How to run free trials and free tastes

1 The bottom line

Spell out how the trial works, remind users before billing, and test length, free tastes and paid intros against your costs. Fear of surprise charges and of paying before use is what stops people.

2 Key takeaways

• Show a timeline and promise a reminder before the first charge. Fear of a surprise charge is the main thing that stops people starting a trial.
  • Spelling out the 7-day trial with a Day 5 reminder and a Day 7 charge led to 23% more trial signups at Blinkist.
  • Adding trial reminders gave a 48% increase in revenue at Koby Conrad, Sunflower46:39.
• Set trial length per plan, and weigh conversion against cash and compute cost. Longer trials usually convert better, but they delay cash and cost money to serve.
  • A 7-day trial on monthly and a 14-day trial on annual led to a double-digit conversion increase at Headspace.
  • Trials of 17 days or longer convert 42.5% versus 25.5% for short ones Jacob Eiting and David Barnard, RevenueCat54:37.
• Give a limited taste of the premium product, especially to people who decline the paywall. Users who have felt the value need less convincing.
  • A reverse trial moved free-user conversion from 0.4% to 4.5% in Steve P. Young, App Masters11:16.
  • Upgrade rates nearly doubled after Grammarly mixed paid suggestions into free users' writing Albert Cheng, Chess.com23:35.
• Test a cheap paid intro or no trial where free trials leak. Free trials attract people who start and never pay, and warm buyers do not need one.
  • A $1.99 first month beat a 7-day free offer, with conversion rising from 19% to 25% Steve P. Young, App Masters6:21.
  • Dropping the trial on the monthly plan lifted monthly subs by 2,000% at Nancy Anderson, Natal1:02:02.
• Track trial start rate and trial to paid rate against benchmarks. They show which part of the funnel to fix first.
  • Trial start rate should be 5 to 10% of downloads Steve P. Young, App Masters7:10.
  • Trial to paid should average 30 to 40% Steve P. Young, App Masters7:29.

You give something, then you take it away.

Steve P. Young, App Masters

2.1 Measure your trial funnel first

Know your start rate and your trial to paid rate before you change anything.

• Read both rates from your trial conversion chart. Steve P. Young, RevenueCat4:11
  • Switch the chart from absolute to relative to see start rate and trial to paid side by side.
  • Ask for the start rate first. If it is low, work on onboarding, paywall layout and pricing Steve P. Young, App Masters7:10.
  • A niche app pushes both numbers higher. His 7-day health and fitness app converts over 60% of trial starters, and seasonal apps like fitness peak in January Steve P. Young, RevenueCat4:11.
• Treat a start rate above 10% as great. Steve P. Young, RevenueCat4:22
  • Above 20% is amazing but very rarely seen.
• If trial to paid is below the normal range, test pricing and trial length. Steve P. Young and Chaim, B7 Dev2:53
  • Look at other variables too, not only those two.
  • A client app reached about 35% after pricing and onboarding changes Steve P. Young, App Masters7:29.
  • A trial never converts 100%. The highest rate Steve has heard of is about 70% Steve P. Young, App Masters22:03.
• Fix trial activation before other funnel pieces. Idan Waller and Steve P. Young, BlueThrone30:30
  • Pick one funnel piece and work on it before the others.
  • Apps BlueThrone bought went from $1,000 a month to $10,000, then $12,000, with no extra downloads or new features.

2.2 Make the trial feel safe

Say exactly what happens and when, and promise a reminder before the charge.

• Show a step-by-step timeline on the paywall. Blinkist
  • Spell out today, the reminder on Day 5 and the charge on Day 7.
  • Add a reminder two days before the end and cancel-anytime messaging, as Morrie Schonfeld, Pingo49:50 does.
  • Be explicit about where users are in the trial cycle, and show one or two examples of what good looks like Oji Udezue, Typeform42:50.
  • Customer complaints fell by 55% after the new trial design Blinkist.
  • Pingo's timeline paywall beat the one-page paywall and the social proof variant. It lifted conversion more than testing value propositions or testimonials Morrie Schonfeld, Pingo49:50.
  • Pages that spell out the timeline won A/B tests at Cem Kansu, Duolingo29:28.
• Promise a reminder, then send it. Shaw Academy
  • Shaw Academy emails users 3 days before the trial ends.
  • Prime users before the paywall with a line like "we'll remind you before your trial ends" Koby Conrad, Sunflower46:39.
  • Explain the trial and ask for notification permission before the timeline paywall Steve P. Young, App Masters5:09.
  • The reminder gave a 14% increase in sales at over 99% significance, with no significant effect on cancellations Shaw Academy.
  • Tying the permission to the reminder raised push opt-ins from 6% to 74% Blinkist.
• Let users pick their reminder day. Anmol Tiwari, Duolingo2:38
  • Duolingo normally sent the reminder about 2 days before the end. The extra step draws attention to the promise, so users feel safe to wait until day 5 or 6 to judge the value.
  • The test gave a conversion uplift Jacob Eiting and David Barnard, RevenueCat46:30.
• Add the reminder line only where it fits. Steve P. Young, App Masters9:17
  • It works best when most buyers choose yearly. Skip it for weekly plans.
  • If you cannot send a push two days before billing, leave the line out.

