October 02, 2017

"Modern" Score

It comes up all the time in project work. There's a merchandising problem, and in the process of researching the problem, a question comes up.
  • "How do we get younger customers to buy this stuff?"
On average, the answer is "you don't get younger customers to buy stuff unless you sell stuff younger customers like, and you sell it the way younger customers want to shop." Which isn't a satisfying answer, because the person asking the question believes there is an answer to the question, an answer that is satisfying and easy to implement.

So we approach this problem from a different angle.

Here's what I did in recent research work.
  • I weighted historical purchases for each customer ... 100% for dollars in the past year, 60% for dollars spent 13-24 months ago, 35% for dollars spent 25-36 months ago, 20% for dollars spent 37-48 months ago, and 12% for dollars spent 49-60 months ago.
  • Dollars were added up by marketing channel (i.e. call center, email, search, online marketing, other online orders), then are divided by weighted historical dollars spent.
  • I multiply factors by each percentage within each marketing channel ... 100% for email, 75% for online, 50% for search, 50% for online marketing, 0% for call center. Then I multiply the resulting percentage by 100 ... giving me a score for each customer between 0 and 100.
  • I then gather three additional metrics ... 35 points are assigned to customers who click through 2+ email campaigns per year ... 1 point is assigned to each day with a website visit annually with a maximum of 35 points ... and 30 points are assigned to each customer who generates half of their visits or more on mobile devices. Then sum the points from each metric ... each customer earns a score between 0 and 100.
  • Take the purchasing score (0 to 100) and the visitation score (0 to 100) ... add them together ... then divide by 2.
The resulting score is what I call the "MODERN" score ... it tells us how "modern" the customer is.

But the metric isn't necessarily used to analyze customers.

Instead, the metric is used to analyze the merchandise customers purchase.

Example:  5 customers purchase this item. Here are their "modern" scores.
  • 0, 47, 55, 75, 100.
  • Modern Score for this Item = (0 + 47 + 55 + 75 + 100) / 5 = 55.4.
Example:  5 customers purchase a different item. Here are their "modern" scores.
  • 25, 35, 0, 50, 50.
  • Modern Score for this Item = (25 + 35 + 0 + 50 + 50) / 5 =  32.0.
The first item has a "Modern" score that is much greater than the "Modern" score for the second item.

In other words, the customers who buy the first item are more "Modern" than the customers who buy the second item.

Tomorrow, I'll explain why this is important.

Until then, perform your homework assignment ... go into your database and assign a "Modern" score to each customer ... and then for all items sold in the past year, calculate the "Modern" score average for that item. What do you observe?


P.S.:  For the portion of the audience who enjoy the thought of launching a catalog, this is for you (click here).


No comments:

Post a Comment

Note: Only a member of this blog may post a comment.

A Post-Tropical Cyclone

A tropical system doesn't stay a tropical system forever. Take Milton for instance. A Category 5 hurricane at two different periods earl...