Line Data Processing
The desirability of adopting an on-line IT system to eliminate special causes of error and prevent yet further variation in the records stems from the fast feedback provided relating to on-line errors in the transactions, and the far greater confidence that the records are up to date when access to them is wanted. Batch data processing, by contrast, means that data 'events' are recorded on a device (whether a paper form or an electronic recording collector) and input to the computer records in bulk. With batch processing, the data transactions are deliberately held back until a reasonable quantity of them have been collected, so that input, although efficient, is usually made several hours after the events to which they relate. Many companies insist that batch data is at least input within two or three hours, and certainly before the start of any new planning run. On-line means that instant access is made to the computer holding the stock records. That is, data is notified to the computer as soon after the event as possible. Except in a totally automated environment, however, there is nevertheless a time delay between the event and the recording of it. (The delay may be seconds, minutes or, in a slacker environment, one or two hours.) A very important further advantage of on-line is that the person responsible for the data is also the one who inputs it to the database. As a consequence, if the computer immediately detects an error in the transaction, an error message is signalled at once to the person who has input it. That person can then correct what was wrong, saving a great deal of time that would otherwise be spent in identifying him and asking him to effect a correction.
Tied up with the question of on-line and batch data processing are the matters of data recording and data collection. Thus:
P a p e r F o r m s.
A dominant method of collecting data about stores and warehouse 'events' is through the completion of paper forms, which must then be taken to a VDU so that the data recorded on the form can be input to the computer.
V D U s.
Within the main stores, the most common method of obtaining on-line access to the computer is via an ordinary VDU situated at a convenient point, which must be connected to the computer by coaxial cable and the use of a computer network.
R D T s.
A radio data terminal is an outstanding alternative to a fixed-position VDU, being essentially a hand-held VDU linked to the computer by a radio signal. RDTs can operate from over a mile distance. A major problem with RDTs which makes them unpopular with storesmen is that they occupy a hand of the operator - materials handling in the stores/warehouse requires the use of both hands. (RDTs are used by HM Customs at ports and airports.)
D a t a C o l l e c t o r s.
These are devices similar in appearance to Psion or Sharp electronic organisers (not dissimilar from large calculators), and are used for directly inputting and storing transaction data. When the devices' memories are full, or perhaps at the end of a shift, the data collected so far are downloaded via an input port to the computer as a stream of data transactions.
L o c a l D a t a C o m m u n i c a t i o n s.
'Local communications' here means within the company - ie between factory, stores and the company's computer installation. Coaxial cabling and radio links are common.
R e m o t e D a t a C o m m u n i c a t i o n s.
Data may be sent to and from a depot or other outlying transmitter or data cable, through a so-called 'value added network', or VAN. The VAN can be regarded as a network of computer post offices. That is, the network subscriber who wishes to send data dials into the system and transmits the data, including the (electronic) address of the intended recipient. The message or data is directed to a computer close to the recipient's geographical address. The recipient accesses his local VAN computer on a frequent basis to see if any message has been sent to him, and, if it has, reads it into his own computer. There are four VAN networks in the UK, operated by INS Tradanet, AT&T Istel, BT and IBM. Compatibility problems between these networks have now been largely resolved. In order to send any data via a communications network, it is necessary to follow strict rules relating to the electronic format of the 'message', so that it can be properly handled by the complex computer software program governing the network. The rules are known as the network's protocol. The standards for the VANs are X.12, X.400 and X.435. The last two can deal with both EDI and E-Mail (electronic mail). The remaining technical issue concerns the software employed by the sender and recipient themselves to translate and interpret the information being sent. A number of 'EDI translation' software packages are on the market. Most of them, such as Sterling Software's Gentran, can access all four VAN networks.
Return to Agenda Return to top of Agenda Go to Top of Course Go to End Return To Glossary
2.5 Bar Coding and RFID Tags
B a r C o d e s
The familiar bar code is the representation of a numeric, alphabetic or alphanumeric code by a pattern of dark and light stripes, with 'start' and 'stop' characters at either end, and which can be interpreted, or read, by a light scanning device called a bar code reader. Bar code readers are either contact or non-contact. Contact readers such as those used in retail shops are also called fixed beam readers, since the device needs to be very close to the bar code. They are comparatively inexpensive (£100 +). Non-contact scanners, or line scan readers, work by repeatedly reading the code with a laser beam fired by a gun, perhaps mounted on a truck, until the reading is error free. They cost about £1000. There are a dozen or so different bar coding systems for assigning a code to a material. One used extensively in the warehousing of consumer goods, including the outside carton packaging of groceries, is termed Interleaved 2 of 5. It is numeric only, and requires the code to comprise an even number of digits. With Interleaved 2 of 5, even numbers are represented by the white stripes and odd numbers by the dark bars. Its advantage is its physical density. In industry generally, however, there is a preference for the Code 39 system. This is capable of encoding numbers and letters. Each character is represented by a group of 5 bars and 4 spaces, and has an in-built check to eliminate mistakes in the physical reading and interpretation of the code by the bar code reader. Other bar code systems are EAN (European Numbering System) and UPC (Universal Product System). UPC was devised by IBM in 1973 and is the one used in groceries in supermarkets. Its advantage is that the code does not need to be on a flat surface to be read by the reader. Bar coding in the stores or warehouse is not always successful even when those attempting its implementation have carefully assessed that it will be. There are three issues.
First, there is the matter of ergonomics. Ergonomics is the science of 'man-machine interaction', and here means how codes are to be assigned, how (literally) they are to be attached to the objects and locations in question, how the codes are to be read, and what equipment is to be used. It also encompasses the nature of the computer system that will read the codes and how associated data, such as quantities, are to be recorded.
The second issue relates to the physical nature of the items actually to be bar coded and seems to be the most critical of the three If there is a wide variety of shapes and sizes, and if many items are of irregular shape or have unsuitable surfaces, it will be difficult to devise satisfactory, consistent ergonomically sound procedures.
Thirdly, if it is intended that incoming raw materials are to be bar coded by suppliers, their competence and willingness to do so must be considered (or, at least, their willingness to apply bar code labels and documents supplied by the company).
If bar coding works well and easily, without a continual struggle to keep it going, there are two advantages to its use. First, self-evidently, material and location codes are read correctly and more easily. Secondly, the reading process ensures that each transaction relating to an activity is indeed raised, and is not forgotten, and that it is then input to the computer system in a timely manner. In other words, it has the ability to eliminate a frequent cause of stock records inaccuracy, namely that of lost and missing transactions.
R F I D T a g s ( R a d i o F r e q u e n c y I d e n t i f i c a t i o n T a g s).
By 'identification' is meant the attachment of a small "tag" bearing the code, and potentially much other data, of what is to be identified, and the subsequent reading of the tag code and data at some later stage by a tag reader. The physical tag attached to the object may commonly be a label, in a flat, thin, flexible ticket, often referred to as a smart label, or may take other forms such as a glass bead, depending on the application to hand. Important attributes of tags are that they are robust and capable of functioning in extremely harsh environments and that they are reusable and can last for many years. The code and other data associated with the tag is read by a tag reader, or special tag interrogator, a primary function of the interrogator, or reader, being to excite a component within the tag termed its antenna. Although the technology associated with RFID tags and interrogators is changing rapidly, as at the date of this on-line course, the microchip used in the tag is likely to be a silicon microprocessor (chip) and the antenna is likely to be formed from conductive carbon ink. The silicon microprocessor or chip will be attached to carbon - ink electrodes at the back of the tag itself, or at the back of the smart label. Note particularly in an RFID tag that a tiny enclosed battery is optional. That is, a tag may have a small lithium battery to boost power. Tags with batteries are referred to as active tags and those without as passive tags. Power is transmitted to the tag in the first place from an electric field created by the tag interrogator. Data is transferred from the tag to the interrogator through the modulation by the tag of the interrogator signal. With their extra power, active tags are able to communicate with an interrogator over considerably greater distances than passive tags (many thousands of feet rather than only tens of feet). Active tags are also capable of carrying and conveying greater amounts of data (thousands of bits rather than tens).
Not surprisingly, active tags are more expensive than passive tags. Cost is currently a major issue in RFID technology, especially as it concerns its widespread adoption in retail. A critical milestone in the practicality and acceptability of RFID technology has been the adoption in late 2005 of the GEN2 data technology standard and the ALE standard. GEN2 governs the basic tag reading technology essential to the production of tags themselves and tag readers. ALE deals with the collection, management and routing of data; it addresses the problem of huge amounts of raw data generated by RFID readers - readers can make multiple readings of the same tag in a fraction of second, so that this "dirty" data must be filtered. In summary the key benefits of GEN2 and ALE are that they give the ability to read RFID tags quickly and simultaneously. Finally, and most importantly, we see from the technical nature of the interaction between the RFID interrogator and the RFID tag, that two major advantages lie with the technology and distinguish it from bar coding. These are:
First, that in order to read a tag, it is unnecessary to have a direct view of it. Communication is by electrical waves and antennas, and line of sight is no more required than it is required of a radio in order to broadcast to it a programme from a transmitter.
Secondly, it is possible easily to read tags which have been attached to a succession of irregularly shaped items which would be unsuitable to bar code reading. The examples in everyday life typical of the application of RFID tags are: car tagging for toll booths; hospital patients; criminals on licence; airline luggage; library books; the tagging of wild and domestic animals; and marathon runners
For the stores, one critical application of tags is in making use of the ability to read simultaneously the identities of all the tagged components of an incoming or outgoing load merely by scanning it from a distance with the tag interrogator.
A second application is the ability quickly and easily to verify and count stored stock, as in cycle counting or in the conduct of an annual stock take. An example relates to a national company distributing wines and spirits, which wished to double check assembled loads for correctness on its vehicles before despatching them to customers, and to a major retailer receiving loads of garments hanging on rails at its major stores from its distribution warehouse.
Besides the examples above, there are an ever increasing number of other applications involving the simultaneous, mass reading of palletised loads at the point of despatch and the verification of loads at their destinations. In order to commence a move to RFID, the stores supervisor might first attend a one-day course on the subject held at the DTI's RFID Research Centre in Bracknell, Berks.. In addition, it is possible to see RFID in action at an RFID demonstration site run by Unipart and others at Oxford. A contact name is Mark Howard at Unipart on 01865-383440. A hands-on introduction to the technology can also be gained from the "RFID in a box" starter kit obtainable from the consultancy firm Manhattan Associates (London 0870-3514770 or Bracknell 01344-318000).
Return to Agenda Return to top of Agenda Go to Top of Course Go to End Return To Glossary
2.6 The Working Environment
Although the environment in which manufacture takes place in many factories is simply a fact of life, the manager should be aware that noise, dirt, heat and a shortage of space will work against the accurate completion and submission of transactions. Every effort should therefore be made to eliminate unnecessarily stressful conditions, and where complete avoidance is not possible, at least to mitigate their effect. The atmosphere in which the storeman is to work - its freedom from disruption and distraction - is as much a part of the stock recording system as the operation of computer programs and the clarity and correctness of the computer transactions. Four areas to observe and, if necessary, reorganise and improve are: (1) goods receiving from outside; (2) goods receiving from the shop floor; (3) the point of physically picking and assembling stock; and (4) packing and despatch. Questions critical to the conduct of activities in these four areas are considered below under a number of headings. In each case, the manager addressing the stock recording system should observe the various areas for himself and listen to the testimony and opinions of those who work there. As stated, the recording and submission of data transactions should be undertaken in an atmosphere free of stress and chaos - it should be a calm atmosphere conducive to error-free, accurate work.
Provision of Ample Space:
The degree to which there is loss of productivity, and indeed the degree to which there is an increase in liability to accidents and errors, increases as the percentage occupation of the facility itself increases. The increase has been graphed. The graph shows that liability to accidents and error does not rise linearly as the percentage occupation of the workspace rises. Instead, liability rises exponentially. That is, as occupation gets higher and higher, the probability of accident and error becomes nearer and nearer a certainty.
Housekeeping:
Just as clearing away clutter and sweeping and cleaning are principal rules of lean production, so they should be in stores/warehouse management to achieve an environment conducive to accuracy. Damaged cartons must be seen to and containers closed. Unstable loads must not be allowed on racking or as floor stock. Untidiness is an invitation to allow stock to "wander" to the next location or behind crates.
Noise, Heat and Cold:
Work should be redesigned as far as possible so that storemen required to perform work under these conditions can at least complete and submit actual transactions back in the office away from the stressful areas. Every storeman should have his own desk.
Lighting:
Inadequate lighting very obviously leads to errors in recording and counting. Have lighting redesigned by an expert - the technology in this area is constantly on the move. As well, remember that natural light is far preferable and far less stressful.
Item Identification:
Ensure that all labelling is clear, employing a large, clear font.
3.1 The Total System
There are certain characteristics of the stock recording system which make a wholly IT oriented approach to its design and implementation quite inappropriate. One characteristic is the detachment of the data processing from the actuality of stock movement - ie no data need be verified in order for a real-world "event" to take place, and no event is ever invalidated or reversed. (For example, stock can be picked and issued without the literal completion of a transaction, and if a transaction were to be completed wrongly, the stock would remain duly picked.) A second is the reliance on a great many system operators, with diverse professional interests and from many different backgrounds, working in a stressful environment - ie storemen, drivers, cycle counters and others. Yet a third characteristic is the potential informality of the data transactions themselves, at least, usually, in the first stage of the data trail as paper forms. As a consequence, it is necessary to consider the stock recording system from a wide, outside in viewpoint, and include for consideration not only data procedures and IT, but such other matters as the recruitment and capability of staff, the procedures followed by external suppliers and drivers, and the data they raise, the role and responsibility of shop floor staff, staff bonus schemes, training and education, links with sales department and customers to whom goods are despatched. As stated, the approach to the stock recording system which encompasses this total viewpoint is referred to as outside in. In considering stock records accuracy, then, all facets must be considered. Thus it is pointless to have excellent data measuring and recording devices but to employ dyslexic staff to operate them. Although the outside in view of the system is important, naturally one cannot sideline the formal data processing system itself. The data processing aspects of the system are considered below.
3.2 The Data Processing Procedures
The data transactions - records of events - are the essential means of ensuring the system itself is backed up by actual data reflecting the real events that happen. It is normal to organise transaction records into files of data by transaction type, and to relate each such transaction record to the main stock record via the Item Code. The stock data on a transaction record (perhaps a transaction dealing with an issue, say "Type 05") will obviously depend on that transaction type and the product in question. If Type 05 was "Stores Issue", the data associated with an "05" might be ... Transaction Type Code 05 : Product Code; Date and Time of Transaction; Issue/Order Number Quantity Issued; Unit of Measure.
The main stock record and the various associated transaction records will be related through their common item code. Other data will be associated with other types of transactions. Some data are present to assist with the task of data reconciliation should things go wrong - see Audit Trails below. A separate computer program is needed in the system for each type of transaction. A large number of transaction types will typically be defined to the system. By checking the names of the transactions it is possible to assess broadly whether the system is sufficiently comprehensive. Failure to provide a comprehensive set of transactions and procedures to meet all circumstances will result in the system being side-stepped and in an immediate deterioration in stock records accuracy. [Note however that there is an alternative view to this. Roger Brooks and Larry Wilson in their book Inventory Records Accuracy state that the number and diversity of transactions should be minimised, with, say, a given transaction serving several situations. For example, Transaction X may relate to the issue of stock from the stores. There may be several different, distinct stores activities which relate to the issue of stock, each activity therefore calling for the use of the common Transaction X. The training of staff in such a system must clearly be thorough.]
T y p i c a l T r a n s a c t i o n T y p e s
Examples of actual transactions in two categories are given below. Each example corresponds to a data processing program that allows the on-line or batch update of the database. It will be backed up by other data processing programs: for example, programs allowing personnel to make enquiries; programs to support cycle counting; by programs producing regular reports (including a report of the stock records accuracy figure itself).
Category 1 (stock movements originating within the Stores)
Issue material from stores/warehouse to the factory;
Issue material to goods outward for despatch to a customer;
Return material to outside supplier (+ reason);
Cycle count of stock;
Scrap obsolete/spoiled stock in the stores (+ reason, such as quality, deterioration etc;
Move or consolidate stock within stores;
Direct stock adjustment (+ reason).
Category 2 (stock movements originating outside the Stores)
Process batch charged / completed;
Receive material from the factory into stores;
Supplier receipt adjustment;
Material placed in quarantine;
Direct stock adjustment (+ reason);
Complete & report works order in the shop;
Start / complete machine operation in the shop;
Start Works Order in the shop;
Deliver goods outward to customer;
Inter-factory movement of material;
Receipt of external material direct from supplier;
Deliver material to stores;
Receive material from outside supplier into stores;
Return of unused material from the shop floor to the stores;
Report loss
Chemical batch formulation / BOM adjustment;
Cycle count (WIP).
Return to Agenda Return to top of Agenda Go to Top of Course Go to End Return To Glossary
3.3 Procedures & Responsibilities for Raising Data Processing
Defining Procedures for Transaction Processing
Stock records accuracy will not be achieved without the understanding and commitment of those involved, backed up by defined responsibilities and physical tools. In this sense it is a mistake to view the task purely as a systems or technical one. On the other hand, it is a mistake also to believe that goodwill alone will accomplish the job. Procedures must be set up which make it easily possible to record what needs to be done. They must be documented, simple to follow and comprehensive. If personnel do not know how to cope with a situation through the formal system, they will create or revert to an informal method in order to make progress with the physical side of their jobs.
Standardising Terminology for Units of Measure
It is required that everone connected with data input should in all circumstances use the same abbreviation for each term denoting a unit of measure, such as cartons, bails, drums, square feet, rolls "each",and so on. The supervisor should publish a list stating the only allowed abbreviations. As well, the publication should standardise the writing of such expressions as "5 boxes of Size A brackets containing 50 in each box" (say, 5 b Size A Brackets @ 50 ea). Thorough practical training must be given to staff in the new standards. Finally, it is recognised that many documents raised outside the stores, say by suppliers or shop floor staff, have marked on them informal manual inscriptions. To make it instantly clear which markings on such documents are the work of stores staff, it has been suggested that stores personnel should always use green pens so that anyone else in the stores/warehouse will know at once what has been incribed by his colleagues.
