Unconverted trials: what are they worth, and should you fix them first?

An unconverted trial is potential revenue, not lost revenue, so I value it separately and keep it out of the headline leak. To size it, add up the monthly plan value of every trial that ended without a paid invoice, then spread that over the months in your window. Whether to fix it before payments or cancellations depends on your own numbers, and this article shows how to compare them.

What counts as an unconverted trial?

A subscription that had a trial, whose trial ended in your window, and that never produced a paid invoice after the trial started. That is the whole definition. Trials still running are not counted, and neither are trials that converted.

In Stripe this is readable from the subscription's trial start and end dates and its invoices. You do not need product data to count them. You do need product data to understand them, which I come back to below.

How do I put a dollar figure on them?

Take each unconverted trial's plan, convert it to a monthly value (list price times quantity, with yearly and weekly plans normalised to a month), add them all up, and divide by the number of months in your window.

The result is a monthly figure for what these trials would have been worth if every one had paid. Read that last clause twice. It is a ceiling, not a loss. Not every trial would have paid, and some were never going to.

That is why I keep the number out of the headline. In my report, the headline is the sum of four realised leaks, and unconverted trials sit beside it as a labelled potential. If you add potential to realised, you get a bigger number that you cannot defend.

What does that look like in practice?

In the synthetic sample report (a fictional company, Lumenfold Analytics, not a real customer), 426 trials ended in the window and 177 never paid, a conversion rate of 58.5%. Those 177 come to about $989 a month of potential, shown separately from the $2,789 a month of realised leaks.

The data is made up on purpose, so that I know what the right answer is. Do not compare your conversion rate to 58.5%. I do not have a trustworthy benchmark, and I would rather say so than quote one.

Why do trials not convert?

The mechanism I work from is simple: trials that reach a first moment of value tend to convert, and trials that never do rarely do. The practical question is how many of your trial users reached that moment, and when you could have told.

Stripe cannot answer that. It knows about billing, not about whether someone created their first project or connected their first account. To see activation you need product events from your own app. That is the main gap between what a billing export shows and what you need to act.

Two other causes sit in billing itself. Some people forget they started a trial and are surprised by the first charge. Some have a card that fails on the first charge. The second one is a payment recovery problem in disguise, and it belongs to the work in how to recover failed payments on Stripe subscriptions.

What can I do about them?

Three things, in rough order of effort:

1. Nudge trials that have not activated, early. Waiting until the last day is too late for someone who never got started. The message should point at the one step that gets them to value, not at a feature tour.
2. Remind people about the card and the first charge before the trial ends. Said plainly, with the date and the amount. It takes the surprise out of the first charge.
3. Look at the trial design itself. Length, card up front or not, what is included. These are product decisions, and I cannot make them from billing data.

I would try the first two before touching the third. They are cheaper and easier to reverse.

How do I check that a trial fix worked?

Treat it like any other change: split the trials at random, send the new nudge to one group and nothing new to the other, and compare how many in each group paid. Keep the split fixed for the whole test. If you change who gets the message halfway, the comparison stops meaning anything.

Decide the metric before you start. I use net revenue per customer, after the cost of any offer and after margin, because a trial that converts only because you gave away a discount may not be a win. And be patient with small numbers: with fewer than about 30 conversions in a group, all you have is an early signal.

One more limit worth stating. Stripe will show you who paid, but it will not show you whether your nudge was the reason. The holdout is what lets you say that, and a before-and-after chart cannot.

Should I fix trials before payments or cancellations?

Compare them in dollars, from your own data, and discount the trial figure because it is potential. There is no general answer, and I distrust anyone who gives one without seeing your numbers.

Here is how I would set up the comparison:

• Failed payments never recovered: realised cash. The customer wanted to pay. Usually the cheapest to fix.
• Voluntary cancellations with no visible save attempt: realised, but Stripe cannot see saves made by email or support, so this is an upper bound.
• Unconverted trials: potential. Count only a share of it, and be honest that you are guessing the share until you have tested.

In the synthetic sample, the payment fix is worth about $1,362 a month, the cancel flow about $1,134, and trials about $989 of potential. In that fictional company I would start with payments. In yours the order may flip. That is the reason to compute it rather than copy it.

What does it cost to find out?

If you do it yourself, the cost is your time and some care. You need the trial and invoice data for the last year, the monthly conversion of each plan, and the rules above applied consistently. I have not timed how long that takes you, and it depends on how clean your Stripe data is. If you have fewer than roughly 2,000 paying customers, I think it is the right call to do it by hand, because there will be too few trials to pay for any outside analysis.

If you want it done, the Revenue Leak Report costs $490. You send a restricted read-only key; the report arrives within 48 hours of my receiving it. It lists four leaks in dollars per month, with unconverted trials reported as potential beside them, and it names the top three fixes. If after reading it you want to test the fixes, I run a 90-day pilot at $1,500 setup plus 20% of the measured uplift, against a 20% holdout. Uplift is net: revenue times margin, minus the cost of the offers. If there is no uplift, there is no percentage.

Whether you buy the report or not, the habit is the same: measure first, split realised from potential, and test one change at a time. If you want to start with the definition of the leak that causes most disagreement, read what involuntary churn is and how to measure it.