You'll have to click on the image here to see it ... lots of data there.
Here's an example ... this company has thirteen Merchandise Categories. In this table, we evaluate customers with three life-to-date purchases and at least one purchase in the past year. We assign Weighted Product Preferences to each customer, based on historical behavior. Then, we measure where customers spent money in the next twelve months.
This portion of the table is for customers who have a primary weighted product preference of category 01. The next column lists the secondary weighted product preference category. Obviously, the biggest cell count happens in 1/99 ... these are customers who only purchase from category 01.
Read across the columns ... the columns represent how much customers spent in each product category in the next year. Read across the 1/7 row ... these customers like product category 01 most, then product category 07. Which product categories, in the next year, have a value above $8.00?
- Category 01 ($51.44).
- Category 02 ($11.18).
- Category 05 ($12.24).
- Category 07 ($11.50).
- Category 12 ($10.22).
For customers who like category 01 most and category 07 second most, the five categories that warrant subsequent marketing activity are categories 01 / 02 / 05 / 07 / 12. That's where you focus your personalized merchandise assortment ... on those five categories. Those five categories generate more than 80% of future spend.
That's how you use Weighted Product Preferences to determine a Personalized Merchandise Assortment. This stuff doesn't have to be complicated. Just do something!!