The Validation of Data
Data should be validated whenever an opportunity presents itself. The scale of the validation and the usefulness of the error messages will be an indication of the quality of the background software. Examples of general data validation are (1) general data field validity; and (2) cross- checking against master data files. Other examples are:
(i) Message Patterns
It is often possible to define a sequence of transactions that must - without variation - be followed by a particular order throughout its life. Then, to ensure that no step is missed along the way, an internal computer check can be made as each successive transaction in the defined sequence is reported. Missing or out-of-order transactions are then picked up on. A typical example relates to the succession of procedures taking place on receipt of an external supplier delivery. Two further examples might be (a) the required put-away transactions issued as instructions in a variable (random) location system, and (b) the required stock retrieval transactions also issued by a variable location system.
(ii) Reasonableness Limits
A not uncommon error in the operation of the system is the entering of a product quantity that is grossly too large or grossly too small. For example, a storesman might receive two boxes of 300 widgets and enter `2' instead of '600'. "2" is not a reasonable receipt quantity for these items if they are usually ordered in multiples of 300. It is suggested that reasonableness limits be associated on the database with each transaction type for each product, especially raw materials. If the limits are broken, the computer can issue a warning.
(iii) Duplicates
If a second transaction is input within (say) 48 hours of a previous transaction. and is otherwise identical to it in every regard, the system can issue a warning that the second one may be a duplication of a transaction already submitted and processed.
Return to Agenda Return to top of Agenda Go to Top of Course Go to End Return To Glossary
3.4a Goods In
C o n d u c t o f G o o d s - I n
The very first action the Goods-In Supervisor must take when a vehicle arrives is to check that the load is indeed for his company. The next action is to count and verify that the number of boxes, crates, pallets etc on the driver's delivery note tally with the number delivered. Next, two actions are taken as follows (though the sequence of the actions might be reversed - see below).
(1) The detailed receipted material is checked against what the supplier says has been sent on his Unpacking Note. (2) The transaction data from (1) is input to the computer. The principal issue in the receipt and validation of incoming loads is the extent to which the Goods In supervisor should go in opening packages and counting out material. The supplier's own assertion of what has been sent should be accepted only if the supervisor has evidence that his (the supplier's) procedures are sound. Even if they are, there should still be spot checks. And if any evidence emerges that the supplier's paperwork differs from reality, the supplier's deliveries should be thoroughly checked until he is back in control. If material is received in standard containers (the "egg box principle"), the stated number should be accepted without counting. Any non-standardised cases/amounts, however, should be counted. If weigh counting is employed, it is likely that, in the event of small discrepancies, the supplier's totals are more accurate than the receiving company's. Weigh counting, incidentally, can be effective for difficult materials, such as powders, liquids and bails. In verifying the received amount versus the stated despatched quantity, a useful tip is to circle (with the stores' standard green pen) the goods despatched figure on the supplier's documentation, making sure this is not confused with the order quantity or any quantity "to follow". The receipt and staging areas where goods are to be off-loaded and counted should be well lit and tidy. And the rule should be instituted that goods received today must be verified today and the database updated today. The rule prevents mix-ups and mistakes, even though occasionally it may be necessary for the stores supervisor to draft in extra staff. See also the responsibilities of the purchasing manager, below.
Note finally that the 'INCOTERM' (International Chamber of Commerce Terms of Sale) agreed by the company with the supplier or customer regarding the responsibilities for either delivery (ie with the customer) or receipt (ie with the supplier) should, if possible, be DDP (Delivered), not EXW (Ex-Works). A major problem with Ex Works* arrangements, by which the company's customers will collect the goods from our loading bay, is that frequently enough he does not turn up at the time, or even on the date, promised. Consequently, we find the goods, which are stacked and waiting for him in the despatch area, to be an obstacle, and we are forced to continually shift them out of the way. If our own company agrees to deliver the goods, however, by specifying the Incoterm'DDP', everything regarding despatch is is under our own control, from picking to assembly to loading (on our own transport). To be sure, there is a cost associated with delivery and this must be taken into account in the negotiation of the selling price. The purchasing manager and the sales manager must be made aware of these points and the impact of the selected INCOTERM on the stores'/warehouse's operations. [* Note that under EXW, (1) the only wrapping required of the goods is 'immediate' wrapping, not wrapping for the journey, and (2) there is no obligation to load the customer's vehicle. If the vehicle is loaded for the customer (as a favour to him), who is to be responsible if it is done negligently? It is urged that the stores/warehouse manager should obtain the booklet published by the ICC, available for about £26 setting out the INCOTERMS terms in full. This may be ordered on the Internet through Amazon.co.uk. It also strongly suggested that the manager wishing to become seriously involved in using an Incoterm obtain the commentary on the Incoterms written by Jan Ramberg, also available from Amazon.
G o o d s - I n T r a n s a c t i o n P r o c e s s i n g
The completion and processing of transaction data relating to goods-in is the biggest source of error in raw material stock records. One reason for errors is that the sequence in which material is listed on the company's purchase order or the supplier's advice note will almost never be the same as its sequence on the Unpacking Note. Another reason is the use by the supplier of different codes and, even, different names. To help overcome these difficulties, it is suggested that a provisional transaction set is raised first, before the physical count, the data required to raise it being obtained directly via computer from the purchase order or advice note. The data from this provisional transaction set must then be printed out by computer as a check list to assist in the verification process of the physical material. (The check list might be augmented with other data such as units of measure and cross reference codes.) Clearly, somewhat complex software is needed to accomplish this. Such software very obviously must also allow the provisional transactions to be amended before confirmation.
The need for an overall system approach was emphasised a couple of years ago by a report relating to lost stock by the MOD. Thus some time ago, the Ministry of Defence admitted that thousands of key parts relating to the UK's Trident Nuclear weapons system had 'gone missing', some of the parts being cable subassemblies required to link different sections of the weapon. According to the MOD, "the problem came to light after checks of incoming supplies at the Trident base in Faslane, Scotland, exposed "discrepancies in receipts. Parts which left the manufacturers in the US were missing when the consignments reached Scotland". The parts were British Government responsibility when they disappeared. Detailed enquiries with the shipping authorities, and exhaustive checks of stock holdings and historical records failed to track down the missing items and they were formally declared 'lost'.
3.4b Put-Aways
C o n d u c t o f P u t a w a y O p e r a t i o n s
In putaway operations, errors can be reduced by the adoption of on-line communication devices such as RDTs and by the adoption of bar coding. To reduce the possibility of lost transactions, the storeman would do best to complete/scan in the transaction data at the time the material is being assembled for the putaway, rather than after the putting away has been accomplished. In putting away in a fixed location facility, errors will be rare since the presence in the location of the very same items that are being put away will help confirm validity. (To assist FIFO, when material is placed in the location, partial pallets and partial cases should be placed on top and/or in front, with labels facing the picking face.)
P u t a w a y T r a n s a c t i o n P r o c e s s i n g
Problems in putaway arise in variable (random) locations, however. An obvious worry is that stock so placed under these systems, but then wrongly recorded, will become "lost", since its actual location is no longer capable of discovery through the system. A number of safeguards are available here.
First, if a storeman is directed by the system to place stock in a location which is supposedly empty, but then finds it is occupied, he must abort the placement and report matters so that an audit of the location can be carried out.
Secondly, periodically, all locations recorded as being empty on the system can be printed out, and a quick visual check made to verify that they are indeed empty.
Finally, a powerful verification system sometimes known as the Magic Number Method, based on secondary location codes can be instituted, as described below.
The Magic Number Storage Method
A unique, randomly generated secondary code is assigned to every stores location, and tagged to the end of the primary location code. For example, the arbitrary secondary code 3729 might be generated and tagged to primary location code A08G4 to give a new code as follows: A08G4-3729 When the storeman puts stock away at a location or retrieves it, he must record on the transaction the full code of the location, including the tag. The computer checks the validity of the code and rejects the transaction if the primary code and the tag do not match. For example, suppose the 200 units of product 'P50' was supposed to be placed at Location A08G4, but instead by mistake had wrongly been placed at Location A08F4, Location A08F4 having tag 6881. The storeman records on his transaction "A08G4-6881"(he thinks he has placed it correctly at A08G4, and he also records the only secondary tag he is aware of from the putaway.) The transaction of course is rejected by the computer because primary code A08G4 corresponds to tag 3729, not 6881.
3.4c Picking
C o n d u c t O f P i c k i n g O p e r a t i o n s
One rule that will promote greater accuracy in picking is that orders should not be released for picking until the day the picking and issue/despatch is to be accomplished, and that when an order is so released for picking, it should be completely processed on that day. Picking orders over extended periods, or, even worse, kitting days in advance, is error prone due to mix ups, the robbery of kits and difficult cycle counting.
A second rule is that no pick should be directed if there is insufficient stock. If there is insufficient stock, the requirement should be written to an "outstanding" file. If a vehicle or container is to be loaded with material that has been picked and staged, and any item is missing because it is still under manufacture in the factory, do not load. Wait until everything is to hand. Similarly, to verify a loaded vehicle or filled container, empty it. (And, by the way, never allow a storeman to add a few more "for good luck".) Because wrong picking is so visible and has such a major bad effect on the perception of service by our company, pickers must be highly focussed on the job in hand. For that reason, order picking should not be interleaved with other stores' jobs such as putting away or cycle counting. (Batch and wave picking are also prone to error for these reasons.) And to help the storeman concentrate when picking a multiline order, a separate tag or ticket should be generated, perhaps with extra information, for each item on the order to be picked. This facility reduces the risk of an order line being missed. Picked orders should be checked by whatever means is best, such as by counting pallets or cases, or even by weighing. RFID tags have been especially successful in pick/issue checking, since the signals from the items loaded can be read to check the items' presence - ie line of sight. is unnecessary.
P i c k i n g T r a n s a c t i o n s
When a picking order contains the location, code and quantity of each item, and a pick has been sucessful, it seems to be inviting error to require the data to be fully input to the computer. Surely, all that is necessary is to confirm the data by exception.
V o i c e D i r e c t e d P i c k i n g
Voice directed picking is a highly effective, highly accurate and increasingly popular technology that has many advantages in both stores and warehouse operations. With voice, workers wear a headset, earphones and a belt-attached portable computer which enable them to hear instructions from the computer and to speak words of confirmation as to action taken. The computer communicates with the company's Warehouse Management System (WMS) to retrieve the locations from which picking is required, then the identities and quantities of items to be picked, and transmits this information as speech to the system. The worker proceeds to the designated location, reads out a check digit to confirm its correctness and receives further instruction as to the identity and amount of what is to be picked. He confirms the pick by voice and the system proceeds to the next location/item. The advantages of voice directed picking operations are that the storeman or warehouseman has his hands and eyes free. Picking accuracy is far higher than with a paper based system, productivity is substantially higher and accidents are reduced.
An alternative to voice directed picking is pick to light. With light directed picking, a small light is illuminated at the places from which picking is to take place, speeding up the travelling process. The system is more expensive and less flexible than voice.
3.4d Other Procedures
Procedures must be laid down for every reasonable situation the storeman will encounter in his day-to-day work. There is only one correct way! For example, procedures must be in place to deal with
returns from customers;
returns to suppliers;
outsourcing;
the direct receipt of materials from suppliers on the shop floor;
items taken for sampling;
consolidation; and
scrap.
Two further subjects are selected at random for comment as follows:
S u b s t i t u t i o n s
The rule must be laid down firmly that substitutions are never to be made by stores or warehouse staff (though staff may wish to use their knowledge of products or shop floor practice).They must only originate with changed official documentation raised by production planning or sales service.
I n n e r P a c k s
Inner packs ("ip") are defined here as packages within a larger box or case, each inner pack containing the same number of identical items. An error can arise when an inner pack is confused with a single item. It is recommended therefore that inner packs be removed from the case packaging in which they have been sent, and counted (as individual packs) into a well-labelled, well-marked container. The total number of individual units can be calculated and checked against the delivery note.
K i t s
Kits are packages or cases each comprising a specified set of different items, each set being identical (A,B,C & D; A,B,C & D etc). An error sometimes made is to remove a required individual item from a kit without realising it is indeed part of the kit. Kits must be clearly labelled as such and the removal of any unit from a kit supported by documentation.
4. Interpreting Transaction Trails
4.1 Introduction to Transaction (Audit) Trails
Suppose that at the beginning of the morning according to an inventory record there were 10 heavy duty batteries in stock. Later, at 10.00am, a system transaction is submitted and processed for a receipt of 60 batteries from the supplier. Shortly after this again, at 10.30am, a transaction is processed for 15 units issued to the shop floor. The record now says that there are 55 heavy duty batteries in stock. The purpose of the transaction trail, or audit trail, is to show to an investigator at some time later how a position such as this was arrived at. In the small example just recounted, the transaction trail will display (if it is output to the VDU) the "before and after" stock levels that were recorded on the record immediately before and immediately after each transaction as it was applied, and the data fields associated with the actual individual transactions that were applied. The transaction trail is usually displayed with the most recent transactions at the top of the screen (ie in reverse order), since this makes 'scrolling' down to older and older transactions easier to follow. That is, the trail should be read from bottom to top. Thus:
Trans. 2 10.30am Issue to Shop 15 units [Stock Before 70; Stock After 55 units]
Trans. 1 10.00am Receipt ex Supplier 60 units [Before 10; Stock After 70 units]
The entries in the transaction trail are a history of what the stock record really said as the result of the modifications, or updates, that were made to the record. The entries relating to the stock record totals are not calculated by an independent program as the audit trail request program is run, as a presumption of what the stock record ought to have said. The totals are what the stock record totals actually were, at the time, as transactions were received. In summary, the stock recording system should provide the easy ability to display and examine the audit trail of any product on the VDU at any time. The maximum number of (increasingly older) transactions capable of being retrieved will clearly be governed by what is on the computer disk. It may be necessary to page back through several VDU screens to see them all. As stated, the order in which transactions are displayed is usually (but not always) the reverse order in which they were input to the computer. That is, the chronological order of transactions is from the bottom of the screen to the top. Note that the data input times recorded by the computer clock relating to the physical receipt of the actual trasnactions are also displayed (not shown above). Except in totally automated warehouses, the times at which transactions are input to the system do not correspond to the times at which the physical events to which they relate actually occurred. They are later by seconds, minutes or hours. More to the point, in batch systems, and even in on-line systems, it is possible for the transactions to be input in a different order to the order of the real events.
4.1a Negative Stock
In the two transaction above, if the storeman in charge of goods-in had been unable to submit the transaction for the batteries receipt until 11.00am, even though the delivery was made much earlier, the audit trail might read as follows instead:
Trans. 2 11.00am Receipt ex Supplier 60 units [Stock Before - 5; Stock After 55 units]
Trans.1 10.30am Issue to Shop 15 units [Before 10; Stock After - 5 units]
We see that the stock record from 10.30am to 11.00am in the case above was negative (ie - 5units). Negative stock balances due to out-of-order or late transactions are potentially very common. It is vital that the stores or warehouse manager should investigate very carefully, and concur with, the action taken by his software when this condition is encountered. Some software packages prevent the condition occurring, by holding in abeyance the transaction that would create the negativity (in the example above, Transaction 1 would be suspended at 10.30am until Transaction 2 was submitted at 11.00am.) Other packages simply block the transaction that would create the negativity, issuing an error message. There is even some software, believe it or not, that processes the transaction, and then overwrites the negative stock figure obtained with 0 (zero stock), on the grounds that 0 must be more correct than the 'impossible' negative figure! (There is much the stores/warehouse manager must contend with.)
In summary, it is important that the person responsible for the stock records system should understand how negative stock balances can arise and what he should do about them. As stated and explained above, the principle cause is a timing issue (ie 10.00am v. 10.30am v. 11.00am above). The general advice if this is the reason is to do nothing, since the record will be corrected when all the late transactions have been received, and if a correction is made in the meantime, when the late transactions are indeed received, it will be found that the correction really has made the record wrong. Note that a common cause of negative stock is found with a variable (random) storage facility where the stock of the same product can be held in several locations. Thus the stock record of Product X is 100 units in Location A. The storeman is required to pick 50 units from Location A, and does so, but by mistake submits an 'issue' transaction of 50 units from Location B. Although the total stock of Product X remains correct, the location record for Location B is negative and incorrect. Similar situations can arise when stock is being transferred from one location to another. Negative stock balances such as those described might be described as location-level negative balance problems. They are generally easy to investigate, since the transaction at fault is usually the one which created the negativity in the first place. Item-level negative balances, however, are where the total quantity of stock is genuinely shown as a negative. The reasons for this are not self-correcting, as they are with timing related errors above. Instead they include a variety of mistakes such as the over-reporting of scrap quantities, cycle count adjustment errors, the over-reporting of production (especially when using back-flushing), the over-issuing of material to production, duplicate transactions and the over-issuing of material to customer orders. Item-level negative balances are not self-correcting, and should be carefully investigated. Again, the culprit is likely to be the transaction that created the negativity in the first place, but the investigator should make sure the wrong transaction has not created other errors elsewhere.
A potential problem with a negative stock balance is likely to be the behaviour of a production planning or stock replenishment system when the negative figure is read by such a system. Naturally, being a planning/replenishment system, the arithmetic of it will create a stock request for more material than is needed in actuality. There are occasions when this might be serious. For example, if the degree of negativity is, say, - 5 units, and logic of the planning system is to call up a complete replenishment lot of 1000 units to cover this, the company will be overstocked for a considerable time if this is a slow moving product. Again, if the product concerned has a deep bill of materials, any unnecessary order initiated by the negative stock may potentially create many further orders for its component products deeper in the bill of materials.
Finally, let it be said that the person responsible for the stock records system should (1) know exactly what the system does in these circumstances; (2) ensure all negative stock balances are identified, either through the system itself or via a separate report, and (3) correct any negative stock balance, and any other damage that might have resulted, only after very careful thought and consideration.
