Retention benchmarks: what a good repeat rate looks like
There is no single 'good' repeat rate. It swings wildly by category, so a number that means thriving for a coffee brand can mean dying for a mattress brand.
The honest benchmark is your own history. Feed AI the raw counts and it pulls the number into shape fast.
Fastest path: one prompt, end to end
🤖 AI prompt — paste into ChatGPT / Claude
You are a retention analyst. Use MY numbers only (invent nothing).
Customers who bought in the last 90 days: [#]
Of those, how many placed a 2nd order within 90 days: [#]
My category: [e.g. consumable / apparel / durable]
Last 3 quarters' 90-day repeat rate if I have it: [%, %, %]
Do this:
1. Compute my 90-day repeat rate (2nd-order customers / total).
2. Tell me the TYPICAL range for my category (consumables high, apparel mid, durables low) and cite the source; if you can't browse, say so and give a directional band, not a fake exact figure.
3. Compare my rate to my OWN trend, not just the generic band, and say if I'm improving.
4. Give one action if I'm below my own trend.
If a number is missing, ask; do not guess.
Output: my rate + category band (with caveat) + my-trend read + one action.
Or do it in 4 steps
- Pick 90-day repeat rate as your working metric. Of customers who bought, what share came back within 90 days. It reads fast and stays comparable month to month, unlike lifetime rates that only look good because they never close.
- Benchmark against yourself first. Your last three quarters beat any generic number, because they hold your category, price, and audience constant. A rising line is the win; a falling line is the alarm.
- Frame it category-relative. Consumables like coffee and supplements can see 40% or more; apparel sits lower; big durables may be single digits and that is fine. Judge against the right shelf, not a blog headline.
- Get the measurement math right. Cohorts and lifetime value are where repeat rate turns into money; see ltv-and-cohorts under money & metrics for how to slice by acquisition month so mixed cohorts don't fool you.
Worked example (labeled): Say 1,000 customers bought in the window and 220 reordered within 90 days. That is a 22% repeat rate.
If your prior three quarters read 15%, 18%, 20%, you are on a healthy climb, even if a generic 'good is 30%' post would call you behind. The trend is the signal. Plug in your own counts.
Benchmark against your own last three quarters first; the category band is context, your trend is the verdict.