Stripe Billing Benchmarking: How Peer Churn Comparisons Work, and Where They Mislead You
Stripe now ranks your churn rate against peer businesses inside Billing analytics. Here's how the matching model works, and its blind spot.
Stripe's Billing analytics dashboard now tells you where your churn rate sits against similar businesses, automatically, without you uploading a survey response or paying for a benchmark report. That's a genuinely useful thing to have next to your own numbers. It's also a comparison built on a peer group with a specific, structural blind spot — one that has nothing to do with how good the matching model is and everything to do with who's allowed into the room.
How Stripe actually builds your peer group
The matching runs on a k-nearest neighbors model, which is a fairly plain way of saying Stripe plots every eligible business in a shared feature space and finds the ones sitting closest to yours. The inputs are your Merchant Category Code, an industry classification Stripe generates by reading your public website, your annual recurring revenue, and your average revenue per user. None of those four inputs know anything about your pricing model, your customer segment, or how your churn splits between someone hitting cancel and a card just failing to charge. Two businesses can land in the same peer cluster with completely different retention mechanics, as long as their MCC, inferred industry, ARR, and ARPU happen to line up.
Once your peer group is assembled, Stripe shows the top, median, and bottom trends across that group and where you sit by percentile — Stripe's own documentation gives the example of landing in the 37th percentile against 200 similar businesses. The metrics on offer are MRR growth, subscription failure rate, and churn rate, all pulled straight from live transaction data rather than a self-reported survey field, which is the feature's real advantage over a static industry report.
The eligibility bar that shapes who you're compared against
Getting a benchmark shown to you requires a peer group of at least 100 similar businesses. Getting counted as one of those 100, for someone else's benchmark, requires three things: at least 100 active subscriptions, more than 5 active subscriptions sustained for over a year, and positive annual recurring revenue. Read that eligibility bar carefully and a pattern shows up immediately — every one of those three conditions selects for businesses that have already survived long enough to look stable.
| Requirement | What it filters for | Who it filters out |
|---|---|---|
| 100+ active subscriptions | Businesses with real transaction volume | Pre-revenue, early-stage, and very small SaaS |
| 5+ active subs sustained 1+ year | Retention that has already held for a full year | Anyone still inside their first 12 months, including businesses that already churned hard and shrank below the line |
| Positive ARR | Revenue-positive businesses | Businesses running at a loss, including ones burning cash on purpose to grow |
Source: Stripe Billing analytics documentation, peer group eligibility criteria.
None of this makes the feature wrong. It makes it a survivorship-filtered comparison, which is a different thing than an unbiased snapshot of "similar businesses." A one-year-old SaaS company that lost half its subscriber base to bad early retention and shrank below the 5-subscription floor doesn't get counted in anyone's peer group. The businesses that do get counted are, by construction, the ones whose churn was survivable. If your percentile comes back looking fine relative to peers who already cleared that bar, it tells you less about whether you're actually healthy than it looks like it does — you're being compared to the businesses that made it, not the full population of businesses that started where you did.
What a single churn percentile can't show you
The benchmark surfaces one churn rate number and ranks it. It doesn't split that number by voluntary and involuntary causes, and none of the four matching inputs — MCC, inferred industry, ARR, ARPU — carry any information about that split either. We've written before about how involuntary churn from failed payments typically accounts for 20–40% of total cancellations at a subscription business, and that ratio varies enormously by payment method mix, geography, and dunning maturity — none of which the peer-matching model sees. Two businesses can land in the same percentile with entirely different problems: one losing subscribers because its product isn't sticky, the other losing the same percentage to expired cards and a dunning sequence nobody's touched in a year. The benchmark treats them identically.
It also doesn't segment by cohort age, which is where the highest-signal churn information usually lives. A business with strong month-12+ retention and terrible first-90-day churn can average out to a percentile that looks completely unremarkable, while sitting on an onboarding problem that a cohort retention curve would expose immediately. A single blended number is exactly the kind of metric that hides two offsetting problems behind one reassuring digit.
