Trial to paid conversion dropped. Find the cause.

When as many trials start but fewer pay, the trial itself changed, or the people starting it did. Find out which, then find the step where they stopped.

Open the live demo Start free

Week by week, Jun to SepTrials and what they became

marked

Trials started68 → 68 a week

Trials that paid35% → 18%

New customers27 → 14 a week

2 Sep · trials kept coming, fewer paidthe week a new onboarding went out Onboarding step 2 where most stoppedPaywall −41% reached it

Drawn with made-up numbers.

What to check, in this order

1Count it over trials that ended, not trials that started: a trial begun yesterday has not had its chance to pay.
2Check who is starting trials: the same sources as before, or a new one that brings people who were never going to pay?
3Find the week it began and what changed that week: onboarding, the paywall, the price, the trial length, a card now asked for or no longer asked for.
4Follow trial users through the steps that lead to paying, and see where fewer of them get through than before.

Where Sumly shows it

Trial conversion over the trials that ended in the last 30 days; trials running on the Overview, trials started on Metrics.
Sources and campaigns of the people who started trials, and of those who paid.
Funnels from the first page or screen to the payment, so the step where people stop shows.

Two reasons conversion falls

Either the trial got worse for the same people, or different people started it. A new source, a launch or a viral post can fill the trial with curious visitors and halve the share that pays without anything being wrong with the product.

So look at conversion by source before you change the onboarding. If the old sources still convert as before, the product is fine.

Mistakes that make it lie

Reading it on trials started this week. Wait until they end.
Comparing a trial with a card to one without. Asking for a card changes who starts, so the rates are not comparable.
Changing three things at once to fix it. Change one, and give it a full trial length to show.

Trials dropping questions

What is a normal trial to paid conversion rate?

It depends on whether you ask for a card. With a card, many paid products convert a large share of trials; without one, a much smaller share of many more trials. Your own rate last quarter is the fairer benchmark.

Where do trials in the app stores show?

Through RevenueCat, and from the App Store read directly, as trials that start paying when they end. Google Play's reports only show a trial once it is paid.

How soon will I see if a fix worked?

After one full trial length, for the trials that started after the fix.

Try it with your own projects

Sign in with Google or a code by e-mail and connect your first project. Free for one project, and 14 days of a paid plan with no card.

Start free

Or look around first: open the live demo.