# analyze the forecasting errors.

L. L. Bean uses several different calculations in order to determine the number of units of a particular item it should stock, whether it is a new item or a never out item. It first freezes a forecast for its demand for the upcoming season. This figure is a result of a consensus between the product people, buyers and inventory managers. Once the predicted demand is frozen, L. L. Bean uses its historical demand and forecast data to analyze the forecasting errors. The forecast errors are calculated for each individual item and a frequency distribution of these is made, which is further used as a probability distribution for future errors.

Thus, if 50% of the errors were within 0. 7 and 1. 6, the forecast for this year would be adjusted accordingly. Next, each item commitment quantity was tabulated using its individual contribution margin and salvage value if any. For e. g. if an item had a margin of \$15 if sold, and \$5 loss if not sold, the commitment value would be 0. 75. Hence the optimal stock to keep would be 0. 75 fractile of the probability distribution of demand. If for instance, the corresponding error for 0. 75 is 1. 3, the optimal stock to keep for that item would be 1.

3 * frozen forecast. Hence, this value is the stock for that item. Q. 2 We explain different scenarios to determine relevant costs and revenues. The first scenario is where the stock kept of a particular item is sold. In this case, all the costs related to buying and selling that product would be included. The selling price of the product, the cost of buying from the vendor, the carrying cost of that particular stock item, and the cost of marketing that item in the catalogue are the relevant costs to be included.

In the second scenario, excessive stock is kept and at the end of the season, it is still not sold. In this case, the relevant costs are the cost of buying the stock from vendor, its storage and marketing cost in the catalogue, and the salvage cost ( if any) of that particular item. In case there is no salvage cost for an item, its redundancy cost becomes relevant. The final case is where the stock for a particular item is not kept, but demand for that item is there, resulting in a stock out. Since L. L.

Bean is a direct ordering company, the cost of a stock out is very high for it in terms of lost goodwill. Hence, the cost of goodwill lost becomes relevant in this scenario as well. Q. 3 Scott Sklar should have data on actual and forecasted demand of the item that were previously introduced as µnew item¶. It will give him an idea about various costs that go on in launching a µnew¶ item. He should have information on selling price which will be determined after consulting with the marketing, sales and production department of L. L Beans.

It will be useful for both the company and Scott. Information on cost of sales, commissions provided (if any) for sales, stock outs and backorders cost are relevant in his decision making. He can get sales information on a new catalogue item by comparing it with that of Bean¶s competitors¶ similar item and by consulting with the marketing department. It will help him understand the current market trends for that particular item. After analyzing the market, he can observe what level of buffer stock he should have to avoid stock outs.

For this, he will have to compare stock out costs and over-stocking costs for that item. To acquire this information, he needs to sit down with his buyers, sales and inventory personnel. Moreover, he should also know what promotion cost each µnew¶ item will incur in getting printed in one catalogue. How much space will it require to be noticed as a new item in a particular category? In other words, he should gather information on the promotion costs of the new item. He should also find the service level based off a profit margin calculation. These bases will help him forecast demand for the new item.

He should also observe what happens to the demand of existing products whenever new products are introduced in the market. If the new product is pulling customers away from the existing products then this factor should be incorporated in the demand forecast of the new catalogue item. Q4. One thing that Fasold need to realize is that we lose more when we have stock outs then when we liquidate our inventory. He can take a simple example that for example selling one item gives you 15rs worth of profit, and the liquidation cost for that item is 5rs.

So if you have inventory then you can earn 15rs as compared to the cost of stock out which is pretty less. So a proper cost benefit analysis could be done for Fasold. But here a proper cost benefit analysis should be made to give a true picture that also has the cost on inventory storing and not just the loss because of liquidation. Also he should be explained that forecasts are never accurate. Also this over stocking is more for the ? new items? , so L. L Bean should introduce them in the catalogues to get a fait bit of idea before actually introducing them.

We can also tell him that because of the large lead time of eight to twelve week we have to keep access inventories as well because if we receive an unexpected order the time to fulfill that order would be very long. But we also need to realize that improvement is needed and Faslod concerns are genuine in the sense that he should know that why is he keeping so much inventory. Also he should be told that we are improving the whole forecasting process to better predict demand so we have less buffer stock, but still as the cost of out-stocking is more the over stocking, we would be applying a more prudent approach.

Q5. There could be numerous improvements that L. L Bean can incorporate in its process. As we read in earlier cases as well that the fashion industry is a fast changing industry and product demands change very rapidly. So only using the past data to predict future is not enough. They should be continuously be updating their forecasts based on latest data as well. Secondly in is mentioned in the case that ? customer behavior is very hard to predict? , so what it can do is that it may involve experts in the whole process, with experts and the data combined they would be better able to forecast.

So a Delphi method is more appropriate here. Another important thing to note is that the new catalogue is introduced in August, which gives the customers little time to place a second order. L. L bean should try to introduce catalogues earlier so that the demand could be more easily assessed and a second commitment could be made. Also with the catalogues we can send questionaries¶ that could be optional and in this way we can gain better insights inside the consumer behavior and thus predict demand in a better way. This should especially be done for ? never outs?

and sent to regular customers and once we know how customers rank the new products in comparison to the new products we can adjust our plans. But a long term plan should be to reduce dependence on foreign vendors and have more local vendors who can supply quickly. Also we can convince suppliers to keep stocks of raw materials that are same for all like zips, buttons shoe laces etc. In this way the response will be faster and L. L bean can plan a second order too based on the new forecast. So the whole supply chain process needs to be improved along with forecasting accuracy improvements.