Here is a story about a made-up brand called Pellwood Greens.

In June, Pellwood cuts its advertising in half to save cash. In July the monthly report has wonderful news. The repeat purchase rate jumped from 31% to 38%. Retention is working. Someone suggests a team lunch.

Nothing got better.

The repeat rate is the share of this month's orders that came from returning customers. Pellwood cut ads, so there were fewer new customers. Fewer new customers means returning customers are a bigger slice of the pie. The slice grew because the pie shrank.

It gets worse. The month Pellwood has its best advertising month ever, the repeat rate will drop, and someone will ask what is wrong with retention.

A number that goes up when things get worse and down when things get better is not a metric. It is a mood generator.

Think of a school. To know if the school is getting better at teaching, you do not average every student in the building. You take the class that started in 2024, see how they did after one year, and compare them to the class that started in 2025 after their first year.

Do the same with customers. Take everyone who first bought in January. How much had each of them spent, on average, after 30 days? After 60? After 90? Now ask the same about February's customers. If February is ahead of where January was at the same age, you are getting better. If not, you are not, no matter what the repeat rate says.

That is called a cohort, and it is the only honest way I know to measure this work.

Paste in your orders and it builds the table and the chart for you. Pellwood's numbers are filled in as the example so you can see how it works before you add your own.

Get the Cohort Tracker

Find the one number your company uses to judge retention. Ask: would this number go up if we simply stopped advertising? If the answer is yes, stop reporting it.

See you next month.

Jason