Trial Conversion: What It Measures and How to Calculate It Honestly

Trial conversion is the proportion of product trials that become paying customers within a defined period of time. It is reported as a percentage of a cohort of trials, not of all signups ever, and it is only interpretable when the denominator and the conversion window are stated alongside it.

That last clause is not pedantry. More arguments about this metric come from undeclared definitions than from genuine disagreement about the underlying business. Two honest analysts can produce two different correct numbers for the same month, and the gap between them can be wide enough to change a decision.

Trial conversion rate = (converted trials ÷ trials in the cohort) × 100.

Numerator: converted trials

The count of trials that produced a paid subscription within the stated window. Decide in advance whether a paid subscription means a successful first payment, an accepted order, or a signed contract, and whether a discounted or internal account counts. Refunds and immediate cancellations should be removed, otherwise the metric counts money the business never kept.

Denominator: trials in the cohort

Here there are two accepted definitions, and both are in common use. Measured on starts, the denominator is every trial that began during the period. Measured on completions, it is only the trials that ran to the end of the trial period. The starts version answers "how good is the whole funnel". The completions version answers "how good is the closing conversation with people who stuck around". Use whichever you like; say which one you used.

The window

The period after the trial starts within which a payment still counts, for example thirty days from trial start or fourteen days after trial expiry. Fix the window, and do not report a cohort until it has fully matured, or recent months will always look worse than older ones for purely arithmetic reasons.

The unit

Accounts or users. In a team product, five people from one company starting five trials is one commercial opportunity. Counting the users rather than the account depresses conversion and hides what is really happening.

The numbers below are invented to show the arithmetic. They are not benchmarks and should not be treated as targets.

A business runs a fourteen-day trial with no card required. In March, 400 accounts start a trial. By the final day of their trial period, 260 of those accounts are still active. Within thirty days of their own start date, 52 accounts have made a first payment, and one of those 52 refunds in week two.

Numerator: 52 minus the refund, so 51 converted trials. Measured on starts, the denominator is 400 and the rate is 51 ÷ 400 = 12.75 percent. Measured on completions, the denominator is 260 and the rate is 51 ÷ 260 = 19.6 percent. The worked example above uses trial starts as the headline denominator, and reports the completions figure beside it so the drop-off before expiry stays visible rather than being averaged away.

Trial conversion drives one decision above all others: where to spend the next unit of effort. A low rate measured on starts with a healthy rate on completions says the problem is early — people are abandoning before they ever reach a decision, which points at onboarding, data import, and time to first value. A rate that is weak on both says the problem is at the end: the value is not obvious, the objection is not being answered, or nobody is having the conversation.

It also sets the economics of acquisition. The amount you can afford to spend attracting a trial depends directly on what share of trials pay and what they are worth. Teams that spend on traffic without knowing this number are guessing, and usually guessing generously.

Attributing payments to the month they arrive

Counting April payments against April trial starts mixes two cohorts. In a growing month it flatters the rate, in a shrinking month it punishes it, and in both it makes the trend unreadable. Payments belong to the cohort of the trial that produced them.

Comparing across trial models

A card-required trial and a no-card trial are different products from a measurement point of view. So are a seven-day and a thirty-day trial, and a self-serve trial and a sales-assisted pilot. Blending them produces a single number that describes none of them.

Optimising the percentage instead of the outcome

Requiring a card, shortening the trial, or gating the trial behind a demand form will all raise the percentage, because they shrink the denominator to people who were already close to buying. Fewer customers can arrive alongside a better-looking chart. Always read the rate next to the absolute count of new customers and the retention of the cohort that converted.

Ignoring the source mix

A quarter where paid traffic doubles will usually show a falling conversion rate even though nothing about the product changed. Segment by source before concluding anything about product quality.

There is no universal good number, and any page that gives you one is guessing. Rates vary enormously by price point, trial model, audience, and whether a human is involved. What is reliable is the qualitative shape.

A healthy trial motion looks like this: most trials reach a real activation event in the first days rather than in the last hours; conversion is stable or improving while trial volume grows; the rates by source are known and different from one another; and customers who converted from a trial retain at least as well as customers who arrived another way. A trial motion in trouble looks like a large gap between starts and completions, conversions clustered entirely in the final forty-eight hours, a headline rate that only improves when the funnel gets narrower, and a cohort of trial-sourced customers that churns faster than the rest of the base.

Most small teams do not need a warehouse to run this well. They need trials to exist as records with an owner, a start date, an expiry date, and a source, so that follow-up can be scheduled and the report can be filtered. HelloGrowthCRM treats a trial as a pipeline stage with exactly those fields, adds sequences and reminders against them, and there is a free plan available if you want to run a first cohort before committing to anything.

Continue with what a CRM is, lead management software, sales automation, AI CRM capabilities, the free CRM plan, and CRM for small business.