Return to Agenda Return to top of Agenda Go to Top of Course Go to End Return To Glossary
4.1b Four Reasons for Investigation of an Audit Trails
Four reasons for conducting a reconciliation are:
1. Investigation of a Count Variance
A count variance either during cycle counting or as a result of an annual stock audit means that there is a difference between the figure recorded on the stock record and the amount of stock counted. Variances will be dealt with in due course. For now, we may simply say that the investigator must use his knowledge and experience to find the reason for it. A primary tool to assist the process is the transaction trail. It is highly desirable that access to the file be on-line and that a sophisticated means should be provided to filter the query. Filtering means accessing only those selected parts of the transaction file of interest - say, only transactions after a certain date, or only transactions relating to raw material receipts. See also part (iv) in Section 5.12.
2. Part of a Control Group
The stock of a very small group of products is counted every day and the counts compared to the stock records. Discrepancies are resolved at once in the factory or warehouse. The purposes of doing so are to monitor the operation of the system and keep a close eye on problem products.
3. Production Reconciliation
The stores stock record is used to track the events on the factory floor from the issue of components to actual manufacture to the final production of finished material. Production reconciliation is needed for costing and similar purposes connected with good factory management.
4. Count Reconciliation
The physical stock of a product present in a stores or warehouse is counted. The count is then compared to the computer record with the objective of correcting it if a discrepancy exists. As referred to above and in due course, out-of-order and "premature" transactions must be taken into account.
IT Footnote: Individual transactions should never be permanently deleted from the transaction trail file. If, for any reason, a transaction must be deleted for IT reasons, a copy of the deleted record should be retained as a 'dummy'. If genuine transactions are allowed to be deleted, the way is opened up for fraud - ie stock may be stolen and all record of its existence removed (except in the invoice accountancy file, of course).
Return to Agenda Return to top of Agenda Go to Top of Course Go to End Return To Glossary
4.2 Count Variance Investigation and the Control Group
As stated above, the transaction file should be accessible on-line. The investigator should be able to make on-line queies of the data so as to focus on particular transactions or transaction groups of interest. Central to this facility is the ability to apply so-called screen filters, to screen out data of no interest (eg transactions earlier than Date X) or focus on subsets only of the data (eg raw material receipts only). Control groups are very small groups of` products (six or ten products at most) which are tracked for a couple of months by repeated checking in order to determine the nature and cause of any errors which may affect them. An audit trail is produced for them automatically. Any deviation between the physical count and the inventory record is picked up at once and its cause exhaustively investigated via the audit trail. Use of a Control Group is especially recommended in the first months of any new system to help check and improve its operation and check the validity of the transactions. See also the cycle counting load of a Control Group.
Return to Agenda Return to top of Agenda Go to Top of Course Go to End Return To Glossary
4.3 Work-in-Progress (WIP)
WIP is a species of inventory account and can be defined as the summarised value of in progress materials (US = "in process" materials). In progress materials are those undergoing physical transformation. As such, they no longer exist in the inventory system as specific SKUs* or specific parts, with part numbers. (*A stock keeping unit, or SKU, is a particular item, or product, perhaps qualified by its location. The term is used widely in logistics, where it is likely to be important to distinguish whether Product X is in the depot in the North (SKU X/N) or in the depot in the South (SKU X/S)) To distinguish SKUs/parts in a stock account from material in the WIP account, a formal definition of WIP is employed that relates to the material's 'transactional status' and to its product structure (ie BOM). 'Transactional status' means a status that stems from the act of issuing a works order (WO) - ie a manufacturing order (MO).
The convention is thus adopted that at the moment that a works order is issued, the material concerned and the necessary components needed to make it (derived from the bill of materials) assume the status of WIP. When the WO is closed, the materials actually made convert back to the status of stock.
For example, suppose one were to pick components according to a BOM and deliver them to the production area. There they are stocked items with part numbers. Suppose, next, one was to issue these materials to a particular manufacturing order. They are now WIP and are no longer stocked items. The vital action that converted them to WIP was not their delivery to production or any other physical action performed on them. It was their issue through a transaction to the specific Works Order.
A "WIP account" has a financial value that incorporates the value of the manufacture taking place - that is, it includes more than merely the financial value of the material involved. For example, it includes the value of labour reported against the production order and the value of overheads assigned through the costing system. When the WO is closed, the finished goods and scrap resulting move out of WIP status and become stock. Their value is consequently deducted from the WIP account. What remains is the residual value of WIP. That is:
The Residual Financial Value of WIP = (value of the starting materials + labour + overheads) - (costed value of finished goods + scrap).
Unfortunately, because of discrepancies and adjustments due to costing adjustments, scrap, inconsistencies in production rates and quantities, labour reporting errors and so forth, there are usually inconsistencies in the maintenance of WIP accounts.
As stated, the key to tracking WIP is clearly the Works Order, or Manufacturing Order. When a WO is released, the BOM is referenced relating to the product to be manufactured and the components are then allocated from component inventory. (Such allocations are specific to the WO - once created they remain attached to it unless manually changed or until it is closed. That is, WIP = WO + issued components.) There is then a need to monitor variances from standard component usages through a "variance by production order" standard report. The report lists the WIP inputs and outputs and the calculated differences. Usually, for short run production, the report is examined after production is complete, so that errors and write offs can be identified.
WIP cannot be "counted" or cycle counted in the same way that stock relating to a stocking account is counted. For one thing, from a system viewpoint, WIP does not exist as discrete units of stock. Instead, what must be done (if anything) is simply to verify that a specific Works Order exists and that all the components issued to the WO are, in fact, with the order.
To accomplish this, start with the WO listing and check. Then go onto the factory floor and compare the listing and the parts. Check that the WO is at the correct cost centre, that the quantities are correct and that the components issued are correct.
Return to Agenda Return to top of Agenda Go to Top of Course Go to End Return To Glossary
4.3a Work-in-Progress and a Transaction Trail
To see how a transaction trail helps us follow stock in the stores, stock on the shop floor and work-in-progress, consider the manufacture of a product 'Flexible Disks' (FD), made from 'Disk Hinges' (DH), as follows:
(a) The initial position is that there are 227 Disk Hinges in stock, in the stores;
(b) There is a stores issue of 200 Disk Hinges to the Shop. (27 DH as stock in the store, 200 DH stock in shop);
(c) Release of Works Order to make 180 FD (now 20 DH stock in shop, 180 FD as WIP in shop);
(d) Production Report of 180 FD (still 20 DH stock in shop, 180 FD as WIP in shop);
(e) Close out the Works Order (still 20 DH in shop, but 180 FD as stock in shop, zero WIP);
(f) Transfer of 17 Disk Hinges from Shop to Stores (44 DH as stock in stores, 3 DH stock in shop, 180 FD in shop);
(g) Transfer all FDs from Shop to Stores (44 DH stock in stores, 3 DH stock in shop, 180 FD stock in stores).
Return to Agenda Return to top of Agenda Go to Top of Course Go to End Return To Glossary
4.3b Backflushing
Backflushing
Pre-Production Relief
When a manufacturing order is created, the basic information associated with it and with the BOM is used to create the "materials list". The materials list is completed by multiplying the quantities involved, by the "quantities per" (as stated in the bill of materials for each component in the bill). The materials are then picked one by one and moved to the production area. Subsequently, all materials are issued to the MO (manufacturing order).
Post Production Relief (Backflushing)
With post production backflushing, materials are not issued via a transaction until production has been reported against an operation. That is, once production is complete, the operator posts production against the order (MO number, quantities, scrap etc). The materials used are then brought up and issued in one transaction. Also see above for information on backflushing. The four advantages of backflushing are:
1. Where scrap is common, production may continue along until enough good material has been made;
2. In point of use materials, it simplifies the issuing process and makes counting easier;
3. In very long production runs, where pre-production issuing of components takes materials out of stock weeks before their use, and transfers them to WIP, it gets things out of WIP faster.
4. Where bulk materials are involved, such that exact quantities cannot be picked beforehand.
Return to Agenda Return to top of Agenda Go to Top of Course Go to End Return To Glossary
4.3c Cycle Counting of WIP
It is clear from the description of WIP above, that accounting for it in financial terms as an asset of the company presents special problems. Yet the value of work in progress (US = work in process) can be immense, and cannot be ignored in either cycle counting or in an annual stock check. The key to its value is clearly the Works Order, since, as stated, by convention and often literally, it is the release of the WO which creates WIP. In general, items in stock on the shop floor are valued at their 'production cost', a cost accounting term, since work on them has been completed up to that point in the bill of materials chain that they have now reached. Work in progress itself, however, is at an indeterminate stage in the bill of materials. It is therefore necessary to estimate its value based on: (1) the starting value of units, at their production costs, before the commencement of processing; (2) an estimate of the amount of cost so far expended - for practical purposes, the accountant may choose to assume that 50% of manufacturing is in fact complete; and (3) their completed value, again based at production cost. Thus if Sub-Component A (£5 production cost) is being converted to Component B (£8 production cost - ie an increase of £3 in production cost), and if there are 100 units of A in progress on the shop floor, their value is deemed to be £650 (100 × £5 + 50% × 100 × £3).
Instructions on stock valuation of work in progress were issued by the Accounting Standards Committee in a 'Statement on Standard Accounting Practice on Stocks and Work in Progress' in 1975, revised in 1988, known as "SSAP9". SSAP9 stipulates that companies should value material at the lower of cost or net realisable value (NRV), where NRV is equal to the market price/selling price less all costs 'to completion' and less those to be incurred in selling, marketing and distribution. The standard further states that in comparing cost and NPV, comparison should be based on individual items, unless individual comparison is impractical. (For example, although stock at a distribution warehouse or retail outlet should normally be valued at the price originally paid to the supplier for it, or at its disposal value, if this is difficult to do because a large range of goods is sold and items in the range are constantly changing, such stock may be valued merely at the selling price less the gross margin normally added for profit.)
In summary, the assessment of WIP in cycle counting or in an annual stockcount is conducted separately from the counting of stock, and by its nature involves costing considerations and such factors as Works Order transactions, established scrap or yield values and other data familiar to the cost accountant. It must clearly be carried out under the direction of the company's cost accountant or chief accountant. Note finally that since the decisions made in this matter affect company accounts and profits, HMRevenue & Customs also have an interest in the matter.
Return to Agenda Return to top of Agenda Go to Top of Course Go to End Return To Glossary
4.4 Reconciliation of a Stock Count
4.4a Late Transactions
In discussing the somewhat complex subject of the reconciliation of a physical stock count with the stock record of the product counted, cognizance must be taken of the relative timing of the receipt of a transaction into the stock recording system and the timing of the conduct of the physical count. In doing so, a distinction must first be made between late transactions and premature transactions. The circumstances of a late transaction are illustrated here, as follows. The circumstances of a premature transaction are illustrated in 4.4b.
Suppose that we follow a number of events relating to a part P from time 3.00pm to time 3.20pm as follows: The physical stock quantities involved, and the changes in the physical stock quantity levels following a receipt and an issue, are as follows.
Starting stock quantity: 1000 units;
1st 'Event': 3.05pm, Receipt from Supplier 500 units (Stock quantity now 1500 units);
2nd 'Event': 3.10pm Issue to the Shop 300 units (Stock quantity now 1200 units);
3rd 'Event': 3.15pm Carrying out of cycle count 1200 units (Stock quantity remains at 1200 units).
Suppose we now follow the progress of the above events from the viewpoint of the Stock Record, not the physical stock, and the Data Transactions raised to reflect the activities. Thus:
Starting stock record quantity: 1000 units;
1st Transaction: 3.06pm Receipt from Supplier Transaction of 500 units (stock record now 1500 units);
2nd Transaction: 3.17pm Submission of Cycle Count reading of 1200 units (stock record now 1200 units);
3rd Transaction: 3.20pm Issue to Shop Transaction of 300 units (stock record now 900 units).
Because the Issue to the Shop transaction had not been submitted at 3.17pm, it appears from the cycle count at that time that the record is in error with a 300 unit excess. The correction is made to the record, but then the late transaction turns up at 3.20 pm to compound the error.
Return to Agenda Return to top of Agenda Go to Top of Course Go to End Return To Glossary
4.4b Premature Transactions
An alternative scenario is presented relating to Part P, this time from 3.00pm to 3.30pm. Again, we start by considering the physical stock quantity and the physical quantities involved in three 'events' affecting the stock. Thus:
Starting stock record quantity: 1000 units;
1st 'Event': 3.05pm Issue to Shop 400 units (stock quantity now 600 units);
2nd 'Event': 3.20pm Carrying out of Cycle Count of 600 units (stock quantity remains at 600 units);
3rd 'Event': 3.25pm Receipt from Supplier of 700 units (stock quantity 1300 units).
Again, we now follow the progress of the above events from the viewpoint of the Stock Record, not the physical stock, and the Data Transactions raised to reflect the activities. Thus:
Starting stock record quantity: 1000 units;
1st Transaction: 3.07pm Issue to the Shop Transaction of 400 units (stock record now 600 units);
2nd Transaction: 3.17pm Receipt from Supplier Transaction 700 units (stock record now 1300 units);
3rd Transaction: 3.30pm Submission of Cycle Count of 600 units (stock record now 600 units).
Because of the earliness, or 'prematurity', of the Receipt from Supplier transaction at 3.17pm - ie its submission before the submission of the cycle count transaction - when a comparison is made at 3.30pm between the cycle count value of 600 units and the stock record at 3.30pm of 1300 units, it appears, wrongly, that the record was in error by an excess of 700 units and the record is then made wrong by 700 units.
Return to Agenda Return to top of Agenda Go to Top of Course Go to End Return To Glossary
4.4c Solution to the two Reconciliation Problems
In assessing the timing of a transaction in relation to a physical stock count, it does not seem possible to identify it as being either late' or 'premature' by the use of software unless further information pertaining to the sequence of physical events is added to the recorded data attached to each such transaction. (That is, if all transactions and stock counts also held the date/time at which the physical events actually occurred to which they related*, the problems in 4a and 4b above could be solved). * The problem could also be resolved if the chronological sequence of the transactions was recorded.
L a t e T r a n s a c t i o n s
Failing help from software, one common solution to the late transaction problem is to eliminate all such transactions by closing off the stock location of each part which is to be counted and reconciled at some point in time before the count takes place, so giving time for all transactions remaining in the system to be submitted and processed before count and reconciliation are made. The close-off duration required can clearly be reduced by the prompt submission of transactions by storeman and by the provision of adequate technology for conveying transaction data through the system. In encouraging the first, the role of transactions must be explained to storemen. (It is very possible that a storeman views a transaction merely as an historical record of what has happened, without appreciating its further role as a means of communicating data through the stock recording system.) As for the second, an on-line IT system is very obviously desirable, preferably one making use of modern technology. It is common practice to "close off" a stock location temporarily from picking and further receipt by suspending a coloured rope or ribbon across it.
A second solution to the late transaction problem which requires assistance from software is to "freeze" the record of the item to be counted. "Freeze" does not mean "freeze"(!) - it means copy. The count recorded by the cycle counter is then compared to the frozen/copied record and any required adjustment made to the copy figure accordingly. If any late physical transactions have impacted the stock in the meantime, and their accompanying transactions have changed the record, these will not affect the copy record and the cycle count adjustment will not therefore be affected. See One-Step Reconciliation below.
P r e m a t u r e T r a n s a c t i o n s ( F o u r S t e p s)
Fortunately, it is not necessary to deal with premature transactions in an analogous fashion to that used for late transactions by keeping the stock closed off until reconciliation following the count. A simple software solution is available and is widely implemented in proprietary systems. This is described as follows in four steps:
Step I.
The stock of each part to be counted is closed off as described above, to enable late transactions to be flushed through the system.
Step II.
Before the commencement of actual cycle counting, and when it is judged that all transactions from before close-off have been processed, an "indicator" is set by computer on the stock record of each part it is intended to count. Once the indicator has been set for a particular part, all future transactions, including the cycle count, relating to the part, are collected in a small, dedicated computer storage area, or 'scratchpad' area, to await the action of the reconciliation program in Step IV.
Step III.
The stock of each part involved is kept closed off only until the cycle count has taken place. Immediately the part has been counted, the location is released for stock picking and receipt in the normal way. Note, however, that the data transactions raised relating to the released part, as well as being processed through the system in the standard way, are collected in the part's computer scratchpad storage area referred to above.
Step IV.
At any convenient time later, the reconciliation program is run. The program accesses the dedicated computer transaction scatchpad storage area of each part, and simply calculates the net stock balance (ie cycle count plus receipts less issues). If the net balance is different from the stock record following this calculation, the record is changed to the net value. The indicator setting is removed and the scatchpad area for the product is deleted.
The four-step procedure above is illustrated as follows for a Product P. The first five lines in the sequence of steps below represent what actually occurs from 3.00pm to 3.25pm. The nine lines which follow below represent the action of the system from 3.00pm to 3.35pm following Steps I to IV above.
3.00pm: Starting stock quantity of P 1000 units;
3.05pm: Issue to Shop 400 units, physical stock of P = 600 units;
3.10pm: Close off Stock Location, physical stock of P = 600 units;
3.20pm: Carry out the Cycle Count, count of physical stock = 600 units;
3.21pm: Release of Stock Location, physical stock of P = 600 units;
3.25pm: Receipt of P from Supplier 700 units, physical stock of P = 1300 units.
In the nine lines which follow, SSA is an abreviation for Special Scratchpad Area (see Step II above).
3.00pm: Starting stock record of Product P 1000 units;
3.05pm: Issue to Shop transaction 400 units, Stock record of P = 600 units;
3.10pm: Close off Stock Location, Stock record of P = 600 units;
3.15pm: Set the Record Indicator (see Step II above), Stock record of P = 600 units, SSA = 0;
3.20pm: Carry out the Cycle Count, Stock record of P = 600 units, SSA = 0 units;
3.21pm: Release the Stock Location, Stock record of P = 600 units, SSA = 0 units;
3.27pm: Receipt from Supplier transaction 700 units, Stock record of P 1300, SSA = 700;
3.30pm: Submission of Cycle Count trans 600 units, Stock record of P 1300, SSA = 1300*;
3.35pm: Run the Reconciliation Program, Stock record overwritten by SSA value (ie 1300).
