January 15, 2018

80 / 50 Rule

Here's something that has changed a bit over time.

Step 1: Build a model ... independent variables = 0-12 month demand, 13-24 month demand, 25-36 month demand, 37-48 month demand, 49-60 month demand, 61+ month demand. Those variables are created as of a year ago. Then select only 12-month buyers. The dependent variable is demand spent in the past year. Create a simple regression equation.

Step 2: Score the file using the equation.

Step 3: Rank scores into 1%-tiles or 2%-tiles or 5%-tiles ... your preference.

Step 4: Sum cumulative annual demand across %-tiles. Document the point at which 50% of annual demand is accumulated.


If you learn that 80% of annual demand happens at the 50% mark or later, you generally have a productive customer file.

If you learn that 80% of annual demand from the 12-month buyer file comes from fewer than 50% of your customer file, then the profit generated by your business comes from a disproportionately small portion of your buyer file.

I just ran this analysis on an e-commerce business ... 80% of demand came from 40% of 12-month buyers ... and 33% of demand came from 4% of 12-month buyers. At this point, you have zealots and a bunch of customers producing minimal value.

Over the past decade, our businesses shifted ... aligning more with zealots. The meaty "middle" of the customer file was ceded to Amazon. Expect more of this going forward.

Memorial Day Weekend

There's the obvious reason why we "celebrate" Memorial Day. And then there's the secondary outcome ... an unofficial ...