'Proplenishment' makes a payoff

Bobbin, May, 1999 by Russell E. King, Robert P. Maddalena, Kara Moon, Mahmut Ali Gokce

A POST-SEASON ANALYSIS OF A PILOT "PROPLENISHMENT" PROGRAM ILLUSTRATES HOW A MAJOR RETAILER AND CHILDREN'S WEAR MANUFACTURER MINIMIZED INVENTORY PROBLEMS AND LOSSES DURING A SLOWER-THAN-EXPECTED SALES PERIOD.

This time last year, Bobbin reported on the pre-season projected advantage of a pilot replenishment program for infants' apparel sold at Dillard's, and supplied by The Warren Featherbone Co. (See "Replenishment Rules," Bobbin, May'98.)

This year's article tells the rest of the story - a post-season analysis of how well the program actually worked. As it turned out, sales were higher than the prior year but below what had been forecasted, yet both the retailer and manufacturer were rewarded with fewer losses and smaller inventory excesses than they would have had under a more traditional retailer-vendor relationship.

For a little background: Research of the pilot replenishment program between Dillard's and Warren Featherbone has been funded in part by a grant from the National Textile Center and the Sandia National Laboratories under the AMTEX-DAMA Project. The replenishment program has been built and analyzed with the aid of "The Sourcing Simulator" software tool, also known as ARMS, which can simulate different product sourcing and allocation scenarios and calculate estimates of return on investment, sell-through, losses due to markdowns, etc. The tool also can measure and compare projected outcomes against real outcomes. (See "Quick Response Beats Importing in Retail Sourcing Analysis," Bobbin, March '97.)

The Case Study

The study considered eight styles of infants' dresses and rompers, offered in three or four sizes, and sold regionally in Texas stores. The merchandise plan called for sales of about 1,000 units in each style over a selling season of about 40 weeks, beginning with an introduction of the styles in mid-December. (Because this type of merchandise typically has been launched in January, the projected seasonality of sales, or percentage of sales expected each week, had to be reestimated.)

Under the traditional retailer-manufacturer relationship between Dillard's and Warren Featherbone, the entire plan would have been delivered prior to the start of the season, with no option for in-season replenishment. For the pilot program, however, the firms opted to implement a replenishment strategy that called for 50 percent of the buyer's pre-season plan to be manufactured and delivered prior to the start of the season, with the rest delivered in a single reorder with the SKU mix and quantities determined based on in-season point-of-sale (POS) data from the stores.

The firms termed this strategy a "proplenishment" plan, realizing that there are virtually unlimited variations of replenishment strategies. For instance, the retailer and manufacturer can agree that only a small percentage of an order be delivered up front, with future demand fulfilled by numerous reorder shipments throughout the season.

Proplenishment Performance

While demand for the infants' wear program came in lower than expected, the assortment (SKU) mix forecast error came in at approximately the predicted level, which is consistent with most retailers' performance in this area.

Estimates of the sell-through percentages for the styles (see Table 1) were significantly higher than the average for the infants' wear department - suggesting that the in-season reorder helped the movement of the merchandise. (Still, we believe these sell-through levels can be improved further, and will offer several improvement methods.)

The study also examined improvement in the gross margin percentage attained for each style under the proplenishment program, and compared this percentage with what would have been achieved had a traditional buying approach been used (i.e., assuming the full order was delivered up front and comparing this with what actually sold). The bottom bar of Figure 1 shows the level of gross margin improvement achieved with the proplenishment program over traditional methods.

Room for Improvement

While the proplenishment strategy lessened the blow of a poor selling season, the manufacturer and retailer still were left with excess inventory. Three alternative strategies have been identified that could lead to less commitment of inventory without introducing lost sales. These strategies are suggested as "Improvements" as follows:

Improvement No. 1: Improve the accuracy of the forecasted seasonality of sales (i.e., the estimated percentage of sales/week). The demand reestimation algorithm of the proplenishment program depends upon this seasonality estimate. In the pilot program, the seasonality of sales projection was not very accurate, in part because the lines were introduced earlier than usual. This scenario limited the ability to use historical sales data as a reference, and in the end, the number of goods ordered exceeded what was needed.

TABLE 1

Sell-Through Percentage Improvement

Proplenishment Program vs. Traditional Program

                      Percent of Improvement in Sell-Through
Style Number               with Proplenishment Program

1                                    34.3%
2                                    11.9%
3                                    36.0%
4                                    28.2%
5                                    24.6%
6                                    30.4%
7                                    28.7%
8                                    18.4%