* The scratch storage area now contains the receipt transaction (+700) and the cycle count figure of +600, total 1300 units. The sratch pad total is arrived at by adding up all the transactions (plusses and minuses) in the area, and, if the total is different from the stock record, replaces the stock record when the reconciliation program is run.
Common Implementations of Cycle Count Reconciliation Software
A number of software packages, especially those relating to cycle counting as part of MRP systems, offer what might be termed "semi-automatic" data reconciliation - basically, refinements of the Four Step Method above. Often, the facility comes with two options for the user: (1) One-Step, and (2) Two-Step.
One-step reconciliation
One-step reconciliation assumes that the user can guarantee that no late transactions will be submitted after the cycle count has taken place. Consequently, when it is known that all transactions have been processed, the computer is notified that a product is to be counted. A "switch" is then set on the record, and all transactions are read into a special area of disk equivalent to the scratch storage area referred to in Step II above. When the reconciliation program has been run, the switch is turned off. The methodology is virtually identical to 'freezing' the record referred to above.
Two-step reconciliation
With two-step reconciliation, it is assumed that there is no guarantee that late transactions can be cut off and prevented. The method proceeds as described above, until the point where the reconcilation program is run. At that point, the systems displays on the VDU all transactions processed since the switch was set, allowing the user to examine each of them on the screen for lateness. (That is, the screen displays the contents of the scratchpad area.) If a late transaction is discovered, it is deleted on the screen by the user. After all deletions, the reconciliation program is run as normal.
Return to Agenda Return to top of Agenda Go to Top of Course Go to End Return To Glossary
5.1 100% Physical Inventories (The Annual Stocktake)
A considerable further advantage of cycle counting besides those set out in due course below is that it should enable the elimination of the annual stock count. (And we should remember that the 'annual' stock count in many companies takes place several times a year. If the company issues quarterly accounts, it will take place four times a year.) However, if the accuracy of the records is still poor and the chief accountant is especially cautious, it may be decided to continue it. (In most companies, the chief accountant will have the final word on this matter, so that if anyone is to be persuaded to give up the annual stocktake and adopt cycle counting, it is he/she. One excellent way of so persuading him/her would be to charge the cost of the annual stocktake to his/her budget.) If a 100% physical inventory is indeed to be held, the first thing very definitely to say is that the counts obtained should not be allowed to update the main stock quantity fields on the stock records. Overwriting the stock records by annual stock count figures is a major cause of errors in the records. Instead, the annual stock record figure should be recorded in a special, reserved data field on each item's record. The reason is the very high degree of inaccuracy of the counts obtained in a full physical audit. We should remember that the accountant is mainly concerned with the total value of stock, not the "plusses and minusses" of individual records. Furthermore, the criterion by which the 100% audit should be judged cannot in fact be applied - ie whether the counts were accurate. Instead, the only criterion actually applied is whether the exercise finished on time. It always does! If the dictum above is followed and the only purpose of the 100% physical inventory is to report the total stock value, it is probably not worth going to great efforts to improve the procedure. If, however, improvement is required, the main point to make is that the preparation for the count should be thorough. Many days will be needed to prepare paperwork, prepare individual counting schedules and to print count sheets and forms. In addition, the stock and environment themselves must be made ready. For example, packages should be straightened and their markings checked for clarity. Location identifiers must also be checked for clarity. Equipment must be made ready, from clipboards to wire cutters to calculators. All preparation must be very seriously made, but it must be remembered that the greatest problem in the conduct of a full stocktake is control of the transactions as they are raised during the process itself, and in particular the reconciliation of variances. Six further points should also be taken account of as follows:
1. Training and selection of Counters.
Counting requires staff who are familiar with the materials being counted and with such matters as stock locations. To be effective, therefore, temporary counters should be (well paid) volunteers from the shop floor, absolutely not juniors drafted in from some hapless commercial department such as HR.
2. Closing off Stock:
Late transactions in reconciliation are often not successfully dealt with in 100% audits. To be sure of a clean break, therefore, , the stock must be closed off well beforehand, and sufficient time given after the close off to submit and process all outstanding transactions up to that point.
3. Variances and Unidentified Stock.
Time must be allowed to deal with these problems, and expertise provided to deal with them on an individual basis. 'Expertise' means the presence of the stores supervisor and his senior staff.
4. Transaction Processing.
If transaction numbers are clearly very large, an on-line reporting system is needed, preferably involving Radio Data Terminals or somesuch. Note that "default" transactions are not recommended in a 100% audit. (A default transaction means allowing the counter merely to put a tick by a count printout indicating correctness.)
5. Safety.
Instructions must be issued and explained on safe practice, and the audit supervisor must rigorously enforce the rules. The stores/warehouse is an unfamiliar environment especially to commercial staff.
6. Valuation of the Stock counted - See Section 5.14 below.
5.2 Definition of Cycle Counting
From time to time and for reasons to be discussed, including any perceived need to conduct an annual stocktake for financial reasons, as described above, the recorded quantities of all stocked items (ie the stock records) in a stores or warehouse are compared to the actual, physical amounts of stock present. If the counting and comparison are conducted as a single continuous exercise over as short a period as practical, the process is referred to as a full physical inventory, or 100% stocktake as described above in Section 5.1. If instead, the considerable tasks of counting and comparing all products are performed in many short, distinct stages over an extended period, perhaps of many weeks or months, the procedure is referred to as cycle counting. The 'extended period' chosen is the cycle, so called because when one cycle is completed, a second one is started. For this reason, cycle counting is also referred to as perpetual inventory checking (or "PI"). For example, suppose it was decided to count 1000 items over a 10 week 'cycle'. Then, in the first cycle, 100 items might be chosen to count per week - that is, 20 items chosen per day. At the end of five weeks, 500 items will have been counted. At the end of ten weeks (the cycle length) all 1000 items will have been counted. If cycle counting is accepted as a continuous activity, we then immediately start all over again on a second counting cycle. Then a third. And so on, perpetually.
5.3 Two Purposes of Cycle Counting
5.3a Identifying Errors
The first purpose of cycle counting is to find records which are in error and then to determine, if possible, why they have gone wrong. Depending on the nature of the error, one of two courses of action must be followed.
(a) 'Special Causes of Error'
Special causes of error in record accuracy (see Section 1.5 above) arise from somewhat 'out-of-the-way' incidents. That is, they stem from occurrences outside the routine of normal working practices or stem from the actions of people who are not part of the normal stores/warehouse team. Two extreme examples of special causes of error are the unauthorised and unrecorded removal of stock from the stores by a shop floor operator, and the theft of stock. The significance of special causes of error is that, once they have been identified, ad hoc procedures can relatively easily be put in place to stop them ever recurring.
(b) 'Common Causes of Error'
Common causes of error are the causes of those errors arising in the ordinary course of the day-to-day operation of the stock recording system. They arise because: (1) the people engaged in the system do not - cannot - work perfectly consistently all of the time, and sometimes make mistakes; (2) equipment does not function perfectly all of the time, and is subject to small but random variation in operation; and (3) procedures are flawed and permit the intrusion of random variation - for example, a during a stores' busy period, there is a likelihood that a recently completed goods-in transaction may become lost. If nothing is done about the system, common causes of error will naturally continue to occur. But by identifying them as part of cycle counting, management can see where effort would best be expended to change the system. Although a residue of common causes of error will always remain, their dominance can be reduced, and with it the susceptibility of the records to the random variation they represent.
Return to Agenda Return to Top of Agenda Go to Top of Course Go to End Return To Glossary
5.3b The Correction of Errors
The second very obvious purpose of cycle counting is to verify the correctness of the records and to correct any which are wrong. This purpose, obvious though it may be, is too often downplayed by experts in this subject.
Return to Agenda Return to Top of Agenda Go to Top of Course Go to end Return To Glossary
5.4 The Conduct of Cycle Counting
5.4a Blind Counting versus Verification Counting
A blind count is one in which the cycle counter has no prior indication as to the number or quantity physically present of the product that he is counting. One reason the manager might demand a blind count is to prevent the counter cheating by conducting a skimpy count - or no count at all - and then merely marking the record as "correct". (The counter might be inclined to cheat if counting is difficult.) A second reason for blind counting might be to avoid the honest mistake whereby the counter "sees what is supposed to be there" - the record might say 12 cases and he "sees" 12 cases, rather than the 11 that are actually present. The principal problem with blind counting is that it takes much longer to conduct than verification counting - ie counting in which the counter has knowledge of the number or quantity on the stock record. For one part, this is due to the frequency of re-counting necessary (perhaps because of a non-match with the stock record or perhaps due to miscounting, rather than a record error). For another, blind counting requires more care and time of itself.
With verification counting, counting to an approximately expected number seems to be capable of being conducted more methodically and slickly. In verification counting, a quick count may first be made. Then : (1) if the quick count reasonably closely matches the stock record, the counter merely deems the stock record to be correct and moves on to the next product; or (2) if the figures do not agree, the counter will look closely for reasons for disagreement - for example, for units that have slipped to the back of the storage bin or ones that have been located in the next storage location. Such investigation should always takes place when the quick verification does not agree with the record. The quick verification count itself could always be in error, of course, and if so the error will be revealed when this investigation takes place. Note that it is most unlikely that the quick count error will be equal to a stock record error (ie suppose there were really 48 units present, and the quick count came to 42 - what is the chance that the stock record is also showing an erroneous value of 42 at this particular time?) To catch anyone cheating with verification counting, the auditor can deliberately remove units of stock from a product's location to see whether the counter enters the stock record quantity or the known reduced number.
Return to Agenda Return to top of Agenda Go to Top of Course Go to end Return To Glossary
5.4b Exception Reporting
The next question with Verification Counting is whether, in the case of a "good" count, it is necessary to record on the system transaction the number counted. Instead, if the count has agreed with the record, why not simply tick a box saying "OK"? In order to conduct exception reporting, it will, naturally, be necessary to have support from software so that only a simplified transaction need be completed. Nevertheless, if we are using the admittedly complex automatic reconciliation procedure described above, we can see that the prompt input of these simplified transaction will allow the immediate release of the stock to further picking and putaway.
Return to Agenda Return to top of Agenda Go to Top of Course Go to end Return To Glossary
5.4c The Storage of Variances in a Special Data Field
If the automatic stock adjustment process described in Section 4.4 is followed, stock adjustments resulting from a variance between the stock record and a count are handled by the IT system and need not be made manually by the cycle counter. However, they must be written to a report file for financial auditing purposes. In addition, differences should also be recorded in a special "variance field" on the stock record. The reason is that oftentimes stock quantities that are missing - ie the stock variances themselves - are found later. For example, missing stock will be revealed by location audits described at the end of this Chapter of the on-line Course. When stock is found, it is helpful to be able to refer to the last stock record variance or adjustment, to see whether it equals the amount just found.
Return to Agenda Return to top of Agenda Go to Top of Course Go to end Return To Glossary
5.5 Cycle Lengths
It is dubious whether the accountant who has been persuaded to discontinue the annual 100% physical inventory would agree to any cycle of a length greater than 12 months. It is also dubious whether any cycle shorter than one month could be justified. Between these limits, the duration of the cycle over which all of the products are to be counted in a stores/warehouse or within a group of products within a stores/warehouse is purely a matter of management judgment. Judgment will be influenced by many factors, including the current state of accuracy, the criticality of accuracy to on-going operations and the number of products in the stores that it is required to count. Above all, it will be influenced by the number of staff required to perform the counting task over the duration decided on and the money it is thought appropriate to allocate to the activity from the departmental budget. To calculate the length of the cycle and the counting load, a two step procedure can be followed:
Step 1: First find the number of counts per day from the effort to be expended.
Before the company can consider either cycle length or the effort necessary to sustain it, it is necessary first to estimate the number of products which can be counted in a given time. This is clearly particular to a given company, and will depend on such factors as the nature of the products, the type and size of the stores or warehouse, the procedures for reporting counts and dealing with reconciliation etc..
Suppose, however, that a cycle counter is able to count 25 products per day, this being a generally agreed target figure, and one suggested by Roger Brooks and Larry Wilson in their book Inventory Records Accuracy, being based on an average taken from surveys on their Stock Records Accuracy training course held in the US. (The figure suggested by APICS is 30 products per day.) This being so, then the first parameter is the number of counts per day possible, given by Equation 1 below.
Equation 1: Counts per day = no. of counters x counting rate per day
This is the critical Equation, and is used to obtain the number of counts per day that will be made employing the cycle counting effort decided on:
Example 1 Suppose that the counter can count 25 parts per day, and there a 2 cycle counters: Counts/day = 2.0 × 25 = 50 counts per day
Example 2 There is one counter for 3 days/week and he can count 50 items/day ... Counts /day = 0.6 × 50 = 30 counts per day
Step 2: Find the cycle length in days.
To obtain the length of the cycle in days, apply Equation 2 ...
Equation 2: Cycle Length = no. of products / counts per day
Example 1: What cycle length would apply to a group of 1000 products if a cycle counter was employed 2 days / week and could count 50 products / day?
Step 1: Counts per day = 2/5 × 50 = 20 counts / day; Step 2 Length of cycle = 1000 / 20 = 50 days (10 weeks)
Example 2.: What would be the cycle length if 2000 products in a stores were to be counted all over the same cycle by one cycle counter, if the cycle counter could count 20 products per day?
From Equation 1, No of counts per day = 1 x 20 = 20. From Equation 2, cycle length = 2000 / 20 = 100 days, or 20 weeks.
Return to Agenda Return to top of Agenda Go to Top of Course Go to end Return To Glossary
5.6 The Classic Approach to Cycle Counting
5.6a Introduction
It would be impractical and of no benefit to assign a separate counting cycle to every product individually. However, rather than impose the same cycle on every product, products in the stores or warehouse are divided typically into three groups termed A, B and C. Those falling into Group A are all counted over one cycle, products in Group B are counted over a second somewhat longer cycle, and those in Group C over a third cycle, being the longest of the three.
Return to Agenda Return to top of Agenda Go to Top of Course Go to end Return To Glossary
5.6b ABC and Annual Value
A very common way of dividing products in the stores/warehouse into the three groups referred to above is by ABC Analysis, or Pareto Analysis, based on each product's annual value. Annual value in units of £ is defined as:
unit cost × annual usage
In exactly the same way as illustrated below in section 5.3, it will be found that if all products are placed in a list ordered in descending order of each product's annual value, then:
(1) the top 20% of the products in the list will account for about 80% of the total annual value of all of the products in the list;
(2) the next 20% of the products in the list will account for about 15% of the total; and
(3) the bottom 60% in the list will account for only about 5% of the total annual value.
Obviously, this scheme proposes that items in the A group are counted more often than items in the B group and that items in Group B are counted more often than Cs. The application of ABC analysis to a group of ten products is illustrated by the two lists in the Tables below. In the first, the identity of each product in the group is shown in column (1) ... P1, P2 ... P10. In column (2) we have the quantity of each product used annually. ('Used' may imply bought in, issued for production, or produced as stock.) Column (3): the unit cost V of each product. Column (4) is the annual value, ie quantity × value (that is, annual quantity x V). Note that the total annual value of all 10 products has been calculated (£4875). The figures in column 4 enable the ten products to be ranked in descending order of annual value. The ranking, or order, of each one is given in column (5).
Table 1
(1) Item (2) Annual Quantity (3) Unit Value (4) Annual Value (5) Order or Rank
P1 .......................10..........................£2.......................£20..............................10
P2 .......................40 .........................£1.......................£40 ..............................8
P3 ........................15 ........................£10 ...................£150..............................6
P4 ........................50 ........................£4 .....................£200 .............................5
P5 .......................115 .......................£12 ..................£1380 ............................2
P6 ....................... 5 ...........................£5 .....................£25 ..............................9
P7 ....................... 30 ........................ £15 ..................£450 ............................ 3
P8 ........................275 ...................... £8 .................. £2200 ............................1
P9 ....................... 15 ........................ £4 ...................... £60 ........................... 7
P10 .......................50 ....................... £7 ..................... £350 ...........................4
Total ...........................................................£4875.....................
To make things clearer, the ten ranked products in Table 1 have been rearranged in order of their ranks, and transferred in order to Table 2 below. Note that 'Cumulative % Value' means the value of this Product expressed as a percentage of the total (ie a percentage of £4875), but as a cumulative figure. That is, the entry in the Table for, say, P5 (second in the table) is added to the entry for P8 (first), and the total of these two (£1380 + £2200) expressed as a percentage of £4875 (ie £3580 / £4875 = 73.4%). Similarly, the entry in the Table for P7 (third) is added to the entries in the Table for P8 (first) plus the P7 (second), so we have a total of £4030, which is 82.7% of £4875. Note that the bottom product in the table (P1) represents only 0.4% of the total annual value and might be relegated to a fourth class, Class D.
Table 2
(1) Item (2) Order or Rank (3) Annual Value (4) Cumulative % Value
P8 ........................1..........................£2200.......................45.1
P5 ........................2 .........................£1380.......................73.4
P7 ........................3 ...........................£450 .......................82.7
P10 .......................4 ..........................£350 ...................... 89.9
P4 .........................5 ..........................£200 .......................93.9
P3 ........................ 6 ...........................£150 ......................97.0
P9 ........................ 7 ............................ £60 ......................98.2
P2 .........................8 .............................£40 ..................... 99.1
P6 ........................ 9 ............................ £25 ..................... 99.6
P1 .......................10 ............................ £20 ..................... 100
Total ......................................£4875...............100
One argument against ABC Analysis based on annual value in cycle counting is that there is no supposing that the stock records of products in Group A are more likely to error than the records of products in Groups B or C. A second argument against it is that products in Group A are typically stocked in larger quantities than products in Groups B and C, and if so will take commensurately longer to count. (The average stock quantity of any product is (R / 2 + S) units, where R = the replenishment lot size and S = the safety stock. Since Group A products have larger annual usages, they will have larger replenishment lot sizes.)