Stripe's benchmark vs. a third-party benchmark report
| Stripe Billing benchmarking | Third-party benchmark report | |
|---|---|---|
| Data source | Live transaction data from Stripe accounts | Self-reported survey data or aggregated customer data from a reporting vendor |
| Peer group | k-NN match on MCC, inferred industry, ARR, ARPU — 100+ business minimum | Fixed segment buckets (e.g. by ARR band, price point, or vertical) defined by the report author |
| Peer group bias | Skews toward survivors — a year of sustained subscriptions and positive ARR required | Skews toward whoever responded to the survey or opted into the reporting tool |
| Refresh cadence | Continuous, tied to your live Stripe data | Published periodically, typically annually or per report cycle |
| Segmentation you control | None — Stripe picks the peer group for you | You choose which published segment to compare against |
| Cost | Included with Stripe Billing analytics | Often free for summary data, paid for full reports |
Neither column is strictly better. Stripe's version is faster, live, and free, and it's measuring what actually happened in your Stripe account instead of what someone typed into a survey field months ago. A published report, on the other hand, lets you pick the exact segment definition that matches your business — something the automated matcher can't do, because it has no idea whether you're a $50/month prosumer tool or a $500/month SMB product that happens to share an MCC with one.
Here's roughly where third-party segment benchmarks put "good" monthly churn, for context against whatever percentile Stripe hands you:
Source: Baremetrics, ChartMogul, OpenView Partners, and ProfitWell SaaS benchmark reports, compiled in our churn rate benchmarks guide.
If Stripe tells you you're in the 60th percentile of your peer group, that's a meaningfully different signal depending on which of those rows you actually belong to — and Stripe's matching model has no way to tell you which one that is.
How to actually use the number
Treat the Stripe percentile as a fast directional check, not a verdict, and run three things alongside it before deciding anything based on it:
- Split it into voluntary and involuntary before reacting. If your blended churn rate looks fine but you haven't checked what fraction is failed payments versus actual cancellations, you don't know whether you have a retention problem, a dunning problem, or a payment-mix problem. Pull that split before you touch anything.
- Check it against a cohort curve, not just the headline number. A benchmark-competitive average can still be masking a first-90-day churn spike. Build the cohort retention curve and look at where the drop-off actually happens.
- Sanity-check your price point against the right published segment. A prosumer tool being compared, by MCC coincidence, against SMB peers on Stripe's model will look artificially bad next to a target that never applied to it in the first place. Our churn calculator is a fast way to get your own number in the same terms before you compare it anywhere.
None of this makes the feature worth ignoring — a live, zero-effort percentile is genuinely more useful than nothing, and most SaaS teams have never benchmarked their churn against anyone. The failure mode is treating one number, generated by a model that has never seen your cancel reasons, as a finished diagnosis. It's a starting point for the same question a cancellation flow answers directly: not whether your churn number looks acceptable next to a peer group Stripe assembled for you, but exactly why the subscribers behind that number are leaving, and whether anything could have kept them.
Frequently asked questions
What is Stripe's Billing analytics benchmarking feature?+
It's a feature inside the Stripe Billing analytics dashboard that automatically compares your subscription metrics — including churn rate, MRR growth, and subscription failure rate — against a peer group of similar businesses on Stripe, showing where you rank by percentile.
How does Stripe pick which businesses to compare me against?+
Stripe uses a k-nearest neighbors (k-NN) model that groups businesses by Merchant Category Code, an AI-generated industry classification pulled from your public website, annual recurring revenue, and average revenue per user. It does not use your churn composition, pricing model, or customer segment as matching inputs.
Why doesn't Stripe show me a benchmark?+
Stripe requires a peer group of at least 100 similar businesses before it will display a percentile. To be counted in someone else's peer group, a business needs at least 100 active subscriptions, more than 5 active subscriptions sustained for over a year, and positive ARR. Younger, smaller, or thinner-margin businesses often fall on both sides of that gap — too small to be shown a benchmark, and structurally excluded from being one.
Should I trust Stripe's benchmark over third-party SaaS benchmark reports?+
Use them together, not as a substitute for each other. Stripe's benchmark draws on live, real transaction data instead of self-reported survey responses, which is a real advantage. But its peer pool is built only from businesses stable enough to clear a one-year, 100-subscription bar, which skews the comparison toward survivors. A third-party report segmented by ARR stage and price point, cross-checked against your own cohort data, still catches problems Stripe's single percentile won't.
Stop losing subscribers today
One script tag. One function call. A live cancellation flow in under 10 minutes.
Start free trial →