benchmarks
How Fast a Saved Cart Actually Converts
We measured the time between a prevention offer appearing and the order landing. Median 4.7 minutes, 94.8% within 30 minutes, from 290 matched orders.
ChatGPT
Perplexity
Gemini
ClaudeMedian time from a prevention offer appearing to the order being placed: 4.7 minutes. That's across 290 matched orders that carried our discount code, measured 2026-08-25.
Almost nobody publishes this number, and there's a simple reason: almost nobody can measure it. An on-site intervention and the order it produces happen in the same session, often the same browser tab. If you're not built to watch both ends of that, the gap between them is invisible to you.
The distribution, not just the median
The median undersells how tight this actually is.
| Percentile | Time to order |
|---|---|
| Median (p50) | 4.7 minutes |
| p75 | 9.9 minutes |
| p90 | 20 minutes |
| Within 30 minutes | 94.8% |
| After 24 hours | 1.0% |
n = 290 matched orders, measured 2026-08-25.
Nine out of ten of these orders are placed within twenty minutes. Almost all of them are placed within half an hour. The tail past a day is close to nothing — 1.0%.
That shape makes sense once you think about what's being measured. The shopper is already on the site, already has the item in their cart, and has already decided to leave. An offer that changes their mind changes it right there, in the same window they were about to close.
Compare that with how email works
A recovery email has to clear a different set of hurdles before it can convert anyone.
It has to be sent — usually after a delay, so it doesn't feel instant or robotic. It has to land in an inbox instead of a spam folder. It has to be opened, which depends on subject line, time of day, and whether the shopper still remembers what was in the cart. Then it has to be clicked, and the shopper has to come back to a device, open a browser, and complete the purchase — possibly hours or days after the moment they were actually deciding.
Every one of those steps is a place the shopper can fail to convert, and every one of them adds time. None of that makes email worthless — it reaches shoppers who already left and have no other path back. It just means the two approaches are answering different questions. Email is measuring how many people you can bring back later. This is measuring how many people never actually needed to leave.
Why this matters if you're deciding what to measure
If you only track abandonment rate and email recovery rate, you're missing the fastest-moving part of the funnel. A shopper who converts in under five minutes never shows up in an email report at all, because they were never abandoned long enough for a recovery message to reach them.
That's the case for watching time-to-conversion as its own number, not just a footnote under revenue. It tells you whether an intervention is catching shoppers at the moment of highest intent — while they're still looking at the page, still holding the decision in their head — or whether it's just another message competing for attention after the moment has passed.