Return to Agenda Return to top of Agenda Go to Top of Course Go to end Return To Glossary
5.6c ABC and Transaction Frequency
To fulfil the two purposes of cycle counting, it is necessary to count products with records that are in error. It is obviously therefore desirable to split the products into A, B and C groups by a procedure such that products in Group A are likely to have the greatest proportion of stock record errors. Logically, the only straightforward way of dividing the products into counting groups having a higher proportion of errors is one which analyses products according to the number of transactions that have been applied against their records. Every transaction theoretically represents an opportunity for the record to be made wrong, so that the records of products which have had the greatest number of transactions applied against them are most likely to be in error. The Pareto effect holds true for the analysis of stock records based on the number of transactions that have been applied in exactly the same way as it does based on annual value described in Section 5.6b above. That is, 20% of the products will account for about 80% of all transactions raised/processed (Group A), the next 20% will account for about 15% of all transactions (Group B) and the bulk of products at the bottom (60%) will account for a mere 5% of all transactions. An example of ABC Analysis, or Pareto Analysis, based on the number of transactions is given in the two further tables below, Tables 3 & 4, again for ten products P1, P2, P3 .... P9, P10. In Table 3, the columns have the following meanings: (1) Item (2) Annual Number of Transactions (3) Order or Rank. (Total number of transactions is 8826).
Table 3
P1 ........................28.........................10
P2 .......................100 .........................8
P3 .......................300 .........................6
P4 ........................420 ........................5
P5 .......................2200 ........................2
P6 ......................... 48 .........................9
P7 .......................1600 ........................3
P8 .......................3200 ....................... 1
P9 ........................ 110........................ 7
P10 .......................820........................ 4
Total ........... 8826.....................
Again, to illustrate the next step in the Pareto process, the ten products in Table 3 have been transferred to Table 4 below, and arranged in descending order of annual number of transactions. That is, Product P8 (1st) is at the top, Product P5 (2nd) is second, and so on.
Table 4
(1) Item (2) Order or Rank (3) Annual Number of Transactions (4) Cumulative % (of Transactions)
P8 ..................1................3200...........................36.3
P5 ..................2 ................2200 .........................61.2
P7 ..................3.................1600 .........................79.3
P10 ................4...................820 ..........................88.6
P4 ..................5...................420...........................93.4
P3 ..................6.................. 300 ...........................96.8
P9 ..................7...................110 ...........................98.0
P2 ..................8...................100 ........................... 99.1
P6 ..................9.................... 48.............................99.7
P1 ..................10 ..................28.......................... 100.0
Total ........................ 8826....................100
Again, note that although the third column of this Table shows the individual part's annual number of transactions, the fourth column gives the cumulative annual number, as a percentage of the total. For example, the entry for P10 , say, reads 820 transactions, so that the cumulative total is (3200 + 2200 + 1600 + 820) = 7820 and the cumulative percentage is (7820 / 8826 x 100% = 88.6%). Note in this example that the split is quite a long way from the "traditional" 80/20. Oftentimes in ABC analysis the split is more likely to be 70/20; in this case it is only 60/20. As before with annual value, when Pareto analysis is performed on the stores' or warehouse's products on a full scale, it will generally be found that the products in the top 20% of all of the products account for about 70% of the total number of all transactions applied to the whole group. These are the "A" products. The next 20% of products further down the list are likely to account for a further 20% of the total number. These are the "B" products. The remaining 60% of the products account for only 10% of the total number of transactions and are the class "C" products.
As we shall see below, there is absolutely no requirement to split the products in the proportions 20%, 20%, 60% given above: the split in fact should be made in conjunction with the calculation of the counting load, discussed in Section 5.7 below.
Finally note that it is common to plot the results of Pareto analysis on a graph. The vertical axis of the graph represents the percentage of total annual number of transactions (ie 0% to 100%). The horizontal axis represents the percentage of the total number of products, typically plotted by a computer graph plotter. The cumulative annual value of each successive 1% of the products, in descending order of annual value, is then plotted. The familiar "Pareto Curve" shows a rapid initial rise in slope followed by a flattening off.
Return to Agenda Return to top of Agenda Go to Top of Course Go to end Return To Glossary
5.6d Alternative ABC Methods
(1) By Unit Financial Value of the Individual Products
The products are placed in (say) three categories A, B and C according to the unit value of the individual items. The A items are (say) defined as the top 10% in value and are counted on the shortest cycle - ie most often. The B's (perhaps the next 20%+ in value) are counted on a medium cycle, and the C's on the longest cycle. This method is probably inferior to annual value.
(2) By Location
When products are held in a very large warehouses or in several separate stores or warehouses, counting and travelling clearly present problems, and especially so when such sites are remote from the organisation's main operations. In these cases, the cycle counter may prefer to count the different warehouse sections or different stores over separate, different cycles. Clearly the actual counting of the different cycles does not take place in parallel in these circumstances, unless there are 2 or more cycle counters.
(3) By Replenishment Leadtime
There is nothing whatsoever inherent in a product's replenishment leadtime that would suggest the record of its stock is either more or less liable to error. In that sense, the method of categorising products into groups based on (a) long, (b) medium and (c) short replenishment leadtimes appears irrational. However, if a stock record is in error, and the error is a stock deficiency (rather than an excess), stock replenishments to make up any deficiencies in products with short leadtimes are clearly not so serious as the replenishments of products with long leadtimes. Consequently, products with long replenishment leadtimes are made Class A products, and should be counted most frequently, and products with short replenishment leadtimes are made Class C and are counted least frequently. Two further points: (1) for finished products, Class A goods with long replenishment leadtimes will usually have zero safety stock associated with them because of the way safety stock is calculated, and (2) Class A goods will often be the most numerous.
Return to Agenda Return to top of Agenda Go to Top of Course Go to end Return To Glossary
5.7 Calculating the Count Load
The basic principle to be applied to determine the total cycle counting effort for three different groups of products, each with a different cycle, is the same as described above in Section 5.5. A problem, however, is the greater degree of arithmetical complexity involved. A useful way of performing the calculations is to set out the facts in a simple table, illustrated below by Table A. As will be described later, the table lends itself readily to being set up as a spreadsheet for ease of manipulation. The essence of the calculations in the Table is to work out the number of counts per day for the different cycles, and then simply to summarise these in a grand total. The illustration below involves a stores with 5000 products which have been categorised into A, B and C in the percentages shown. Initially, the company wishes to find the effort that will be necessary to count the products in the various categories over the cycles shown in Table A: ie the A's monthly, the B's quarterly and the C's annually. Note that the last column (Coumn (6) (Counts per Day) is obtained by dividing the Counts per Year (Column (5)) by 240, it being assumed that there are 240 counting days in a year. It can be seen that the proposed cycles for the A, B and C Class products require in total approximately 79 counts per day. If a cycle counter is capable of achieving 25 counts per day, the counting effort required is approximately 3 persons.
Table A (5000 products)
Category ...% of Prods .... Number of Prods ... Counts per Year/Item ... Tot Counts/Year .. Counts/Day
A .....................20% ...............1000 ................................12 ................................12,000 ...................50.0
B .....................20% ...............1000 .................................4 ...................................4,000 ...................16.7
C .....................60% ...............3000 .................................1 ...................................3,000 ...................12.5
Totals .............100% ..............5000 ...............................n/a.................................19,000 ...................79.2
A procedure to find cycles which give an acceptable counting effort, such as 25 counts per day, is both to vary the cycle lengths of the A, B and C counts, and to vary the percentage A, B and C subdivisions (ie to change the 20%, 20% and 60% split). If the split between A, B and C is amended to A: 4%; B: 8% and C: 88%, and the cycle lengths are changed to 3 months (A), 6 months (B) and 12 months (C), the results shown in Table B are obtained and the total effort needed is exactly 25 counts per day (ie "one person"). Note, incidentally, that actual execution of the A, B and C counts, when they are undertaken in the physical sense, should be in parallel. That is, if we were to go ahead with the scheme in Table B, each day three A products would be counted, three B products, and 18 C products.
Table B (5000 products)
Category ...% of Prods .... Number of Prods ... Counts per Year/Item ... Tot Counts/Year .. Counts/Day
A .....................4% ................. 200 ..................................4 .................................... 800 .......................3.33
B .....................8% ..................400 ..................................2 .....................................800 .......................3.33
C .....................88% ...............4400 .................................1 ...................................4400 .....................18.33
Totals .............100% ..............5000 ...............................n/a...................................6000 .....................25.00
While the stores manager may be satisfied with the overall pattern of cycle counting and the split between A, B and C chosen initially, he may also wish to explore alternative possibilities. That is, he may wish to increase or reduce the total counting effort, or make other changes such as to cycle lengths. To perform the calculations necessary in an easy way, it is strongly suggested the stores manager should set up a customised "cycle counting spreadsheet". The advantages of doing so are not simply that it will save time, and allow 'What If?' simulations. It will also allow ready changes to be made over the years as circumstances change. New circumstances include: squeezes on manpower and budgets, improvements or deteriorations in the stock records accuracy, changes in the stock profile or the size of the counting task ... Whenever necessary, the spreadsheet is reviewed and new, more appropriate cycles and counting effort worked out.
S P R E A D S H E E T S
In this context, the Cycle Counting Spreadsheet will consist of a grid or matrix of 20 rows and 10 columns on a computer screen, the 200 (20 x 10) intersections of the rows and columns being called 'cells'. There are two steps involved in making use of the spreadsheet:
1. SETTING UP THE SPREADSHEET: To "set up" the spreadsheet, formulae are entered in the cells to express the relationships described above relating to Tables A and B. Descriptive headings are also entered.
2. PUTTING IN THE NUMBERS Once the spreadsheet has been set up, it is "ready to go". The formulae and expressions referred to above disappear from the cells on the screen. The cells seem empty, but in fact the formulae are, as it were, `behind' the cells. Now, actual numbers can be entered into those cells. For example, if we had 5000 products, we would enter 5000 into the cell reserved for ''Number of Products'. We might also enter the value '4' into another Cell, corresponding to the percentage of A products. Immediately that was done, the value '200' would appear in another cell reserved for the 'Number of A Products', since the Spreadsheet program would multiply 5000 x 4% (5000 x 0.04). Clearly, the stores manager or cycle counter can enter alternative values for the split of items into A, B and C, and alternative value for the frequency of counts. As he does so, the spreadsheet will work out automatically the total counting load, as illustrated in Table A and B above.
FOOTNOTE
It should be remembered that if there is to be a `Control Group' as described in Session 4.2, which is counted every day for the reasons stated there, that this constitutes a group of its own. Furthermore, the counting demands per day of a control group can be considerable. Although allowance should be made for the speed with which very familiar products can be counted, the load of only (say) 5 products, counted every day, is 1200 counts in all ... 20% of the total effort of one cycle counter.
Return to Agenda Return to top of Agenda Go to Top of Course Go to end Return To Glossary
5.8 Low Stock Counting
If products are selected from three ABC groups randomly, then a major objection to this procedure for cycle counting is its poor productivity. That is, the amount of stock on-hand to be counted will be the "average amount" (ie R/2 + S, where R = replenishment quantity and S = safety stock). But there is no merit in counting for the sake of it. Low stock counting, described here, represents a complete alternative to ABC. Increases in cycle counting effectiveness which are obtained from low stock counting have been quoted by Roger Brooks and Larry Wilson (see Inventory Records Accuracy, book referred to above) from 25 items per day to an astonishing 300 items per day. Two other advantages of low stock counting are that (1) in counting fewer items, there will also probably be greater accuracy in the counts; and that (2) if, in fact, there is a discrepancy in the stock record, then counting the item when the record and the actual amount of stock are low, may enable the stock to be replenished before the next required issue of the material - since the next issue is likely to result in a stock-out. Note that although theoretically low stock counting could be applied to ABC, the software system required to support the two in parallel would be rather complex, or at least, rather messy, as might be its day to day control. It is suggested, therefore, that ABC ideas be abandoned. With low stock counting, the selection of items to count from day to day within the stores is made on the basis of the comparative stock holding of the products. In short, items are chosen with the least stock. After counting, the products counted are transferred to the "already counted" list. The next day, all the products not on the 'already counted' list are examined again, and again the ones with the lowest stock are selected for counting. (As the cycle proceeds, there is less and less choice, so that, at the end of the cycle, the products being counted will have an average amount of stock, not low stock.) One obvious question with low stock identification is how "low stock" is to determined. That is, we must decide not only how "lowness" is to be decided as it concerns a given product, but how "lowness" can be compared from one product to another. There are two solutions to the problem.
Return to Agenda Return to top of Agenda Go to Top of Course Go to end Return To Glossary
5.8a Computer Identification of Low Stock Items
The first solution to the problem depends only on knowing, for each product, the product's approximate maximum stockholding. A product's maximum stockholding might be found by an analysis by computer of the stock records each day over the course of two or three months. Alternatively, after discussion with others, one might simply plump for `the item's replenishment quantity + 10%'. The maximum stockholding should be held on the computer stock record as a distinct data field. The stockholding of a product - relatively high or relatively low - is then compared to the stockholding of another product by simply calculating and examining the fraction:
Current Stockholding / Maximum Stockholding
For example, consider Drawing Pins with a maximum stockholding of 100,000 units and current stock of 6,000, and Size 3 Engines with a maximum stock of 200 and current stock of 70. The two fractions are (1) for Drawing Pins, (6000 / 100000), or 0.06 (6%), and (2) for Engines (70 / 200), or 0.35 (35%). Drawing pins are therefore the lowest stock product. Note first, that when the percentages or fractions for the stock records as a whole are analysed, they will not show an even spread from zero to 100% (or 1). Because items are subject to large replenishments and withdrawals, then for each item there will be large gaps in the range. That is, the percentage for a product will be at, say, 95% one day and 25% the next, following an issue that day to the shop floor. This unevenness for one product will be reflected in all of them as a group. The second point is that sometimes for a product the percentage calculated will be greater than 100%, showing that the maximum stockholding on the computer record has been set too low. There is some merit in getting the maximum as correct as possible. Consequently, if percentages over 100% are detected, the maximum stockholding on the product's record should be amended accordingly. Finally note that a product's stock may in fact be physically low, but its record may indicate it is high. If so, the product will not be selected for counting in the normal course of events. However, as we will see below, it will eventually be counted.
The selection process by computer is now quite straightforward. The program calculates the percentages of all products in the 'still to be counted' list and selects the required number with the lowest percentages. (As stated previously, as one gets nearer to the end of the cycle, there will be fewer on the list and the choice will be correspondingly smaller. In the last count of the cycle there will be no choice and the products to be counted will be liable to have average stocks.)
Return to Agenda Return to top of Agenda Go to Top of Course Go to end Return To Glossary
5.8b Identification of Low Stock Items by Patrol
The cycle counter deals with a defined area of the stores each day, and is given a list of all parts and quantities in the area. He counts and records only those parts which he observes, literally, are easy to count, ie those seen to have very low stock or seen to consist of a very small number of distinct packages. (However, if he spots that the stock quantity of a part is different from the recorded quantity on his list by an order of magnitude, so that the record is obviously in error, that part is also included in his daily count.)
Return to Agenda Return to top of Agenda Go to Top of Course Go to end Return To Glossary
5.8c Ensuring All Items are Counted
In implementing patrol identification in 5.8b above, a careful note must be kept of all the areas dealt with from day to day to ensure complete coverage of the stores/warehouse in a reasonable time. An area may be visited several times in a month, at each visit different products being at low levels and having become candidates for easy low stock counting. The cycle counter submits transactions each day relating, of course, only to the low stock counts he has made (plus any parts counted because the records were suspected as being in error). The computer system records all parts counted and submitted over a period (say, six months), and a report is produced at the end of the period listing any parts in the stores which have been missed, usually because their stocks were never low but perhaps for some other reason. These parts are separately and specially audited. In this way, it can be ensured that all parts are counted, having expended the minimum overall effort.
Return to Agenda Return to top of Agenda Go to Top of Course Go to end Return To Glossary
5.9 Other Cycle Counting Product Selection Methods
As with low stock selection, other methods by which products might be selected for day to day cycle counting do not necessarily mean that ABC should be abandoned, but a warning is given that any scheme devised should be relatively easy to apply and manage. For this reason, the manager should think most carefully about 'opportunity selection' in (i), (ii) and (iii) below, even though they seem to be logical and uncontroversial.
(i) Prior to a Re-order of stock
This is an "opportunity count" used in conjunction with the materials planning system.
(ii) Prior to a Receipt of stock
An 'opportunity count' as above.
(iii) Prior to the issue of a production / works order.
An opportunity count as above.
(iv) Random Selection.
Items are chosen from one day to the next by random selection. To do so, the items are assigned ordinal numbers from 1, 2, 3 ..., and a random number generator is then used to choose the items to be selected for the day in question. The rationale behind random selection is that of the spot check audit - to defeat collusion by staff who might be attempting to fix the records. The method has little to recommend it except for this. ABC however is easily possible.
(v) Permanent Fixed Order
A permanent counting order is assigned to the items which is maintained from one cycle to the next.
(vi) Selection by Location
Even though the actual physical counting route is sorted into location order, this might still result in a very considerable distance having to be traversed if the stores or warehouse is an extensive one. Cycle counters might therefore choose the sensible and attractive method of selection based simply on the location of the products. That is, products are selected literally in their physical order of storage. ABC is possible with a little ingenuity.
In summary, two alternatives are clear: either the cycle counter must abandon ABC and rely on low stock, or he must get his IT colleague to write a special program that combines ABC with perhaps several of the opportunistic identification methods out lined above. Any such program will be convoluted but not necessarily complex.
