Loss framing on a basket reminder
Hypothesis
Framing the reminder around what is about to be lost — a reserved item — will outperform a neutral reminder that the basket exists.
ABANDONED BASKET EMAIL
- Effect
- +0.4%
- 95% CI
- -1.9% … +2.7%
- Sample
- 22,100
- Duration
- 21 days
- Confidence
- high
This is the kind of test that gets reported as a win when nobody writes down the confidence interval.
The point estimate was positive. The interval comfortably contains zero, at a sample size well past what the test was powered for. The honest reading is that loss framing did nothing here, and a fourteen-day extension did not rescue it.
Worth recording precisely because it is the sort of result teams quietly drop. The next time someone proposes loss framing on a reminder for this client, this entry is what they will find.
Published August 2, 2026. Anonymized. Sector and surface are retained; the client is not.