Measuring trial conversion

A free trial is a promise: try it, and if it is right for you, stay. Whether that promise works shows up in one number, the share of people who start a trial and go on to become paying members. It sounds easy to measure, and most membership software will show you something labeled trial conversion. But the headline figure is often calculated in a way that flatters or confuses, and on its own it cannot tell you why trials fail.

Measured carefully, trial conversion becomes one of the most useful diagnostics you have, because the trial is the moment a stranger decides whether your membership fits their life.

Define a conversion before you count one

Write down exactly what counts, and keep it fixed so your figures stay comparable over time. A sound definition has three parts:

• The start. Count everyone who began a trial, including people who cancelled in the first hour.
• The finish line. A conversion is a first payment that actually succeeded, not a trial that simply ended without being cancelled. A declined card is a failed payment to chase, not a new member.
• The settling period. Leave out anyone who asked for a refund within a few days of the first charge. If you charge a card automatically when the trial ends, some people will convert only because they forgot to cancel.

Trial conversion rate = trials that led to a successful, unrefunded first payment ÷ trials started in the same period

Count by the week or month trials started

The most common mistake is dividing this month's new paying members by this month's trial starts. Many of this month's conversions began their trials last month, so the two numbers describe different people. When trial starts rise or fall sharply, the result can be badly wrong.

Instead, group trials by the week or month they began, and wait until every trial in the group has ended and the settling period has passed. Then count how many from that group converted. The figure arrives a little later, but it describes real people making a real decision.

A worked example: finding where trials stall

Take a hypothetical membership that prepares people for their amateur radio license exams, run by an instructor named Ingrid. It offers a 14-day free trial with no card required. The figures are invented to show the method.

In April, 300 people started a trial. Ingrid follows them through each stage:

1. Started a trial: 300.
2. Completed a first practice exam during the trial: 210. The other 90 signed up and barely looked around.
3. Chose a plan and paid by the end of the trial: 98.
4. Payment succeeded and was not refunded: 90. Five cards were declined and three people asked for their money back.
5. Still paying after the second billing date: 81.

Her trial conversion rate for April is 90 ÷ 300, or 30%. The stages reveal much more than that single figure. Of the 210 who took a practice exam, 84 converted, 40%. Of the 90 who never did, only 6 converted, under 7%. The biggest loss is not at the payment step at all. It happens in the first few days, when nearly a third of trial users never reach the one activity that shows what the membership does.

That points to a clear fix: get more trial users to that first practice exam, through the welcome screen, the first email and a reminder on day two. It is exactly the kind of problem a well-planned trial email sequence is built to solve.

Compare trial types on the number that matters

Trials that ask for a card upfront almost always show a higher conversion rate, because fewer and more committed people start them, and some convert by forgetting. So the trial conversion rate alone cannot tell you which approach is better. Compare paying members per 100 visitors to the trial page instead.

Suppose Ingrid tests a card-required trial on a later month. Illustrative results:

• No card: 5,000 visitors, 300 trials, 90 conversions. That is 1.8 paying members per 100 visitors.
• Card required: 5,000 visitors, 150 trials, 75 conversions. That is 1.5 paying members per 100 visitors.

The card-required trial converts at 50%, which looks far better than 30%, yet it produces fewer paying members. Check the second billing date as well. If more card-required members leave after one payment, the gap widens further. Always follow each group at least one payment past conversion.

Other trial numbers worth watching

• Time to convert. Do people who pay early in the trial behave differently from those who wait until the last day? Early payers often found value quickly, which tells you what to show everyone sooner.
• Cancellations during the trial. With a card-required trial, a cancellation on day one usually means someone wanted to avoid an automatic charge, not that they disliked the membership. Cancellations late in the trial are the ones to investigate.
• Conversion by source. Trials started from a partner's recommendation may convert very differently from trials started from a social post. A drop in the overall rate may simply mean your traffic mix changed.
• Revenue per converted trial. If most trial users choose your cheapest plan, conversion can rise while revenue stays flat. Compare the plan mix of converted trials with your overall average revenue per member.

Keep small numbers in perspective

If you start 40 trials a month, one or two extra conversions will swing your rate by several points. Before reacting, combine two or three months, or look at a rolling three-month figure. Make one change at a time, such as a new welcome screen, and note the date so you can compare the groups that started before and after it.

Setting up your trial conversion report

1. Write down your definition of a conversion, including the refund window.
2. Export trials with their start date, source, key activity during the trial and outcome.
3. Group them by start month and wait until every trial in the group is complete.
4. Calculate the overall rate, then the rate for people who did and did not reach your key activity.
5. Track paying members per 100 trial-page visitors alongside the rate.
6. Follow each group to the second payment, and pick one stage to improve next.