A Category Making Progress

Over time I've changed how I evaluate winning items.

In the past year, I've analyzed items sold on a daily basis. Each day is analyzed separately. If a bad item is featured in an email campaign, the item may well be a winner for a day or two. May as well credit the item as a winning item for that day or two, correct?

I rank-order all items during the course of a day, then cumm total sales. If the best selling item represents 10% of sales that day, that item is assigned a value of 0.10 for that day. If a lousy item sold 1% of total sales and was in the 74% percentile, the lousy item is assigned a value of 0.74.

Each item is coded separately, each day.

There's a thousand things I can do with this data.

For one category, I analyzed the percentage of sales on an annual basis for top 20% of items, 21% to 40% of items, 41% to 60% of items, 61% to 80% of items, and 81% to 100% of items. Here's the table.

This category doesn't have a lot of best sellers - if it did, 20% or more of annual sales would be from items in the top 20%.

However, this category is changing. Look at the 41% to 60% column - average selling items. The fraction of sales have increased each of the past three years ... from 17.2% to 18.8% to 20.7% to 22.0%. Same thing with items in the 61st to 80th percentile. Meanwhile, items selling in the bottom 20% of the sales distribution used to comprise 40% of sales (three years ago) but only comprise 34.2% today.

The merchant in charge of this category is doing something right - the results are not dramatic, but there is clear improvement here. Good job!