Pick your SaaS free trial length from your activation data
Set your SaaS free trial length from how fast users reach activation. In a two-year field experiment, 7-day trials beat 30-day ones by 5.59%.

The right SaaS free trial length is the shortest window in which a typical new user reaches your activation moment, plus a few days of margin. For a product people can set up alone in an afternoon, that is often 7 days. For a product that needs a data import, a teammate or a sales call, it is 14 to 30. Pick it from your own activation timing, then test it against what you run today.
The default answer is 14 days. ChartMogul's conversion report found 62% of free trial products use 14 days, with 7 and 30 days at 14% each. A default tells you what other teams picked. It says nothing about how long your users need.
What the evidence says about SaaS free trial length
The cleanest study we know of is a field experiment published in Management Science in 2023. A large software firm that gave every new user 30 days randomly assigned new trial users to 7, 14 or 30 days and followed them for two years. Giving everyone 7 days lifted subscriptions 5.59% over the 30-day baseline. Fourteen days did not move subscriptions significantly.
The authors found two forces pulling against each other. Longer trials meant more total usage, and usage predicted paying. Longer trials also meant fewer active days per user, less use per day, and a longer gap between the last active day and the end of the trial. That idle stretch at the end went with lower conversion, and in their data it won.
A smaller dataset points the same way. MadKudu looked at nine SaaS companies on 14-day trials, 30-day trials and freemium plans, and Tomasz Tunguz reported that in all but one, half of the paying customers had converted within 30 days. In all but three, half had converted by day 14. His read: trial length on its own is unlikely to change conversion much. What happens inside the window does.
Measure your time to activation first
Your trial length comes from one number: how many days a new user takes to reach activation. If you haven't named that moment yet, start there; we wrote about defining it in the growth metric SaaS teams track wrong. Then pull the last 90 days of trial signups and read four numbers.
- Median days to activation. Where the typical user lands.
- 75th percentile of days to activation. Where the slower, serious users land. This is your floor.
- Last active day for users who didn't pay. If most stop by day 4 of a 30-day trial, the other 26 days are idle time.
- Day of payment for users who did. If payments bunch on the final day, the deadline is doing the work, and a shorter one brings it forward.
Set the trial at the 75th percentile, rounded up to 7, 14 or 30 days. A product whose 75th percentile lands on day 3 has no reason to offer a month.
When a longer trial earns its place
Some products need the time. Three signs yours is one of them:
- Value shows up only after data piles up: an analytics tool needs a week of traffic before its reports say anything.
- The buyer isn't the user: a team lead tries it, then needs finance or IT to sign off.
- Setup needs someone else: an integration, an SSO config, a CSV of last year's data.
The same experiment found that experienced users responded better to longer trials and beginners to shorter ones. A targeted policy built on that lifted subscriptions 6.8% over 30 days for everyone, against 5.59% for a flat 7. So if your signups split cleanly, say a self-serve tier and a team tier, each can get its own length. If they don't, one short trial is the simpler bet, and in the paper it came close.
Test the new length before you commit
Trial length is one of the few pricing levers you can test without touching the price. Run the current length against the new one on new signups only. Judge both arms on paid conversions per signup, counted the same number of days after signup (45 days works for a 7 versus 30 test), so the longer arm isn't caught mid-trial.
Keep the card rule out of the same test. ChartMogul puts a good card-required trial at 25-35% trial-to-paid and a good card-free one at 4-6%, so changing both at once tells you nothing about length. And if you get a few hundred trials a month, a clean answer takes a while: our guide to A/B testing with low traffic covers the options, and the sample size calculator tells you how many trials each arm needs.
While the test runs, put the days to work. Send the email that pushes toward activation on day 1. Show the days left inside the product. Reach the users who went quiet before their trial ended, because that idle stretch is the one the experiment tied to lost subscriptions.
If you'd like help reading your activation data and setting a trial that fits it, that's the kind of work we do.