Return to Agenda Return to top of Agenda Go to Top of Course Go to end Return To Glossary
5.10 Location Auditing
An obvious problem with cycle counting is that it does not find stock which has been placed in the wrong location entirely. The cycle counter looks in the place designated by his system to check that the correct stock is present. (However, if he does find other stock there which he sees should not be present, it is essential that he should raise a transaction to report it, so that the matter can be investigated.) He will not, of course, find lost or missing stock if he is not directed to its location. For this reason, it is strongly recommended that there should be a permanent ½-day week location audit programme. The procedure is first to designate groups of locations ('areas') which cover the whole stores/warehouse such that every location in a group or area can be visited and approximately checked in half a day. What is then done each week is to run a stock records report listing everything recorded as being present in the locations in a given group/area. One group/area of locations is checked each week. That is, in a particular week, the auditor responsible for this programme visits each location in the group or area designated for that week, and (1) first makes a note of what is stored there, and then (2) checks that everything that is stored in the location is on the computer generated list. Note carefully the sequence of these two events. The cycle counter is looking at the location and asking "Is this product on the list?". He is not first looking at the report and asking "Is this product (on the report) in the location?" (*). Note also that the auditor is not counting the product - location auditing is performed to find lost product, not count product which is not in fact lost. (* If the cycle counter sees, however, that there is a major and obvious discrepancy between the record and what is physically present, a follow-up count should be initiated.)
Return to Agenda Return to top of Agenda Go to Top of Course Go to end Return To Glossary
5.11 Cycle Counting Variable (Random) Locations
5.11a Conventional Cycle Counting
If the stock of a given product under the random location storage system is stored in more than one location, as usually it will be, it would be a mistake to count all of the product's stock, at all of its locations, in order to carry out the cycle count. Such a procedure would be relatively slow, since, despite a superior picking rate for the reasons explained below, each location would need to be visited both for actual counting and in the prior closing off of the various locations before the commencement of such counts. Instead, the stock of a given product at each location where it is stored should be regarded as a separate "SKU" (stock keeping unit), and cycle counting performed at the SKU level. That is, what is counted is not all the stock of the product, it is the stock of the product merely at the designated location. And what the count is compared to is not the complete company stock record quantity of the product, it is the quantity on the record corresponding to that location only.
One method of minimising the counting load in such a set up is to select, for counting, low stock SKUs as described earlier for a fixed location store. An alternative scheme, to minimise the amount of walking between SKUs and to make for a practical daily system, is to designate a particular, defined segment of the stores for SKU counting each day. Thus suppose that Rack R is chosen for counting on a particular day. The computer now prints out the locations, parts and stock quantities of all locations within Rack R:
A R 03 B1 ....... Part Q ........ 59 units
A R 03 B2 ........Part S ..........69 units
A R 04 C1 .........Part Z ......1080 units
Companies which count individual SKUs at individual locations, such as the three above, often seem to find it hard to resist the temptation of going on to check the entire stock of the product if a discrepancy is found. For example, a cycle counter counting Location A R 03 B1 above who finds a discrepancy there in the record quantity of Part Q of 59 units, feels obliged then to count all other locations where Part Q is stored. This clearly destroys the whole notion of SKU checking, although it must be admitted that one way of resolving the variance at Location A R 03 B1 might be to examine the stock records of Part Q at its other locations - see Section 5.11d. Note that although there are proportionately more SKUs in the Random Location store than there are products in the Fixed Location one, the amounts of stock in the two systems are clearly the same. Consequently, the amounts of stock at the SKU locations will be proportionately less than the amounts of product at the fixed locations. Consequently again, the counting rate of the SKUs will be proportionately higher than the counting rate of the Fixed Location products.
Return to Agenda Return to top of Agenda Go to Top of Course Go to end Return To Glossary
5.11b Zero Stock Identification
In order to put into effect this powerful method , a required procedural rule of the stores must be that when any location is emptied in the course of withdrawing stock, this fact must be reported to the inventory records system (in addition to the quantity actually withdrawn, of course). This report, in the form of a transaction or a special field on the withdrawal transaction, is required so that the record can be compared to what is now known to be a physical count of zero. As stated, to assist in this, there should be an additional field on the "Issue Transaction" which asks the question 'Was the location emptied by this withdrawal?'. It is at this point that the real cycle count / stock reconciliation takes place. If the location was indeed emptied by the withdrawal, the computer notes this "cycle count of zero" and the value of the stock record at the time (taking into account the withdrawal notified through the transaction itself, of course), and notifies the system controller that he should examine the stock record's transaction trail (at that location only) in order to make an adjustment as described in the following example.
EXAMPLE: Suppose Item A's stock record showed the following data:
Location 12P: ... SKU A-1 ....500 units ... most recent stock
Location 19R: ... SKU A - 2 ....485 units ... next oldest stock
Location 22Q: ... SKU A - 3 .... 65 units ... oldest stock
Suppose that it is required to issue 80 items of A. Based on the stock record, the picking document requires the storesman/warehouseman to pick as many of the 80 as possible from 22Q (oldest stock) and the remainder from 19R. There are three possibilities:
(1) The stock record at Location 22Q is correct. This will mean 65 is duly picked from Location 22Q, emptying it, and the balance of 15 is taken from Location 19R, so reducing the stock at Location 19R to 470 units. Since Location 22Q is emptied, this fact is reported as such to the system. Since exactly 65 units of A were picked, as shown on the record, the stock record of zero (65 - 65 = 0) now tallies with the 'count' of zero.
(2) There is more stock actually at 22Q than it says on the record (say, 73 units). The storesman, knowing he must pick 80 items, picks all 73 items from Location 22Q (ie as many as he can get, provided he doesn't get more than the 80), and the balance of the 7 units from the next location, Location 19R. The stock at 19R is now down to 478 units (485 - 7 = 478). Since he removes all the stock from Location 22Q, he again reports in that the location is empty. In other words, we now have the following: Location report for stock at 22Q: 0 units (ie empty): Calculation: Stock on record (before picking) = 65 units Stock on record (after transaction) = - 8 units (65 - 73). There is now a cycle count adjustment to stock record at 22Q of + 8 units (ie adjustment up, a gain)
(3) There is less stock actually at 22Q than it says on the record (say, 58 units) The storeman, knowing he must pick 80 items, picks all 58 items from Location 22Q, and the balance of the 22 units from the next location, Location 19R (reducing the stock at 19R down to 463 units ie (485 - 22 = 463). Since he removes all the stock from Location 22Q, he again reports in that the location is empty. In other words, we now have the following: Location report for stock at 22Q: 0 units (ie empty): Calculation: Stock on record (before picking) = 65 units Stock on record (after transaction) = 7 units (65 - 58). Cycle count adjustment to stock record at 22Q = - 7 units (ie adjustment down , a loss).
Similar calculations would be made if it was decided to consolidate stock from one location to another. For example, consider the three situations above if was decided to consolidate the stock in 22Q into Location 19R.
Return to Agenda Return to top of Agenda Go to Top of Course Go to end Return To Glossary
5.11c Periodic Checks to ensure all Locations have been are Verified
If a location is never emptied by a storesman, there will be no cycle count check as described above. Consequently, it is vital that some means be found for recording all locations over a period of time (say, six months) which have not been recorded as being empty, and then, separately, auditing the identity and quantities of stock there at such locations. Two methods of doing so have been employed.
(1) The easiest way of finding which locations have never been emptied over a 6-month period is by computer program. When a transaction is submitted as above saying a particular location has been emptied, the location code is written to a data file. At the end of six months, the location codes in this data file are sorted and de-duplicated, and the resulting list of codes compared to a list of all warehouse locations codes. The warehouse locations not on the data file are reported for subsequent auditing.
For example, suppose the following locations are reported as having been emptied, and are on the data file before sorting and de-duplication: C30, A10, F60, B20, C30, E50, A10, B20, E50, F60,
After sorting and de-duplicating, we have A10, B20, C30, E50 and F60
Suppose the complete list of warehouse locations is as follows: A10, B20, C30, D40, E50, F60, G70.
Then the following locations have not been checked, and should be audited:
D40 and G70
(Just because they have never been emptied does not mean there is a problem. The locations might, literally, never have been used, or there might be genuinely slow moving stock there.)
Note: A danger exists if the system thinks that a location is empty when in fact it is occupied by stock. The principal means by which this situation will be detected is by the normal operation of the variable (random) location 'putaway' program: thus suppose the putaway program directs new stock of some other product to be placed in what it believes to be the empty location, but when the storesman attempts to do so, he finds the location already occupied. It is vital in the operation of such the system in these circumstances that he should report the situation back, so that the location can then be audited.
(2) If it is possible easily to tell, physically, whether a location is empty or full, an alternative method exists for identifying locations other than through software and records history. Suppose that a chart, or map, of locations is drawn up. Every morning, the supervisor walks the aisles and places an X in the locations on the map corresponding to the physical locations that he sees are empty. At the end of a suitable cycle - say, six months - we note those locations against which an X has never been recorded. These locations are now audited. (Actually, they may have been emptied, but they were never empty when the supervisor walked by). This method was devised by Mr Alan Shea, then a manager at ICI Organics Division, for the high bay racking warehouse at the division's large distribution warehouse at Heywood, Manchester (UK).
Return to Agenda Return to top of Agenda Go to Top of Course Go to end Return To Glossary
5.11d Variable Location Cycle Counting - Reporting & Auditing
It is important to be aware that in operating the cycle count system either under 5.11a or 5.11b that there will not be a guaranteed total stock check of an individual product at any one time. With the Variable/ Random Location system, it can be guaranteed however that by working through the system and checking the required number of racks each week ... plus a physical audit of non-emptied locations at the end of the period ... a full stock check of each product will have been carried out
Note that if a discrepancy is discovered between the stock record and physical stock when checking empty locations throughout the year, every effort should be made to discover the cause of the stock record error and correct the underlying reason. This will clearly be more difficult to do when checking the non-emptied locations once a year when the audit trail has gone cold. A further refinement to the system augments the reports in 5.11a and 5.11b above if abnormal discrepancies are found. A reasonable system might suggest that ...
if 1 - 2 errors are discovered in a rack, ... no further checking need be done,
if 3 - 5 errors are discovered in a rack, ... an additional 10% of locations should be counted,
if 6 - 8 errors are discovered in a rack, ... perhaps an additional 20% of locations might be counted,
The system described in Sections 11b and 11c works well in practice and is robust; it provides the information needed by both management and auditors. In a large warehouse it can typically reduce the cycle counting effort - physical and clerical - by up to 80% compared to the normal method described for fixed location systems. In a fast moving warehouse such as one storing "FMCG" (ie fast moving consumer goods), it is the only method which is really effective and practical. Like any other system, however, management and staff alike are required to exercise discipline and control to ensure that it is satisfactorily operated. This is especially true with regard to the real cycle counting that takes place - ie the reporting of the zero stock locations by storesmen and warehousemen when they are emptied .
Return to Agenda Return to top of Agenda Go to Top of Course Go to end Return To Glossary
5.12 Reconciling Variances
Note that variances should always be subjected to a reconciliation process even though the error may be within tolerance. The investigation of a product record that is incorrect but within tolerance may evenso yield information leading to system improvement. The root cause of the error may be other than a difficulty in physical measurement.
A "variance" is a discrepancy between the count and the record. "Reconciliation of a variance" means, first, making an adjustment to the record, and second, discerning the basic cause of the error. There are two points:
(1) Making sure that there really is a variance.
This requires the recounting of the item, preferably by a second cycle counter. What the cycle counter must avoid at all costs is making a correction to the stock record when the first count is wrong. The first count might be wrong because, say, of the misidentification of parts; or the failure to find parts in a nearby location; or confusion about units of measure. Very often, a cycle counter can spot a "bad count" at once. It is the failure to recount which is at the centre of the cause of errors to stock records accuracy arising from the annual stock count (see Error 9 in Section 6.5).
(2) Determining the cause of error is as much art as science.
Oftentimes, an experienced cycle counter will be able to to assign a cause to a variance in just a few minutes or even just a few seconds, based on his knowledge of the product and its history. (The facility of immediate access to the transaction trail on-line is highly desirable to assist in the search for clues.) To be successful, he must have an intimate knowledge of the procedures described earlier, and bring to that knowledge judgment, quick thinking and inspiration.
Four methods by which a specific error might be investigated are described below. All have proved valuable in the past. See also Section 6.4 below.
(i) Brainstorming and the Fishbone diagram.
In generating and discussing ideas, it is important to contribute creative and even inspired thought rather than to recycle the past. "Brainstorming" is simply a controlled team get-together to elicit and encourage thoughts and ideas. The "apparatus" necessary to carry it out is likely to consist of sticky notes and a flipchart. The person leading the team must control the meeting well (ensuring, for example, that the less pushy types have their say, perhaps by asking for ideas from each person in turn). Someone should be appointed to write down progress. Suggested guidelines for brainstorming are as follows:
Discussion Rules: suspend judgment; strive for quantity; generate wild ideas; build on the ideas of others.
Team Leader's Rules: be enthusiastic; capture all the ideas; ensure there is a good skills mix at the meeting; push the team for quantity; strictly enforce the rules; keep the intensity high; get participation from everyone.
A so-called fishbone diagram is an aid to the brainstorming as described. The simple diagram comprises a straight central line (the backbone of the fish) and half a dozen lines at 45 degrees above and below the centre line (the bones of the fish). Use of the diagram began in 1952 in Japan in connection with manufacturing quality problems, to enable possible reasons for effects to be recorded easily and in a way suitable for the openness of team working. Five categories of 'cause of error' might be used to guide the group to a solution, each cause being represented by one of the 45 degree bones of the diagram, Traditional causes used in fishbone diagrams are: materials, measurement, machine, man and method. That is, when an idea is mooted during the team discussion, it is written on one of the five fishbones depending on the category into which the idea falls.
(ii) Pareto Analysis.
When causes of error have been determined, each cause can be given a category number. It will then typically be found that a small number of causes give rise to a large number of actual incidences of error (ie what will be found is the Pareto effect). The cycle counter should concentrate on the small number of these "Class A errors" in determining what action is to be initiated to reduce errors.
(iii) Flip Chart Suggestions.
The Cycle Counter sets up a flip chart, stand and pen located within the stores at the entrance. A specific, difficult incidence of error that cannot be resolved is recorded on the flip chart, and storemen are invited over, say, a week to write their ideas as to cause on the chart. Storemen are also encouraged to write on the chart other causes they know of besides the current one. The suggestions are reviewed at regular stores meetings.
(iv) Scrutiny of the Transaction Trail.
The knowledgeable and intense scrutiny of a transaction trail has already been touched on above. The audit trail must be accessible on-line and optionally as printed reports, with the use of data filters provided as an easy-to-use option. Data filters are an essential tool in using the transaction trail, to in order to screen out unwanted data, to enable people to focus of groups of data of interest and in order to list the transactions in whatever sequenceis best for the investigation on-hand. The data connected with each transaction record will include the transaction type, item, time and date, quantities, locations etc..
The outcome of the reconciliation process is, first, a correct record, and, second, an enumeration of the type of error responsible. Note that what is needed is an indication of the type of error (for example, 'misrecording', 'lost transaction', 'units of measure' ...). Usually, this will allow the cycle counter to pinpoint the actual moment or incident when the error occurred. However, since the objective is not to chastise a member of staff but potentially to take action in relation to the system, pinpointing the incident is not actually necessary. The manager is not there to blame the storeman. The storeman is a (fallible) human being who works within the stock recording system provided. The system is the responsibility of management and management alone. Note finally that a time limit should be put on individual investigations - many causes can never be discerned and are part of random variation.
Return to Agenda Return to top of Agenda Go to Top of Course Go to end Return To Glossary
5.13 Performing the Cycle Count, Documentation & Final Audit
Performing the Cycle Count
1. Choose the Items to be Counted and Close off Stock
The computer program which identifies the items to be counted should be run each day (*). It is then necessary for the cycle counter to seal off the items from further picking until the cycle count has taken place. (see ABC, Low Stock etc.)
2. "Select" the Stock Records
Some time after the close-off of the location, when it is estimated that all old, outstanding transactions have been processed, the program is run which sets special software "triggers" on the database to prime the system for an impending automatic reconciliation requirement as described in Section 4. (If it cannot be guaranteed that old transactions have been cleared, 2-Step Reconciliation must be used instead, so that any late, old transactions can be excluded from the reconciliation - see Section 4 again.).
3. Count the Stock and then release it
The counting proceeds, each count made being entered as a normal transaction, either direct into an electronic data recorder (batch), into an RDT (on-line) or onto a paper form (batch). If default reporting is to be used, the cycle counter need only tick a box to say "OK". In most systems using automatic reconciliation as described in Section 4, the material can be released as soon as the count has taken place.
4. Reconcile any Variance
5. Post Variances and Corrections
As explained in Section 4, the Reconciliation program can be run at any time after the count has taken place - if transactions have been recorded on an electronic device, the cycle counter may leave them to be input overnight, the input to be followed by execution of the reconciliation program.
Note : Before starting the full programme of cycle counting, it is advisable first to select a small, but representative sample of items - say, 500 in all - as a test. The purpose in doing so is to validate the system, the computer programs and the various procedures, including reconciliation. All snags can be ironed out before expanding to the complete stores. If cycle counting is being introduced at the same time as a new stock recording system, then shortly after successfully 'going live' as described, it will next prove necessary temporarily to accelerate the programme. If this is not done, the system will continue to report poor accuracy stemming from before the commencement of the improved system.
Documentation
First, to satisfy the auditors on the points above, and for the sake of good management, it highly desirable that the cycle counting system be fully documented and that a record be kept of the adjustments made to stock levels (including those adjustments expressed in financial terms).
Final Audit
Finally, if the auditor is still reluctant to give up his annual stock check, as a final gesture, a financial audit of the stores/warehouse stock can be quickly carried out as follows.
First, analyse the products by calculating, for each, its annual value. As previously explained, this is defined as
a product's annual usage x its unit value (so 100,000 per annum × £3 each = £300,000 annual value).
This analysis is commonly done in ABC analysis as described above. It will be found that the top 20% of the products (called the ‘A' class products) will account for about 80% of the total annual value.
