The "loyalty problem" that was a measurement bug

The "loyalty problem" that was a measurement bug

A content site panicked over a returning-visitor rate of just 11%. Leadership read it as a retention crisis and greenlit a costly membership push. We checked the cookie story first.

The clue: the site had aggressive cookie consent and a 7-day cookie window. Returning visitors who cleared cookies, switched devices, or declined consent were counted brand-new every time. The 11% measured cookie persistence, not loyalty.

Cross-referencing logged-in users told the truth — actual return rate was near 34%.

We extended the cookie window where consent allowed and added lightweight logged-in tracking. Measured returning visitors rose from 11% to 27% with zero change in real behavior, and the panic membership project was shelved.

—
Lesson: new-vs-returning measures cookies, not people. Before you diagnose a loyalty problem, check whether your measurement can even see the loyalty.