2.3 Set trial length by plan and by cost

Longer trials tend to convert better, so choose length per plan and test it.

• Give annual plans a longer trial than monthly plans. Headspace
  • The longer trial makes the bigger commitment more appealing.
  • Test a wide spread such as 7, 14 and 30 days to see where a longer trial stops helping Headspace.
  • Baselines: 7 days on yearly, sometimes 14, and 3 days on weekly. Do not jump to 30 days, which mostly won't work Steve P. Young, App Masters11:11.
  • Headspace judged the test on conversion lift and accepted slower cash timing, because the extra conversions outweighed the delay Headspace.
• Test which plan carries the trial. Steve P. Young, App Masters13:18
  • One app kept prices at $50 a year or $6.99 a week and moved the trial from the 7-day yearly plan to a 3-day weekly plan. Sales tripled with similar downloads.
  • A toggle lets users switch between the weekly plan with a trial and the yearly plan without one. More people took weekly and fewer cancelled. Lifetime value rose because users forget or stay a few weeks. It suits single-use apps like photo, video, AI and PDF converters Steve P. Young, App Masters24:59.
  • In the controlled test, the weekly 3-day trial brought $4,000 in three months. Variant A made about $2,000 in total, although A works well for most of Steve's clients Steve Young, App Masters19:40.
  • Genie's top earner is the $12.99 Advanced weekly plan, the only one with the 3-day trial Steve P. Young, Genie6:06.
• Weigh the conversion gain against cash flow and compute cost. Jacob Eiting and David Barnard, RevenueCat54:37
  • The long-trial lift may be selection bias, since different kinds of apps offer long trials. Many apps moved to 3-day trials for cash flow and faster data for onboarding and paywall tests.
  • For AI apps, 30 days of free usage lets heavy users run up compute costs and then churn. Tolen started at 7 days and adjusted over time Phil Carter, Elemental Growth56:05.
  • Use a longer trial only if your audience already trusts you. YNAB hard-gates behind a one month trial and is still a top grosser, while 14 days is typical for Steve's clients Steve P. Young, YNAB6:55.

2.4 Give a free taste of value

Let users feel the premium product, either just before the paywall or just after they decline it.

• Run a reverse trial for users who dismiss the paywall. Steve P. Young, App Masters8:02
  • Say "enjoy a free week on us" and let them in with Pro. Use secondary onboarding to get them adopting Pro features, then say Pro ends in 5 days and offer a small discount.
  • It works best when your app has low usage costs Steve P. Young, App Masters11:16.
  • BoldVoice gives speech results, a practice set and a coach video free on day one, then requires the 7-day trial from day two Anada Lakra, BoldVoice24:40.
  • Apple's average install to paid is 2 to 3%, against 4.5% for the reverse trial Steve P. Young, App Masters8:02.
  • Counterpoint: a one-week reverse trial benchmark did not hold for ElevenReader, so test it Jack McDermott, ElevenLabs49:10.
• Let users sample before the paywall appears. Peptalk
  • Play a premium video for 10 seconds, then overlay "Enjoying our exclusive video content?" with a "Continue to Watch" button that opens the paywall.
  • Do not add a details screen first. Extra information did not build enough desire and added a step Peptalk.
  • On a hard paywall, give the first scan free. If they close the paywall, push about five minutes later with "your first snap is on us". Then show the paywall again with no trial Steve P. Young, Cal AI9:49.
  • The 10 second preview converted significantly better than the details screen.
• Cap a real taste of paid features inside the free tier. Albert Cheng, Chess.com23:35
  • Tie the taste to usage, not time. The free product should show everything the product can do.
  • Flighty gives the first flight free with pro features, then locks them Jeff Morris, Tinder16:25.
  • Routinery lets non-payers watch an ad to unlock analytics and summaries for 24 hours, then shows a 24-hour discount right away. Most users never come back for a second open TP Young, Routinery8:43.
• Show a longer trial, not a discount, when users exit the paywall. Steve P. Young, App Masters11:11
  • Show a second paywall with a 14-day bonus trial. His usual rule is to discount decliners, but some clients found the longer trial worked better.
  • Pingo shows a 3-day trial first, then a 7-day trial if the user abandons. It plans to test 10 or 14 days, after checking how those users convert post-trial Morrie Schonfeld, Pingo52:10.
  • Internationally the 7-day trial converted about 2% higher than the 3-day. In the US the gap was 0.3 to 0.7 percent Morrie Schonfeld, Pingo52:10.
  • The longer trial works because the user still has to subscribe afterward Steve P. Young and Flo, Sick Days24:58.

