Risk perception: Operational decision making implications; food world application

Abstract

The Indian retail industry loses Rs. 32 crores on an annual basis due to forecasting and indenting errors. It’s a value drain which caused by sub-optimal methods used for projections and forecasting. Such forecasting and operational problems get compounded in perishable goods as products have a short lifespan and flexibility required in forecasting such items is much more than others. ‘Foodworld’ is a retail chain that concentrates on food items. Almost 40% of perishable food items are fruits and vegetables. This study targets to improve operational efficiency within ‘Foodworld’ by primarily looking at ‘Forecasting errors’ and ‘Store specific operational errors’ in the F&V section. The study looks at primarily 3 stores of ‘Foodworld’ in Bangalore in Whitefield, R.T. Nagar and B.G. road. The method used for the study is 1) Information Analysis: Interviews with store managers to understand the local nuances of decision making in stores 2) Data Analysis: Analysis of sales of high ticket items sold between 01-Jun and 18-Jul of 2009. Information analysis showed that a certain process/rule of thumb(captured common information of interview section) of indenting was used uniformly across ‘Foodworld’ stores which did not factor short term fluctuations in demand and also did not factor in shrinkage patterns. Moreover, they did not look at the ‘expected utility’ of the store manager. Apart from this, operational problems of stealing, low consumer insight and improper handling of perishable goods were discovered through interviews. Thereafter, I analyzed store data with ‘Utility theory’ in mind. There were problems of over-indenting on regular basis in some SKUs and also uncontrolled shrinkage of some SKUs that contributed to a large portion of sales, for example bananas, oranges etc. I tackled these problems by looking at the store manager’s expected utility. If the indenting/shrinkage was not in line with the store manager’s expected utility then the results would indicate that some other process/factor was driving the decision on indenting and leading to losses via shrinkage. I understood the store manager’s risk perception by administering standard evaluation questions (Keeney and Raiffa). It was determined through the answer of these standard questions that the store managers were risk averse. By applying store sales and shrinkage and distribution probabilities on a utility function(X^(1-a)), I was able to derive their expected utility. These values have been captured in Appendix 2. The results gave a preference order of SKUs based on expected utility (captured in Appendix 2). It also showed that sales of certain SKUs like Banana and Orange were high on the expected utility parameter but were showing very high shrinkages. Hence, this analysis shows that new measures need to be put in place to control shrinkage in such high expected value SKUs. In the interview process, I had got to know that the sales incentives were on a store basis. Hence the losses made by one department could be compensated by others. So, according to data analysis, high value items were recording high shrinkages and on the other hand there were no incentives to control it on a department basis. This forms the base of my major recommendation. I recommend that new incentive mechanisms should be introduced which factor the store manager’s expected utility and hence incentivize him to reduce shrinkage. I also propose other remedial actions for better operational in the summary sections of data and information analyses.