## Abstract

We have had occasion to refer to the role of inventory a number of times already in this book, and specifically we have suggested that the degree of control which is exercised over the levels of inventory held within a marketing logistics system is a key influence on the reliability of product supply within the system. That is to say, inventory control is primarily concerned with ensuring that stocks of a company’s products are made available on a consistent basis in the light of the company’s service policy to its markets and the behaviour of market demand. The major part of this chapter is concerned with describing and illustrating the basic principles of inventory control in the context of differing assumptions regarding knowledge about demand, and the lead time between placing an order for stock replenishment and actually receiving the order. In one or two instances we shall assume perfect knowledge about demand in order to more effectively portray a concept, but in reality this situation is extremely rare. In fact, we can distinguish between different classes of knowledge about demand in the future, in accordance with familiar decision theory practice. It does not happen often that a decision maker, whether he is concerned about sales planning or inventory control, is either completely ignorant about future demand, or has perfect knowledge of future demand levels. We can think of one or two situations, perhaps, in which these conditions might apply: for example, the demand for a completely new type of product may not be known at all, though it would be a foolish company which would market such a product without any prior research into consumer attitudes and behaviour. At the other extreme, that of certainty, demand may be known as the result of contracted supplies to a known and defined market.

## Keywords

Lead Time Order Quantity Inventory Control Mean Absolute Deviation Safety Stock## Preview

Unable to display preview. Download preview PDF.

## Notes and References

- 1.For a detailed treatment of a number of key sales forecasting techniques, the following references are recommended: Albert Battersby,
*Sales Forecasting*(London: Penguin, 1968)Google Scholar - Gordon J. Bolt,
*Market and Sales Forecasting — A Total Approach*(London: Kogan Page, 1971), esp. chap. 7Google Scholar - John C. Chambers, Satinder K. Mulllik and Donald D. Smith, ‘How to Choose the Right Forecasting Technique’,
*Harvard Business Review*, vol. 49, no. 4 (July–Aug. 1971) pp. 45–74.Google Scholar - 2.In this context, care should be taken to distinguish between forecasts which are conditional upon marketing plans, and those which are unconditional — see Harper W. Boyd and William F. Massey,
*Marketing Management*(New York: Harcourt Brace Jovanovich Inc., 1972) pp. 145–6 and chap. 7.Google Scholar - 3.In fact, demand forecasts obtained by the exponential smoothing method tend to lag in the context of a strong secular trend, whether upward or downward, and this would explain the poor performance of α = 0.1 in the example. The lag can be adequately compensated for by obtaining an estimate of the trend.Google Scholar
- 4.Wendell M. Stewart, ‘Physical Distribution: Key to Improved Volume and Profits’,
*Journal of Marketing*, vol. 29 (Jan. 1965) pp. 65–70.CrossRefGoogle Scholar - 5.John E. Sussams, ‘Some Problems Associated with the Distribution of Consumer Products’,
*Operational Research Quarterly*, vol. 19, no. 2 (June 1968) p. 161.CrossRefGoogle Scholar - 6.For a more comprehensive treatment of stockout costs, see Mark Ellsmore,
*Evaluating Logistics Service — A Model Building Approach*, unpublished M.Sc. dissertation (University of Bradford Management Centre, 1974).Google Scholar - 7.For a summary analysis of the normal distribution, see John E. Freund and Frank J. Williams,
*Modern Business Statistics*(London: Pitman, 1958) pp. 145–52.Google Scholar - 8.The standard deviation is a common measure of variation used in statistical analysis; it is sometimes referred to as the root-mean-square deviation, because it involves taking the square root of the mean of the squared deviations of a series of values from the mean value. See Freund and Williams,
*Modern Business Statistics*, pp. 84–95.Google Scholar - 9.The exception, of course, is where demand is influenced by the increase in inventory locations.Google Scholar
- 10.This example is adapted from Russell L. Ackoff and Maurice W. Sasieni,
*Fundamentals of Operations Research*(New York: John Wiley & Sons Inc., 1968) pp. 63–4.Google Scholar - 11.Normative models attempt to specify how problems should be solved if some value criterion is to be optimised from the decision maker’s point of view. Stochastic simply means probabilistic, as distinct from deterministic. See David B. Montgomery and Glen L. Urban,
*Management Science in Marketing*(Englewood Cliffs, N.J.: Prentice-Hall Inc., 1969) pp. 9–17.Google Scholar - 12.For an excellent summary of the expected value criterion for decision making under conditions of risk, see Philip Kotier,
*Marketing Decision Making: A Model Building Approach*(New York: Holt, Rinehart & Winston Inc., 1971) pp. 257–60.Google Scholar - 13.For a description and evaluation of two different types of P-system model, see Ballou,
*Business Logistics Management*, pp. 303–9. Ballou makes the point that the P-system policy under risk conditions differs from the Q-system in that all demand fluctuations (both order period and lead time) must be protected against: the Q-system policy affords protection against demand variability during lead time only.Google Scholar - 14.In Figure 5.4(a), an order quantity of 80 units means that, on average, 12.5 orders are placed each year. Total ordering and carrying costs are therefore equal to (10 × 12.5) + (40 × 1) = £165. With EOQ, total costs are (10 × 7) + (71 × 1) = £141.Google Scholar
- 15.Table 5.3 could also be derived from an analysis of a sample of past demand data during lead time. The example shown here has been deliberately simplified, in order to facilitate illustration of the procedure.Google Scholar
- 16.The notation used here has been adapted from James L. Heskett, Robert M. Ivie and Nicholas A. Glaskowsky, Jr.,
*Business Logistics*(New York: The Ronald Press Co., 1964) pp. 286–7.Google Scholar - 17.An example of this iterative procedure is given in Ballou,
*Business Logistics Management*, pp. 299–302.Google Scholar - 18.Notice the insensitivity of EOQ in our example, where Q* = 64, 65 and 66 for service levels of 100%, 95% and 87% respectively. A relatively higher stockout penalty, π, would generate greater differences in Q* for varying service levels.Google Scholar
- 19.The technique is discussed more fully in IBM,
*Wholesale Impact — Advanced Principles and Implementation Manual*(White Plains, New York: IBM Technical Publications Department, no date), chap. VIGoogle Scholar - Philip Schary and Keith Howard, ‘Logistics Strategy and Inventory Decisions’,
*International Journal of Physical Distribution*, vol. 1, no. 1 (Oct. 1970) pp. 31–8.CrossRefGoogle Scholar - 20.Representing the mean and one standard deviation either side of the mean along the probability scale.Google Scholar
- 21.See IBM,
*Wholesale Impact*, p. 131.Google Scholar - 22.Strictly speaking, the MAD value should be calculated with respect to a time period which represents average lead time. Where annual sales figures are used, the measure of variability may be expressed in terms of forecast errors during the year (see IBM,
*Wholesale Impact*, pp. 129–31).Google Scholar - 23.For examples, see Schary and Howard, ‘Logistics Strategy and Inventory Decisions’, pp. 34–8; IBM,
*Wholesale Impact*, pp. 133–7.Google Scholar