2.5 Trim the free trial where it leaks

Remove or replace the trial for warm users, and for channels where trial starters never pay.

• Test no trial for users who start using the app unsubscribed. Steve P. Young, App Masters22:03
  • Keep the trial in onboarding, then show a no-trial paywall to people who declined and started using the app. This needs about four plans, with and without a trial.
  • If you lock content you still need trials, because users already in the app are either in or out Steve P. Young, App Masters17:40.
  • A friend's app saw roughly 4x more sales, reporting $122,000 without the trial versus $4,000 with it.
  • A private photo vault app made $12,000 with no trial versus $4,000 with a trial Steve P. Young, App Masters17:40.
• Try a cheap paid intro instead of a free trial, especially on web. Igor Lyubimov, web2wave12:45
  • Free trial to paid on web is two to three times lower than on mobile. A paid intro brings money right away and gives Facebook faster signals.
  • Offer $1.99 for the first month, renewing at the yearly price. Paying users show willingness to pay and are easier to upsell Steve P. Young, App Masters6:21.
  • Try 99 cents for the first week, or $2.99 to $3.99 for the first month on yearly, for apps people use over a year Steve P. Young, App Masters27:53.
  • Rise offers a half-price first month on web, which stops the ad algorithm finding people who cancel a minute after starting Leon Sasson, Rise Science19:40.
  • Coursiv says a money-back guarantee works better than a trial on web checkout CP Young, Coursiv12:04.
• Do not stack a trial on top of a discount. Steve P. Young, Flo Period and Pregnancy Tracker12:51
  • Keep the trial on the initial paywall and remove it from the discounted one, since the discount is already the incentive.
• Drop the trial on a plan where trusted, high-intent buyers just buy. Nancy Anderson, Natal1:02:02
  • For these buyers the trial is only a delay. Natal made the change permanent after dropping the trial on its monthly option.
  • Quarterly subs rose 46% and annual 21% after the change.

2.6 Where they disagree

Whether to have a free trial at all, how long it should be, and whether to take a card depend on your product and audience.

• A free trial can help, or it can add nothing.
  • An invoice maker app launched without a trial, and sales rose 420% when it added one, because most people buy during onboarding Steve P. Young, App Masters16:00.
  • Duolingo's trials helped even with a free tier, because they remove payment hesitation Cem Kansu, Duolingo27:50.
  • Headway does not recommend free trials at all, though 99 cents for a week or three days can work Igor Lyubimov, Headway36:54.
  • Skylight's trial test, with and without a card, gave no lift, because trial takers were the same people who would have subscribed anyway Michael Segal and Mark Ungerer, Skylight57:27.
  • Migraine Buddy has no trial and makes $50,000 a month, because it solves a real pain Steve P. Young, Migraine Buddy5:28.
• Short trials create urgency, but long trials convert better.
  • Outseta chose 7 days on purpose so users explore quickly instead of putting it off Outseta.
  • Longer trials have converted better every year of the report, yet the share of short trials rises each year. Weigh the gain against the cost of free usage instead of defaulting to short Rik Haandrikman, RevenueCat30:27.
• Take a card up front, take a small payment, or take nothing.
  • Ladder takes no card, reasoning that a better product benefits from more people in it, and puts product effort into the first couple of workouts Greg Stewart, Ladder38:58.
  • Honeydew asks for no card before the first doctor visit, at the cost of more no-shows David Futoran, Honeydew35:24.
  • Noom asks most users to pay as little as 50 cents for the trial, which gives them skin in the game Brent Zajaczkowski, Noom25:01.
  • The 50 cents did not change conversion much, but it improved retention per exposure.