Sports Unlimited Case Study

Case Synopsis

Sports Unlimited is a company founded in 1977 and operated more than 500 stores in 37 states by the year of 2011. This company owned not only national brands, but also carried private label which estimated at almost 20% of revenues. Since seasonality played an important role in affecting customers’ buying behavior for their product, optimizing markdown schedules were especially important.

The executive vice president of merchandising, Bob Thompson, was not satisfied with company revenue growth which he believed was related with their current markdown policy. Current markdown decisions are made at the item level by buyers with 50% initial markdown, each buyer has a certain markdown budget. In terms of the timing of markdown, buyers use price optimization (PO) tools, if items’ sell-through rate were lower than 20% by weekly ranking, they would be recommended for markdown. There were also point of sale(POS) discounts (60%) employed for some holiday sales or select customers.

In this study, we need to propose a new markdown strategy on magnitude and timing of the first markdown based on the 2011 historical data. Instead of investing some expensive optimization software, Thompson would like to keep their current PO tool and wish to produce an efficient and easy-to-follow solution for buyers, which could increase revenue by 20% - 25%.  

Data Analysis

The historical data from year of 2011 provided contains 4080 items, along with 14 variables that describe product features, sale and markdown performance.

To compare revenue performance for items with different ticket price, we developed a new standardized value “Revenue Ratio” ([pic 1][pic 2]It is to evaluate how much percentage of the maximum revenue is achieved by certain sale solution for each item.

[pic 3]

Based on investigations from pilot project, lifecycle is selected as influential factor to revenue, because their customers are young and sensitive to fashion trend. Old style apparel should not stay more than three weeks on the rack. Also, different products have different lifecycles, so brand is a key factor for revenue. Beside these two, we consider sell-through rate is critical in distinguishing popular items. Here, based on analysis, sell-through in the third week sale is a good indicator of total sell-through rate.

[pic 4]

After selecting brand, lifecycle length, and sell-through rate as our influential factors, and Revenue Ratio as our standardized revenue result, we fit a linear regression to discover if those three factors statistically influence the Revenue Ratio. From p-value below, we discovered only lifecycle length, and sell-through rate are significant factors in influencing revenue, which will be used to segment our products in the following markdown strategy.

[pic 5]

Figure 1

Description of the approach

Before clustering items, based on selected two significant factors, we visualize the distribution of lifecycle and sell-through rate, and make sure to categorize to three groups with nearly equal number of products. Therefore, we separate lifecycle into 3 intervals [7,12], (12,21), (21,47), respectively categorized into Short, Medium, and Long lifecycle. The sell-through rate is into [0, 0.075], (0.075,0.175), (0.175,1), respectively categorized into Low, Mid, High sell-through rate. Here is the population in each category by histogram below. [pic 6]