Credits, not trials: let the value land first
Our 14-day, card-upfront trial ran out right when users were getting ready for premium. I replaced it with monthly credits people spend at their own pace. Free-to-paid conversion went from ~4% to ~7%.
At a glance
A 14-day premium trial at signup, card required. People balked at the card, and the trial ran out before they were ready for premium.
Most users need one to two weeks to get comfortable with the device, the app and a routine. We were optimising for trial activation, not for discovering value.
Monthly credits every user can spend on AI Coach, Sportstech Move and other premium features whenever they're ready. Upgrade prompts come after use. No hard paywall.
Free-to-paid conversion went from ~4% to ~7% (+75%), measured within 60 days of signup.
Premium was sold against a deadline.
Sportstech sells smart treadmills, walking pads and strength equipment. The app has a free tier. Premium unlocks the features we'd spent the year building: AI Coach, Sportstech Move and other advanced experiences.
The way into premium was the standard playbook. At signup, every user got a 14-day premium trial, and they had to add a credit card first.
Free-to-paid conversion and subscription revenue were my KRAs. Conversion was sitting at around 4%.
The trial ran out before the user was ready for it.
A card wall on day one
People who had only just started with the product didn't want to hand over card details for features they hadn't tried yet. Many never started the trial.
The clock started too early
Most users took one to two weeks to get comfortable with the device, learn the app and build a basic routine. By the time they wanted advanced features, most of the trial was gone.
We were selling the trial. We should have been selling the value.
Talking to users changed the question. The objection wasn't to premium itself. It was to being asked to commit before they knew what it did for them, on a timeline that ignored how long it takes to settle in with a new machine.
A trial starts the clock when the company is ready. I wanted the ask to arrive when the user was.
“We were optimising for trial activation rather than helping users discover the value of premium features.”
Give every user a wallet, not a deadline.
Instead of a trial, every user gets a fixed number of credits each month to spend on premium features: AI Coach, Sportstech Move and other advanced experiences. They try whatever they want, whenever they're ready.
I worked with design and growth on the upgrade prompts, so the ask comes after someone has used premium, not before. There's no hard paywall.
Asked for a card to start a 14-day premium trial. Many users stop here.
14 trial days left10 credits in the wallet. No card, no clock.
10 of 10 credits left this month60 days covers the settling-in weeks plus a full month of credits. That's long enough to see whether using premium turns into paying for it.
- AI CoachSetup plan & questions
- Sportstech MoveA 14-day movement goal
- Advanced workoutsIntervals & programmes
Spend credits on anything premium. Nothing to cancel later.
Liked what you tried? Premium keeps all of it going.
- One currency, every premium feature. Credits work across AI Coach, Sportstech Move and advanced workouts, so people find the feature they would pay for.
- The prompt follows usage. The upgrade ask shows up once credits are spent, when the value is fresh. It never shows up at signup.
- "Not now" always works. There's no hard paywall. When credits run out, the free app keeps working and the wallet refills next month.
A 14-day window would have measured the problem itself.
- Definition
- % of free users who converted to a paid subscription.
- Cohort
- Users grouped by signup month.
- Window
- Conversion within 60 days of each user's signup.
- Comparison
- Signup cohorts from before the credit model launched vs cohorts from after.
Why 60 days, not 14The whole insight was that users need one to two weeks just to get comfortable with the device before they're ready to explore premium. Measuring at 14 days, the old trial's window, would have measured the exact problem we were trying to fix.
Conversion went from ~4% to ~7%. That's +75%.
Free credits could have replaced premium. The metric was built to catch that.
Giving premium away
The risk: credits satisfy people, so they never pay. That's why success meant paid conversion, not credit usage. If credits had cannibalised premium, the headline number would have stayed flat. It rose.
Grading on a moving clock
The 60 days run from each user's own signup date, inside signup-month cohorts. Every cohort, before and after launch, gets the same amount of time to convert.
Asking too early, again
Moving the ask to day 30 would just repeat the trial's mistake. So the upgrade prompts, built with design and growth, follow usage, not the calendar.
Pricing is a timing problem before it's a price problem. Ask for money when people can feel what they're paying for.