Know Who Buys Again

What it is

Kevin Hillstrom runs MineThatData, a consultancy, and has written its blog since the 2000s. His first book, Hillstrom’s Database Marketing: A Master’s Complete Method for Success, came out in 2006, followed by books on multichannel and merchandise forensics.

The theme across all of it is that the customer file is the business. Read it closely enough and it tells you who buys again, who’s drifting, and whether your marketing did anything at all.

What you’ll take from it

First, define value by what customers do next. He defines a loyal buyer with a probability instead of a feeling:

“I define a loyal buyer as somebody who has a 60% or greater chance of buying again in the next year.” MineThatData blog, October 20, 2026

Second, test honestly. In 2008 he published the data from an email test on 64,000 customers who’d bought in the past year. They were split at random into thirds. One group got a men’s email, one got a women’s email and one got nothing, and everyone was tracked for two weeks. Comparing each email group with the group that got nothing shows what the email actually changed, rather than who was going to buy anyway. The dataset is now a standard example for teaching uplift modeling. On testing in general, he put it this way recently:

“A/B tests are humbling. They force you to address your personal marketing belief system.” MineThatData blog, October 21, 2026

Where it shows up in your numbers

There are two habits here worth using every week. Judge customers by what they’ll be worth, and that starts with whether they buy again. And judge your marketing against a holdout, so a campaign only gets credit for the buyers it actually moved.