Background
LL Bean is a retail company that specializes in outdoor and indoor apparel for men and women, footwear, and outdoor equipment for camping. A major strategic tool adopted by L.L. Bean is the use of its directive marketing approach. The directive marketing approach helps bring about better customer satisfaction amongst its competitors, which include Land’s End, Eddie Bauer, Talbot’s and Orvis. Their approach consists of mailing seasonal catalogues to L.L. Bean’s customers. Forecasting and stocking decisions is then based on the catalogues that L.L. Bean manufactures and sends to their brand consumers.
Product Line
LL Beans product line is staged hierarchically. At the top of the hierarchy are the merchandise groups: men and women’s accessories men and women’s apparel, men and women’s footwear, camping equipment, etc. Within each group was a demand center; for example women apparel had demand centers that consisted of knits, sweaters, jackets pants, etc. Afterwards, demand centers were broken down to item sequencing. Lastly, item sequencing was further broken down into individual items that were distinguished by color. The items were then presented in catalogs.
Forecasting
The major L.L Bean catalogs are Spring, Summer, Fall, Christmas, and some specialty catalogs – Spring weekend, Fly Fishing, etc. LL Beans best customers received all catalogs, and as for other customers, LL Bean used past purchasing behavior to send specialty catalogs that were more tailored to the customer. The estimated gestation period for the catalog was nine months. In the beginning of the gestation period was the initial theme for the season. Following the initial conceptualization was layout and pagination. In the middle of the gestation period, item forecasting was constantly revised and then frozen. Towards the end of the gestation period product mangers would hand off their product line to the inventory managers. Lastly, at the end of the gestation period, the catalogs were being sent to customers.
Calculations for Forecasting
LL Bean uses two different calculations in order to determine inventory for new and never out items. The first calculation involves historical forecast errors expressed as A/F ratios being computed for each item in the previous year. Forecast errors are calculated for each item, and a frequency distribution is computed, which would be used as a probability distribution for unrealized future forecast errors. For example, if 50 % of forecast errors for new items were between 0.7 and 1.6 in the past year, it was assumed that with a probability of 0.5, the forecast for new items in the current year would be between 0.7 and 1.6 as well. Therefore if a particular item were 1,000 units, it was predicted that with a 50 percent probability, the demand of the new item would fall between 700 and 1,600 units.
The second calculation involved each item’s commitment quality determined by balancing its individual contribution margin against its liquidation cost. For example, if Bean sold an item that had a margin of $15 if sold at demand, and $5 sold at its liquidation value, the commitment value would be 0.75. Therefore if the 0.75 fractal of distribution of forecast errors was 1.3, and frozen forecast was 1,000, the 0.75 fractal would be 1,000 x 1.3, equating to 1,300. Hence, Bean would make an inventory commitment for 1,300 units.
Problem
L.L Bean must make its stocking decisions based on their catalogues that are placed every season. In several cases, orders have to be placed out several weeks before the catalogue reaches the hands of customer, and commitments can’t be altered afterwards. As a result, there is a wide dispersion in forecast errors for never out and new items led to unsold inventory and stock outs. Annual costs of lost sales and backorders were about $11 million, and costs associated with having wrong inventor was an additional 10 million, totaling $21 million. There is a major mismatch between LL Bean’s supply and demand, which also was a reflection in their sales.
Solutions
L.L. Bean bases one of their equations on past sales which can be a problem. Reason being is it is difficult to calculate consumer behavior on new sales based on past ones, especially in retail market. Consumer preferences and demands are constantly shifting; therefore basing future forecasting over past forecasting is not an accurate approach. Forecasts should be updated continuously throughout all data, including the latest data. Combining all data may lead to better forecasting.
Creating more local vendors can help cut down lead time. The typical production lead time for LL Bean’s domestic orders takes eight to twelve weeks. Having more local vendors would satisfy consumer demand when an order needs to be placed due to demand. Also, this would limit “one shot” commitments.
In order to determine accurate demand for items, LL Bean should compare sales from its catalogues to similar items being sold by its competitors. It will help give more understanding to market trends for certain items. After gathering data from analyzing marketing trends for certain products, it will help LL Bean detect which levels of buffer stock they should obtain in order to abstain from stock outs. Determining accurate demand also allows LL Bean to adjust pricing of certain products in their catalogue. This allows them to limit price reductions and cost liquidations, thus allowing for more potential revenue.
תגובות