Next audit approximately half of the A class products, calculating for each one the financial variance between its record value and its actual value. Thus suppose the record value was £600,000 and the total variance was - £250 (allowing any positives to cancel any negatives). As a percentage, the variance is thus - £250/£600,000 or - 0.0417%. Now calculate the total record value of all the stock in the stores. Say, this is £3,000,000. Using the financial variance figure just estimated, the stock variance figure from £3,000,000 is thus - 0.000417 × £3,000,000, or - £1251, giving a revised estimated stock total of £2,998,749.
Return to Agenda Return to top of Agenda Go to Top of Course Go to end Return To Glossary
5.14 The Valuation of Stock
Note that this Section is entirely concerned with the value of what is properly designated stock. The designation and valuation of work-in-progress is a separate, specialised matter, and is dealt with in Section 4.3 and, particularly, Sub-section 4.3c.
The valuation of stock must pay regard, as ever, to the accounting principles of consistency and prudence, and for stock these may be summarised by the rule that value should be based on the lower of a material's manufacturing cost (or purchase cost) and its net realisable value.
Thus for raw materials, any spoiled or obsolete material should be valued at its scrap or sale value. That apart, if a bought-in part can be individually identified with the item as it was specifically delivered by the supplier, its valuation may be made by reference to the supplier's invoice. Even if it cannot be individually identified, but its purchase price is known not to vary, valuation can be made on this basis. For raw materials liable to variation in price but which cannot be individually ascertained, three principle possibilities exist: FIFO, LIFO and average price . FIFO (First In / First Out) is a commonsense method and the one most usually used. Issues to the factory of raw material stocks are assumed to have a value corresponding to the earliest possible price paid for them, regardless of their physical identity. For example, suppose 200 items were bought in January at £1 each and 150 in February at £1.50 each, making a total stock of 350 units. In March, 250 units are issued to the factory. It is assumed that the value of the units issued is £275 (200 of the older units, at £1 + 50 of the later ones, at £1.50). With LIFO (Last In / First Out), issues from stock are assumed to be at the latest price that was paid. In the case of the foregoing items, the value of the 250 units issued to the factory is £325 (150 of the newer units at £1.50 + 100 of the earlier ones at £1). In the average price method, an average price is calculated by dividing the total amount paid by the total number of items bought. In the instance given, this is £1.21 (£425 / 350) and the value of the units issued is deemed to be £302.50.
Finished goods are valued at their production cost (ie based on their direct cost and production overheads - note that the production cost, is not a fully absorbed cost, such as may be required for price setting or investment appraisal purposes - still to be incorporated are general company overheads not concerned with production, such as sales expenses, R&D, and central administration). It is wrong to value a finished good at its fully absorbed cost, let alone its selling price, since in the first case the full cost incorporates period costs not necessarily incurred in the period of actual manufacture, and in the second case, even if a sale seems certain, the selling price clearly incorporates value from an activity that has not in fact taken place. Two fundamental principles of accounting are (1) that stock should be regarded as an asset until a sale has been made, and (2) that costs incurred in manufacture may only be classed as expenses in the profit and loss account when they relate to material that has actually been sold. If the costs incurred relate to stock, they must be incorporated in asset value. The costs of manufacture migrate in the accounts from being assets (having potential value) to being expenses (something used up and of no value) when, and only when, the sale is made and the asset's ownership is transferred to someone else.
Components and 'Intermediate' Goods. Almost all of the comments above relating to finished goods apply to components and intermediates. That is, they are valued at their production cost, as determined by the company's standard costing system.
Return to Agenda Return to top of Agenda Go to Top of Course Go to end Return To Glossary
6.1 Personal Accountability
6.1a The Stores/Warehouse Manager
The duties and responsibilities of the manager may be summarised under three headings:
First, there is his concern with the technical procedures and apparatus inherent in the job - materials handling, racking, equipment and safe systems of work, for example.
Secondly, there are man-management responsibilities. Differences here are enormous, from the supervisor of a small stores in charge of two or three staff to the head of department in a major warehouse employing perhaps two or three hundred staff.
Thirdly, there are the duties and responsibilities of the manager in relation to the stock recording system. These may be summarised under two headings:
1. The smooth operation of the system from day to day, and the system's improvement so as to reduce the incidence and impact of common causes of error. As stated previously, common causes of error are natural phenomena inherent in the system's operation (as it stands at a particular time). They arise because of the way the system works, because of the nature of the environment in which it operates and because of the fallibility of people as they work in any human-centred system. If the effect of common causes is to be reduced and the system's performance thereby improved (ie accuracy improved), management must take action to change the way the system itself works, defining "system" again here in the broad terms introduced in Section 3. Only management can put these changes into effect. Examples of changes that might be made to the system are the introduction of a policy of limited access to the stock and the purchase of better weighing and measuring machinery. Note that the natural capability of the stock recording system must be determined statistically, by plotting accuracy figures obtained over time, as illustrated in Section 6.3 below. Management must not attempt to attain improved accuracy by interfering in the operation of a stable system by storemen or warehousemen, in order to "encourage" them to work more carefully, or harder, or otherwise differently. Action of this nature taken on a stable system will simply destabilise it and make matters worse.
2. To identify and eliminate special causes of stock records inaccuracy. Special causes are essentially "one off". As with a state of stability, the intrusion of special causes must be detected statistically (see below) ... they result in a temporary level of inaccuracy outside the natural limits of accuracy of the system. By definition, a special cause is not part of the system. When its presence has been identified, the nature of the cause must be determined and action taken to prevent it ever occurring again. An example of a special cause is the loss of transaction data due to a computer malfunction.
As well, the duties and responsibilities of the manager include the management of stock itself through his use of the system. Reports and queries may be generated to identify and investigate slow moving items, parts nearing their expiry date and the percentage occupation of the facility's available storage space.
Return to Agenda Return to top of Agenda Go to Top of Course Go to end Return To Glossary
6.1b Stores/Warehouse Staff
It is essential that everyone capable of affecting stock records accuracy, but especially those working in the stores or warehouse, be made aware through a programme of education why data must be accurate, and what the penalties to the company are of poor data accuracy. The operation of the stock recording system must be seen as a primary task of the job. The environment of very high levels of stock records accuracy changes many things. And because of these changes, it is essential to define the new job responsibilities of staff in some detail, and to make the bounds of authority clear. Examples of job headings to be incorporated in the job description are:
* completing transactions correctly; * submitting transactions to the system in time; * ensuring items are correctly labelled and located; * keeping basic data accurate; * location records; * making decisions on quality/scrap; * maintaining a tidy stores; * picking and kit marshalling; * approving unplanned removals of stock (say, for tests); and * correct materials handling and storage.
It is suggested that the new regime will require the storeman's and warehouseman's job to be formally re-evaluated through a point-factor job evaluation scheme (such as the Hay System), paying particular attention to that element of the job demanding clerical aptitude. In the great majority of companies in the UK, the storesman or warehouseman's job is graded, for the purposes of deciding remuneration, by means of a traditional, but crude, job grading, or classification scheme that is totally unable to take account of the responsibilities of the job holder in the matter of stock records. With a point-factor job evaluation scheme such as the Hay System, instead, however, subtleties such as maintaining stock accuracy can be incorporated into the evaluation process, as can other requirements of the job such as the need for clerical aptitude.The Human Resources Manager, who is ultimately responsible for job evaluation, should have the sophisticated and central nature of the stock recording system brought to his attention. He should be made aware of the critical role of stock records accuracy in other company wide systems. He should be aware of the complexity of the system and its dependence on valid, timely transaction data. The grading of the stores job by a simplistic - and superficial - grade evaluation system will inevitably result in the job being "under-evaluated" - that is, having a rate of pay attached to it that is significantly less than is merited and, perhaps more to the point, is required to attract the right candidates.
Two further issue relating to staff are recruitment and selection, as follows:
Staff Recruitment
Staff recruitment means persuading people to apply for a job in the first place (compare selection, which means selecting someone from among the candidates who have applied.) Recruitment is an activity always carried out by HR department. It is a truism that if recruitment does not attract the right candidate, there can be no possibility of a right selection. Given the inherent uncertainties of the selection process, therefore, a good deal more effort should be expended in recruiting people to apply for a vacancy than currently seems to the case. In particular, the company must consider carefully whether the practice of filling jobs in the stores/warehouse by internal recruitment, often from the shop floor, only can possibly be justified. Storemanship is a profession, and there is a clear case for external recruitment if this is what it takes to find the right personnel. Working in the stores is not an easy option. Advertisements should stress the professional nature of the job and the need for someone who fulfils all the desirable attributes, and should be placed so as to attract a good response from qualified applicants. If the care taken in recruiting is a reflection of the 'cost of making a mistake', HR department should reflect on the cost to the company of a poorly operated stock recording system, the cost of careless counting and the cost ultimately of wrong stock records.
Staff Selection
Selection of the recruit from those applying might be made on the basis of the following four criteria: (1) a rigorous assessment of the truth of the attributes and experience claimed on the application form, (2) demonstrated practical ability and knowledge in materials handling, as manifest in a practical test administered at the interview - for example, a ten minute fork truck driving test; (3) the satisfactory completion of a written aptitude test examining numeracy and clerical ability (*); (4) satisfactory eyesight and colour vision (dyslexic storemen are not a good idea). [* One simple test administered by one company is to give the applicant a sheet of paper on which is printed a list of, say, 30 product codes. The candidate is required to copy the codes onto a second sheet of paper! Not too many applicants can do so correctly.] It was stated in Section 3.1 that the stock recording system should be viewed from the "outside in" - ie viewed in its entirety. One of the required views is very obviously the calibre and suitability of the staff to be recruited into the stores in the first place. Clerical aptitude and good vision are minimum requirements, and it is up to management to see that the required standards are reached, by careful selection, broad recruitment and commensurate pay.
Return to Agenda Return to top of Agenda Go to Top of Course Go to end Return To Glossary
6.1c The Cycle Counter
One of the primary benefits of cycle counting described in Section 5 is the establishment of a centre of excellence in stock checking, in the form of the activity of the cycle counter himself. For this to become reality, however, requires that the cycle counter's appointment be made with care. The personal attributes required of the job holder are that he should be thorough and committed, have a good clerical aptitude and possess patience. The attributes of the job holder related to his knowledge are also essential: he must know the products he has to count and must know the stock recording and cycle counting 'systems'. These systems, in part, are (1) the paperwork involved in the stores and in the shop; and (2) the stages of a product's life from shop to stores, then back again to the shop and then back once more through manufacture. Remember also that the cycle counter must have the intelligence to be able to get to the bottom of record variances.
Appointment to the cycle counting job should also be accompanied by education in the theory of cycle counting itself. This will be very much easier with the existence of a cycle Counting Manual as described in Section 5.13 above. It is suggested also that close study should be made of the Course Notes at this Internet site. It goes without saying as above that the appointee should be thoroughly familiar with the stores' or warehouse's stock recording system and transactions. Note that the cycle counter is responsible merely for his own counting, not for the accuracy itself of the records. The practice of rotating the job of cycle counter between all members of the staff is deplorable and inimical to the achievement of eventual expertise in the job. Even more deplorable is the practice of having all members of staff 'cycle count' together on (say) one half day of the week, or for an hour at the beginning of the day. These schemes are often in place because the manager has had no guidance on how to set up the system properly, or because he believes he cannot find the time to institute a proper system. There is no harm in the cycle counting job for financial reasons being part time, as described below, but whenever the activity is undertaken, it should always be by the same person selected as being temperamentally capable of doing it and who takes responsibility and pride in doing it well, and consistently well. In addition, if the cycle counting job is viewed as an audit function, as indeed some might wish to view it if the accountant has given up his annual stock check, it is suggested that the cycle counter should respond managerially to the internal auditor or even to the accounts manager. If he found irregularities in stock recording, it would be difficult if he responded to the stores supervisor. If the job is viewed merely from the standpoint of records accuracy, however, the cycle counter might respond to the supervisor. In this case, the cycle counting system, documentation and programme should be subject to periodic audit by accountants as suggested in Section 5. 2.
Return to Agenda Return to top of Agenda Go to Top of Course Go to end Return To Glossary
6.1d The Purchasing Manager
As if to reemphasise the 'outside in' view of the stock recording system, a note should be made of the responsibility of the company purchasing manager with regard to the receipt of raw materials. What is required of incoming material from suppliers is that it should be (1) the right material, as ordered, (2) on time, (3) of the correct, stated quantity, and (4) of the agreed, stated quality. If deliveries are not on time, the daily planning of the stores/warehouse is disrupted. As well, to protect itself against lateness, the company may order goods for delivery before they are actually needed, thereby unnecessarily increasing the overall stock holding. Deliveries which are not of the correct quantity also cause problems. A decision must be made whether or not to open, count and inspect incoming packages. If experience indicates a supplier to be unreliable, it is likely that counting and inspection will be undertaken. Again in the matter of quality, the percentage of non-conforming parts sent by the supplier may also be inconsistent, calling for the sampling and inspection of incoming units. The sampling of incoming units is a complex subject, resulting sometimes in incoming loads which fail inspection being returned to the supplier. Quality inspection on arrival, or the placing of incoming material in quarantine for later inspection, results when the supplier has failed to control his own manufacturing process.
All of the problems and challenges above fall to the purchasing manager. However harsh it may be to say so, it is his responsibility to source material from suppliers who are competent in the matters of quality and correct delivery. And since this is not a perfect world, the purchasing manager should develop close relationships, Japanese-style, to bring suppliers up to the level of performance needed by the company. The Goods-In section of the stores acts as a back-stop only, not a first line of defence. See also Gooods-In, above,. Also see the purchasing manager's necessary collaboration with the sales manager, regarding the selection of an INCOTERM.
Return to Agenda Return to top of Agenda Go to Top of Course Go to end Return To Glossary
6.2 Training
A responsibility of the stores supervisor is to provide training for his staff. It is not acceptable that staff should carry out tasks week after week and even year after year in the wrong way. By the "wrong" way is meant by a method that is more susceptible to the making of mistakes or which takes longer or which in some other way is inferior to the one approved procedure. In giving training, the supervisor must make best use of his staff's time by thorough preparation and by ensuring that the stores' work that would otherwise be performed by those being trained is undertaken by other staff on a temporary basis. What is highly undesirable is that staff under training should be continually worried and distracted by work piling up for them outside, as the training proceeds. Staff respond differently to different styles of training, so it is best to mix styles.
Classroom Training.
This is usually the most effective and the best use of training time. Procedures are introduced and described at length, with a variety of presentation aids (overhead foils, dummy packages, demonstrations of equipment ...). The trainer must make it clear that any procedure laid down vis a vis a transaction* is to be the only way of carrying out the task. If a storeman knows a "better method", the place to examine it is at a staff meeting. *Also see the important explanation of the role of transactions in Section 4.4.
On-the-Job Training.
For too many stores, on-the-job training, so called, is no training at all. For them, what 'on-the-job training' amounts to is hoping the storeman will pick up the job as he goes along. On-the-Job training should involve the appointment of a mentor and the issuance of checklists and procedures manuals. There should also be a formal review and assessment at completion
Self-StudyTraining.
Self study usually involves the personal study of a step-by-step manual and guide, followed by the completion of a self-test (multiple choice) exercise. The stumbling block to this, very evidently, is the availability or otherwise of the study guide. A long term goal of the supervisor might be to compile such a guide over the years. Computer training means training on a training database, a training version of the stock records database which is reversed to its original state at the end of the training session (see next sub-section).
Training Database.
A training database means a training version of the stock records file and training versions of the transaction programs required to update them. The trainee is presented with a variety of situations relating to stock and stock movements, and he must raise (training) transactions and update the (training) stock records. His work can be readily examined by his mentor through an audit trail, and at the end of the training session the database can be returned to its initial state. To be effective, the training database must look exactly like the live file. Training versions of the database and transactions can be readily created merely by copying the live data.
CYCLE COUNTING TRAINING
Training specifically for cycle counter must clearly involve procedures training as described above. As well, it must include counting training and instruction in the use of such equipment and measuring devices as scales . Instruction and guidance must be given on the policy of opening and counting packages, moving and organising material prior to a count and the counting of items which are inherently difficult to count. Finally training must be given in relation to the reconciliation of variances, including the interpretation of transaction trails.
Return to Agenda Return to top of Agenda Go to Top of Course Go to end Return To Glossary
6.3 Determining the Limits of Performance of the Current Stock Recording System
The first task prior to commencing a programme of system improvement is to bring the operation of the current system under control and so determine its 'capability'. When the system is under control, it will be found that the records accuracy attains an average value with an equal variation between consistent limits above and below the average. To determine what the average is, and determine the upper and lower limits of variation, it is suggested that 25 parts are counted every day over the course of a month and the % accuracy each day recorded in the normal way. The percentages may be plotted on a graph over the month. When this is done, the graph will take the form of a zig-zag between the upper and lower limits, spread about a central average. The average line and the upper and lower limits will be different for every company ... 80% ± 7%, 70% ± 10%, 68% ± 7.5% etc. What is important is that the results show stability between the two limits. The upper and lower accuracy limits of the graph cannot be simply guessed by drawing straight lines at what seem to be reasonable places (or places where you would like them to be). They must be calculated from the actual readings of accuracy over time. To do so requires the stores manager to work out the standard deviation of accuracy, which we will refer to here as "s", and then draw the lines at (3 × s) below the average and (3 × s) above the average. For example, if the standard deviation s was calculated (as described below) to be 1.1 and the average records accuracy was 80.0%, the upper limit would be drawn at 83.3% (80.0% + 3.3%) and the lower limit at 76.7% (80.0% - 3.3%).
The following six step procedure can be followed to find the accuracy limits of a system from 10 accuracy measurements.
Step 1. First, add up all 10 measurements and divide the total by 10 to find the average, referred to as the mean.
Step 2. Next, write down the "absolute deviation" of each of the 10 accuracy numbers from the mean. To do so, find the difference between each reading and the mean. If the difference is a negative, ignore the negative sign.
Step 3. Now calculate the average, or mean, of the 10 differences from Step 2. This result is the mean absolute deviation, or MAD.
