“Unlock freemium engagement, turn it into App Store rankings and organic traffic, and grow revenue from a bigger, stickier base.”
Replacing the 5-minute limit that told most of our users to leave.
Drops monetised time: five minutes of free learning, then a ten-hour wait. It worked as a paywall, but it meant the majority of users ran out of time and were told to come back later, every day. The bet: a freemium model that lets people learn as much as they want could grow learning, App Store rankings and revenue at once. High-risk and high-reward, so I ran it as a staged experiment to de-risk a change to our core monetisation lever. We validated the upside, tested two replacement strategies, and parked it, leaving a reusable framework for testing any future lever.
We monetised time, which meant telling most of our users to leave after five minutes.
Drops put a hard limit on free learning: five minutes, then a ten-hour wait before you could meaningfully progress again. We advertised it as five minutes a day, and many users assumed it was a whole day. It had been our core monetisation lever for years, and the idea of removing it had sat in the backlog for over two and a half years, never prioritised because no one was confident the product or the business was ready to let it go.
Time actually did two jobs, and we’d conflated them. As a gameplay mechanic the countdown worked well. It created urgency and a tidy daily ritual, and most timer feedback was positive. As a monetisation wall it did real damage: against competitors who don’t cap free learning at all, it left users with a poor impression of how much they could learn with us, throttled activation while motivation was highest, and drew a steady stream of negative reviews every month. Reviews and churn both feed App Store rankings, so the limit wasn’t just costing us conversions. It was quietly suppressing our organic growth. More than 70% of our users were freemium, and every one of them hit that wall daily.
Treat a high-stakes change as a series of cheap experiments, starting with the riskiest assumption.
The hypothesis was a flywheel. If we replaced scarcity with a model that let people engage as much as they wanted, engagement would rise. More engagement means more learning and more “aha” moments, which brings people back for more. A stickier, more-used app is a quality signal to the App Store, which lifts rankings and brings in more users. And if we held or improved monetisation across that larger base, the result is more revenue and more growth, funding the next turn of the wheel. The catch: this meant replacing our core monetisation lever, which is about as high-risk as product bets get.
So I refused to bet big in one move. I broke the work into staged experiments, each one buying down a specific assumption before we spent more. First, validate the cheapest, riskiest belief: that removing the limit actually lifts engagement. We tested it in low-revenue but high-install markets, where we could learn from real behaviour with almost nothing at stake commercially. It did, sharply. Then the harder question: what replaces the limit? An effective freemium model needs some friction, or no one upgrades, but too much friction churns users and tanks reviews, the exact trap we were escaping. I scored every candidate strategy on that balance, on whether it fit our principles, and on whether it could move the core loop, then sequenced what survived by build effort: test the cheapest idea first.
- STEP.01 Validate the upside Remove the limit in low-revenue, high-install markets and watch engagement. It rose sharply, with little commercial risk, confirming the core assumption before we built anything to replace the lever.
- STEP.02 Find the right friction Score every replacement candidate on a single tension: enough friction to drive upgrades, not so much it churns users or hurts reviews. Filter against our monetisation principles and whether each could actually move the core loop. Two strategies survived.
- STEP.03 Test the cheapest first Content capping (a set number of new terms a day, with review left unlimited) was quickest to build, so it ran first. It failed, and failed clearly enough to tell us not to iterate on it, but to move to the stronger bet.
- STEP.04 Build the higher-belief bet Gamified content limits: an energy bar that drains with mistakes and recharges through review, so progress is paced by learning rather than a clock. Higher effort, but this is where our conviction was. Gamifying the limit had real USP potential. Designs were completed and shipped to test.
- STEP.05 Bank the reusable framework Building content capping forced us to map the key moments and surfaces where any new lever gets introduced, explained and triggered. That became a monetisation-testing framework that made every future strategy far cheaper to trial. It’s the most durable thing the project produced.
Not shipped, but the upside was real, and the framework outlived the project.
- the growth upside
- Validated
- strategies progressed
- 2 of 7
- for future tests
- 1 framework
- after the gamified test
- Parked
The staged approach did its job: we proved the central assumption (uncapping engagement lifts it, and meaningfully) without risking the business to find out. Content capping then failed as a replacement, clearly enough that the right call was to stop, not iterate. The gamified energy-bar model was more promising and got as far as completed designs and live testing, but for a small team it wasn’t returning the dividends we’d hoped for in the time we had.
So we parked it. The reason wasn’t the bet. Our strategy shifted for unrelated reasons, and the resources this needed went to another initiative. What stayed behind was worth keeping: validated and invalidated assumptions that de-risk the next attempt, and a reusable framework for introducing, explaining and triggering any monetisation lever, so whenever we want to test a new one, the expensive groundwork is already done. It’s in the backlog, ready to pick up again.
The premium experience is where the gamified model came undone.
Gamifying the limit created a risk we could see coming: if the energy loop was genuinely fun, what happens when someone upgrades and it vanishes? So we tried to keep it for premium users, repurposed from a paywall into a guided learning loop: make enough mistakes to drain your energy and we’d suggest reviewing to shore up what you’d learned before moving on to new content, but you could override it. In testing it confused premium users. It made them feel they hadn’t fully removed the friction they’d paid to remove. If I picked this up again, that premium experience is the first thing I’d redesign.
A smaller miss taught the same lesson about clarity: we visualised energy as a battery, which sat right next to the phone’s own battery indicator on screen. Two batteries, two meanings, one of them ours and confusing. Obvious in hindsight. A monetisation mechanic has to read instantly, or it adds friction of the wrong kind.
A bet can be worth making, and worth stopping, and still pay off.
This sharpened how I de-risk big swings. The instinct on a high-reward idea is to build the whole thing and find out. The discipline is to isolate the single riskiest assumption and buy it down as cheaply as possible. Here, that meant proving the engagement upside in markets where being wrong cost almost nothing. Sequencing the replacements by build effort rather than by conviction meant our first real-world failure was also our cheapest, and a failure that tells you clearly to stop is a genuinely good outcome.
The other lesson is that not every high-potential project ships, and that’s fine if you run it well. We left with validated learning and a reusable framework that makes the next attempt faster. That value survives a parked project. