Trial-to-paid conversion rate: the full guide

What the rate measures, and what it hides

Divide the customers who paid by the trials they came from and you have the trial to paid rate. It is the number every trial business quotes, and it is easy to get wrong because three decisions hide inside the division: which trials count, which payments count and how long you wait. Change any of the three and the same business reports a different rate. The glossary entry gives the short definition; this guide is about the decisions behind it.

ChartMogul's survey asked what portion of leads or free signups turn into paying customers inside six months, which is a long and generous window for a trial that lasts two weeks. If you compare your fourteen-day rate to a six-month figure, you are comparing different things. Before you read any benchmark, write down its window and its denominator, and compare only like with like.

Pick the denominator before the number

There are three common denominators, and each answers a different question. The table uses the calculator's default cohort, with an illustrative number of trials still running to show the effect.

Using trials started keeps the rate honest because nothing is excluded after the fact, but a young cohort will look worse than it is, since some of its trials have not ended. Using trials ended fixes that and is the better choice for a recent cohort. The third row is a diagnostic rather than a headline: it separates people who never opened the product from people who tried it and said no, which need different fixes.

What a 200-product study found

The ChartMogul Conversion Report, published on 4 February 2026 with analysis by Kyle Poyar, surveyed 200 B2B software products in January 2026. Free trials were the primary way new customers start at 57% of products, more than twice the 26% that lead with freemium. The median free-to-paid rate across all products was 8%, but the report stresses that very few products sit near the median: there is a 10x gap between the top fifth and the bottom fifth of self-serve products.

Read the spread before the medians. A product at 3% and a product at 30% are both inside the study, and both can be healthy for their model and audience. The honest use of a benchmark is as a sanity check: if your rate is far outside the ranges for your type of trial, look for a measurement problem before you look for a product problem.

Why signups matter as much as conversion

The report layers signup rates under conversion, and the result changes the question. For every 1,000 website visitors, it shows how many create an account and how many become paying customers under each entry model. The last column below is computed here: paying customers divided by visitors.

Two readings follow. A trial that needs a card collects fewer signups than one that does not, 35 against 45 in the report, and still produces more paying customers because the share of signups that pay is so much higher. And the highest conversion rate is not the highest number of customers. Judge a trial change by paying customers per thousand visitors, which is the only measure that already contains both effects.

Where the rate leaks

A trial to paid rate is the product of several smaller rates, and the one that moved is rarely the one you assume. Sketch the path your own customers take and measure each step.

1. Visitor to trial start: does the page persuade, and is the form easy? See the visitor to trial rate.
2. Trial start to first session: do people who sign up actually open the product within a day or two?
3. First session to the moment the product works for them: how long, and how many do not reach it?
4. Moment of value to the end of the trial: do they come back, or does attention fade?
5. End of trial to a successful charge: does the card work, and does the buyer still want to pay? The failed payment rate captures the card part.

Only the last step is a payments problem, and only the first is a page problem. The middle three are product and onboarding problems. A team that blames the checkout for a rate that actually breaks at the second step will spend weeks on the wrong fix.

Reading cohorts without fooling yourself

Group trials by the week they started, and follow each group for the same number of days. A group is mature once every trial in it has ended and the charge has had time to succeed or fail, which for a trial of any length means waiting past the end date plus your retry period. Stripe documents how its retry schedule runs over days or weeks, so a cohort that ended yesterday is not yet final.

• Never compare a mature cohort with a young one. The young one will always look worse.
• Report a rate with its cohort size. A rate of 20% from ten trials means nothing next to a rate of 15% from four hundred.
• Split by acquisition channel when the sample allows. A blended rate can sit between a very good channel and a very poor one and describe neither.
• Keep a changelog of page, price and onboarding changes next to the weekly rates, so a movement has a possible cause written down at the time.

What a few points are worth

The trial conversion rate calculator turns a rate into customers and recurring revenue. At its defaults, 60 paid out of 400 started is 15.0%. Adding 5 points would bring 20 more customers and $980 of extra MRR a month, and adding 10 points would bring 40 more customers and $1,960. Percentage points are not percent growth: the same points add more customers on a smaller base, which is why the calculator also shows growth in paying customers.

Put the answer next to what a customer is worth over time with the lifetime value calculator. A rate improvement that brings customers who churn quickly is worth less than the same improvement that brings customers who stay, and the lifetime value figure is where that shows.

Levers, by the stage they act on

This table lists places to look, not promises. Neither the sources above nor this guide promise that any single lever lifts the rate by a set amount. Run one change at a time, keep the cohort comparison clean and let your own numbers decide.

Mistakes that make the rate wrong

• Counting a plan change or a reactivation as a new conversion.
• Dropping trials that were refunded or whose first charge failed from the paid count without saying so.
• Mixing monthly and annual buyers in one rate when they pay at different moments.
• Comparing your trial rate to a freemium rate. They start from different pools of people.
• Quoting a benchmark without its window, its denominator and its read date.

A ninety-day plan

1. Weeks 1 to 2: define the rate in writing with its denominator and window, then rebuild the last twelve weeks of cohorts to match.
2. Weeks 3 to 4: add the stage rates from the leak list so you can see which step is weakest.
3. Weeks 5 to 8: change one thing at the weakest stage and compare against the previous cohorts, not against the middle of last quarter.
4. Weeks 9 to 12: check that paying customers per thousand visitors moved, not only the conversion rate, and keep or revert the change.

Sources

• ChartMogul: The Conversion Report, read 19 Oct 2026
• Stripe: Configure free trials on subscriptions, read 19 Oct 2026
• Stripe: Revenue recovery, read 19 Oct 2026