Step 4. Multiply the MAD by 1.25 to obtain the standard deviation s.
Step 5. Calculate the upper boundary of records accuracy as being the average from Step 1, plus 3 × s.
Step 6. Calculate the lower boundary of the records accuracy as being the average from Step 1 less 3 × s.
Note that if the records accuracy system is fully under control, all 10 accuracy readings will be between the upper and lower limits - these are the stock recording system's limits. If there are any readings outside the limits, these are due to special causes of variation - ie the system at those points was not under control. Find out why, and remove the cause, as discussed further below. Follow the example below.
Example
To find the records accuracy of the group of records below over a ten week period, we first work out the average, or "mean", accuracy over the period by adding up the ten figures and dividing by 10. The total comes to 670, giving an average, or mean, of 67.0%. Next, we record the absolute deviation of each of the 10 readings from the mean. These figures are given in brackets after each accuracy figure. Remember that negative signs are ignored.
Week 1 .. 66.7% (0.3%) ........ Week 2 .. 67.3% (0.3%)
Week 3 .. 65.9% (1.1%) ........ Week 4 .. 68.6% (1.6%)
Week 5 .. 67.8% (0.8%) ........ Week 6 .. 66.4% (0.6%)
Week 7 .. 66.3% (0.7%) ........ Week 8 .. 68.0% (1.0%)
Week 9 .. 67.5% (0.5%) ...... .. Week 10 .. 65.5% (1.5%)
The figures in brackets add up to 8.4, so to find the mean we simply divide by 10, to obtain 0.84. This is the mean absolute deviation or MAD. Consequently, from Step 4 above, the standard deviation "s" is 0.84 × 1.25, = 1.05. Therefore 3 × s = 3.15, so that from Steps 5 & 6 above, the upper and lower accuracy limits are 67.0% + 3.15% and 67.0% - 3.15%, or 70.1% to 63.9% (rounded). This is the capability of this stock recording system.
Return to Agenda Return to top of Agenda Go to Top of Course Go to end Return To Glossary
6.4 Bringing the System under Control by eliminating Special Causes of Error
A constant problem in dealing with stock recording errors is determining what the actual causes are. A special cause error is defined here as a discrepancy between the stock record and the physical stock quantity of a particular product, (other than a discrepancy due tolerances or out-of-order transactions) and which does not arise from the normal operation of the system. Unless there is a well-known computer system or computer program fault, which would allow someone investigating the transaction trail to detect the point where the discrepancy occurred, it is extremely unlikely that the cause of a given error will ever be found with any certainty. For example, if the record says 40 units and the stock count is 50, how did the difference arise? Last January, perhaps the receipt of 100 units was counted in as 90 by mistake ... or maybe it was two receipts of 70 and 100 counted in as 75 and 85. One problem with the causes of error, then, is finding out what they truly are, so that steps be taken to eliminate them (special causes) or change the system (common causes). The question of reconciling variances was dealt with in Section 5.12, which also gave four methods 5.12 (i) to 5.12 (iv) which might be used to help determine them.
Return to Agenda Return to top of Agenda Go to Top of Course Go to end Return To Glossary
6.5 Reducing the Effect of Common Causes - Management Action to Improve the System
The bulk of the task of system management is identifying and dealing with common causes of inaccuracy. And since common causes are inherent in the very operation of the system, "dealing with" them means that management must alter the system - for example, by changing procedures; by adopting new rules; by amending software; and by purchasing new apparatus. The list below identifies 23 causes of error, in order of their prevalence, these being mainly common causes. The list was obtained over many years by David Crabtree, by asking attendees on his UK Stock Records Accuracy training course to vote which causes were most prevalent in their various places of work.
1. Incorrect Counting ................................................. 11.8%
2. Missing Transactions .................................................. 9.0%
3. Insecure Stores .......................................................... 8.6%
4. Incorrect Recording .................................................... 8.1%
5. Incorrect Picking (ie wrong item) ............................... 6.9%
6. Stock placed in Wrong Location .................................. 6.6%
7. BOM and/or Backflushing Errors .................................. 5.6%
8. Slow Reporting of Transactions ................................... 4.8%
9. Errors introduced by the Annual Stock Count ............... 4.3%
10= Working under too much Pressure ............................. 4.1%
10= Misidentification of Parts ........................................... 4.1%
12. Data Keying Errors ....................................................... 4.0%
13. Scrap or Production missed .......................................... 3.9%
14. Use of the Wrong Transaction ....................................... 3.1%
15. Wilful lack of Care ......................................................... 3.0%
16. Units of Measure ........................................................... 2.8%
17. Duplicates (ie same transaction submitted twice) ......... 2.1%
18. Misidentification of Locations ........................................ 2.1%
19. Failure to count Raw Material Receipts .......................... 1.9%
20. Shrinkage (unreported deterioration/write offs) ........... 1.6%
21. Incorrect Picking List issued .......................................... 1.1%
22. Theft .............................................................................. 0.8%
23. IT/System faults ............................................................. 0.1%
It is the job of the storeman to work within the system as it is defined and provided. Only management can make the changes to the system necessary to eliminate the causes giving rise to errors and variance.
Return to Agenda Return to top of Agenda Go to Top of Course Go to end Return To Glossary
6.6 Action to Eliminate Typical Errors
Ideas for change may spring from the list above, or from scrutiny of the overall system flowchart. The following actions related to the error list above might be taken to change the system and reduce the number and impact of common causes:
Incorrect counting/recording: (1) Administer a suitability test involving clerical aptitude, eyesight and colour vision to all job applicants; (2) simplify the stores codes (eg shorter product codes, simpler location codes); (3) adopt the following do's and don'ts:
If items are in layers, rearrange them so there is only one odd layer;
Count items from one container to another;
Count in multiples (2, 4, 6 ... );
Count small parts on flat surfaces;
Touch and count (large parts);
Standardise the formation of letters and numbers;
Data keying errors: Install an on-line system with immediate data feedback. It is very likely that if the staff keying in the data are having to read transactions prepared by others, that they are misreading what has been written, so .... Give handwriting practice to staff who complete transactions. Give them paper forms divided into squares, and instruct them as to how they should form capital letters and numbers. They have not received instruction such as this since they attended primary school.
Missing transactions: Introduce Radio Data Terminals and bar coding;
Duplicate transactions and units of measure: Program software checks;
Incorrect Picking: In a random location facility, introduce the 'Magic Number' location check digit system described in Sections 3.4. Also, have strictly one part per location;
Misidentification of parts: Improve labelling and lighting;
Use of the wrong transaction: Set up a training database, and make sure everyone follows refresher training from time to time.
Return to Agenda Return to top of Agenda Go to Top of Course Go to end Return To Glossary
7.1 The Role of Management
Before work starts on the project to improve the company's stock records accuracy, it is essential that senior management should accept three crucial premises:
(i). The project to improve records accuracy is a formal company commitment of major importance;
(ii). The goal of the project - high stock records accuracy - is a critical management and company statistic;
(iii). The responsibility for managing the system, for ensuring obstacles to achieving the accuracy goal are eliminated, and for the removal of common causes of error belong to management alone.
The improvement required to make a substantial difference in stock records accuracy will come about only by treating the programme of activities necessary as a fully-fledged, sanctioned and controlled project. Casual exhortations to "sort out that mess once and for all" will not bring about the required results. Prior to the presentation of the cost/benefit case, therefore, it is desirable to educate management about the company-wide repercussions of poor records accuracy ... on the shop floor, in marketing, in manufacture and in logistics ... and to show why high accuracy is a prerequisite for success in any planning system, be it MRP or any other. As part of the proposal, the "project champion" will need to evaluate it in traditional cost/benefit terms. The problem with all such proposals is quantifying the money that will be spent and - very difficultly - quantifying the financial benefits that will be achieved. On the costs side, items will include securing the stores and purchasing and implementing the system (WMS ... warehouse management systems ... can be costly). The benefits include the tighter inventory control possible; superior factory control; reduced stock losses; realisation of the benefits of effective planning systems; and reduced storage requirements. Stock records accuracy would make a perfect 'Six Sigma' project.
7.2 The Data Processing System
Clearly, it should be recognised that there is a need for formal, properly designed data processing procedures to support the system. The procedures must be well-structured and designed. And because of the speed of events affecting stocks - "the speed of data" - they must be supported by on-line computer processing. Among the many constituent features of a DP system is the system manual - a formal statement of its objectives, procedures and facilities, and descriptions of the responsibilities of staff whose jobs are to operate it. The manual should be up-to-date, easy to comprehend and not necessarily short. A problem well-known with systems, however, is describing them. The system described in the manual will be characterised as stated by data procedures, transactions, information flow, decision points and the activities and responsibilities of the people who operate them. But the English language is a inappropriate tool for conveying this information. We must therefore turn to tables, concise notes, tree diagrams, decision charts and - above all - the flowchart. This is a fundamental starting point in systems design, and a powerful means of ensuring understanding of the operations and information flow inherent in the final system. There are a number of flowcharting techniques in use. One of the most powerful is structured flowcharting, a methodology first devised by Gane & Sarson.
The Structured Flowchart
A structured flowchart shows the flow of a typical set of activities involved in a system. The representation is drawn using a data flow diagram, a diagramming technique which involves only three symbols and an arrow. The symbols are (1) a process or activity, an upright oblong box with rounded corners, (2) a data store, a thin, elongated shape that is open on the right and (3) a source of data external to the system, being a box with the two shaded sides. In fact, three or four 'levels' of flowchart are necessary, as follows. The first (top) level flowcharts on a single piece of paper the main, overall system activities, and typically consists of 5 or 6 process symbols, to be numbered 1, 2, 3, 4, and 5. The second level is drawn on several more pieces of paper, each one expanding the details of one of the process steps at the top level. Thus the first piece of paper at the second level expands Process 1 at the top, with new, detailed sub-processes 1.1, 1.2, 1.3, 1.4 and 1.5. Again, the second piece of paper at the second level expands Process 2 (at the top) with detailed sub-processes 2.1, 2.2, 2.3, 2.4 and 2.5. The third level of detail consists of numerous diagrams at the 'sub-sub-level'. For example, there may be a sub-sub-process consisting of the steps 1.1.1, 1.1.2, 1.1.3, 1.1.4 and 1.1.5 ... and so on until no further breakdown of detail is warranted. The flowcharts from the top overall level to the final level of detail are said to be structured.
Walkthroughs and Inspections
A problem with any system is that it is difficult to gauge its potential impact and operation until the time and money necessary for its installation have been spent, by which time it is too late!. Compounding this are the dangers springing from the closed, creative nature of the design process and the specialised language that design solutions can be couched in. There are two powerful techniques available to help matters.
The Structured Walkthrough
The first of them exists to block the possible adoption of a poor design. It is to subject the design being proposed at an early stage in its life to a 'structured walkthrough'. The structured walkthrough is a half- or one-day meeting at which the designer explains to a small group of staff including the stores manager the system's workings and the associated data transactions in order for him (the designer) to hear, and take account of, the group's comments and criticisms. The explanation starts with the top level flowchart of the structured flowchart (as above) and proceeds over time to the bottom. It includes a synopsis of the transactions to be provided. Note that until the structured walk-through has taken place, there is little point in making a fine job of the actual dataflow diagram drawing.
Inspection
While the structured walkthrough can help catch errors, its main targets are the quality of design and the general acceptability of the design solution. The purpose of this second technique by contrast is purely the elimination of errors. The 'systems inspection' method was devised originally by Michael E. Fagan of IBM in 1976. A system inspection takes place at several points in the system's life, whenever it may be appropriate, and consists of the examination over two hours of the system's documentation and diagrams by a 'moderator' and one or two other inspectors. The design and project staff are not present at the meeting. The purpose of the examination is to see whether the documentation that constitutes the system proposal and is the basis of implementation is clear, complete and unambiguous. The inspection team may refer to any background material it wishes. Careful written records are kept of the inspection. When anything is found that must be corrected, it is noted by the moderator and the correction by the systems staff followed up by him later. The moderator, a kind of one-man QA department, can call for a new inspection at any time. The employment of system inspections to check the emerging stock recording system might be viewed by some users as simply a welcome blow for democracy against being railroaded by computer people. It is more than this. Its use is supported by two facts:
(a) Over 60% of the errors in operational computer software were there before the programs themselves were ever written;
(b) The cost of removing errors in a system by inspection is less than 30% of the cost of removing them by conventional means that rely solely on system testing. The reason for this second statistic is that inspection finds errors far earlier in the system's life. Comparative costs to remove an error at various stages in the development cycle are well-known. If the cost of making a correction at the design stage is, say, £x, the cost of making the same correction after implementation is £500x.
Return to Agenda Return to top of Agenda Go to Top of Course Go to end Return To Glossary
7.3 Implementation of the IT System
The conventional approach to the implementation of a system as difficult as this is by following a familiar series of six steps in the following chronological order: (a). Systems analysis; (b) Systems design; (c) Selection of the Software Package; (d) Specify and plan the implementation; (e) Test the system; (f) Transfer the data and 'go live'; (g) Post installation audit & documentation. Control of a the project is usually best through a two-level management system consisting of a steering committee and a detailed project team. Members of the steering committee may be (in a manufacturing environment) the stores manager, materials manager and (say) the purchasing manager. In logistics/distribution, they will be the warehouse manager, distribution manager and logistics manager. The project team will vary in size depending on the nature, scope and ambition of the task. The steps to be taken are well-known to systems professionals. Remember, there is a need for feedback, either through a structured walkthrough or through 'inspection' (above). The preparation leading up to the change to the new system should include countless practices of data switchovers and dummy runs. The big day should therefore be seamless. The usual purpose of these days of parallel running is not as a back-up in case the new system goes wrong, but to provide a comparison of the new output with the old.
Return to Agenda Return to top of Agenda Go to Top of Course Go to end Return To Glossary
7.4 Education
When asked what they would have done differently if they had to do it all again, time after time project leaders put the provision of earlier, and more, education at the very top of their lists. Education should be thoroughly prepared by the stores/warehouse supervisor, and may consist of much of the contents of this on-line course. Other training will consist of explanations of how the system works and sessions at a VDU with a special training version of the database and system, as explained in Section 6.2. One of the purposes of education is to impress on staff that management has high expectations of their performance and that competence shown in this field will meet with high approval. Continuous education will also remind stores and warehouse staff that their responsibilities and personal expertise differ from that of other staff within the company and of the standards expected of them in the field of records accuracy. Implied in all this is that internal stores/warehouse courses, presentations and discussion meetings should continue permanently, after the target level of accuracy has been achieved.
Return to Agenda Return to top of Agenda Go to Top of Course Go to end Return To Glossary
7.5 Selecting Software
It is a regrettable fact that off-the-shelf systems relating to stock records accuracy are very frequently chosen which simply do not do what is wanted of them. The functions of the system may be ill-matched to requirements, or the software may simply be poorly thought out. For this reason, it is essential that software is selected with the greatest possible care so that the chance of selecting a poor package is minimised (*). A frequent complaint made by British managers seems to be that the software system they are required to use has been thrust on them by their American parent company. ("It worked in the US, so it must be right for the UK.") In the event, the system proves unsuitable in the UK, because the UK and US company environments are different.
A particularly effective means of making a choice is by adoption of the Weights & Marks Method. There are eight steps as given below. And remember, cheap software that is wrong, is mightily expensive; expensive software that is right, is cheap.
Step 1:
A software selection team of two or three people should be appointed from among the personnel likely to make up the steering committee and project team.
Step 2:
The software selection team must thoroughly familiarise themselves with the subject matter of stock records accuracy. It is suggested that this should entail a close knowledge of this on-line course.
Step 3:
Make an initial selection of packages from the current edition of The Software Users' Yearbook. The initial list may comprise (say) a dozen packages.
Step 4:
Write to the vendors requesting information on their packages and eliminate those which appear least promising, to leave a a short-list of about 4 or 5 packages.
Step 5:
Invite the vendors of packages on the short-list to present their products. Get to know the products somewhat more thoroughly through a deeper level package documentation. It is likely that at least one package will be eliminated from the short-list.
Step 6:
Compare the packages on the remaining short-list using the 'Marks and Weights' method. In this, the selection team must discuss and list the important attributes that relate to the potential stock records system, and assign to them relative weights of importance out of 10. For example, If the team was comprised of the people A, B and C, the first four attributes might be weighted as follows:
..........................................................................................................................Andy Bob Charlie Average
On-line Reconciliation ........................................................................................4 ......3 .....4 ..........3.7
Random or Fixed Locations ...............................................................................6 ......5 .....5 ..........5.3
Two levels of Tolerances ....................................................................................9 .....10 ....9 ..........9.3
Multiple units of measure ...................................................................................7 ......7 ....10 *
* There is a major discrepancy between A, B and C in their perception of this attribute. Before proceeding, it is absolutely essential that this is resolved by discussion. It can be seen that the weighting process is as much education and mutual examination as anything else.
Step 7: The assignment of marks to the software packages in the final list is done by assessing how well each of them matches the attributes described above ... irrespective of the weight of importance of that attribute. This is again done as a team to avoid misunderstandings, and, again, the process is one of mutual education and the sharing of knowledge. If the package names are S1, S2 ...
...................................................................................Supplier Attribute 1 Attribute 2 A B C Avg A B C Avg
........................................................................................S1 ..........3................. 3 ................4 ..........3.3
........................................................................................S2 ..........6 .................5 .................7 ..........6.0
........................................................................................S3 ..........9 .................9 .................7 ...........8.3
........................................................................................S4 ..........3 .................1 .................6 ...........3.3
Step 8: It is essential that the selection team should visit at least a couple of users of the most favoured package, or even two packages and ask penetrating questions. One of the problems with software is that, even if it is functionally what is required, one simply cannot appraise how well it works until he uses it. By this time, it is too late. Listen very carefully to what the users say. Listen especially carefully, because they will not wish to admit they have made a mistake in their choice.
Return to Agenda Return to top of Agenda Go to Top of Course Return To Glossary
The End
![[Homepage]](butt1on.gif) |
![[About GMCS]](butt3on.gif) |
 |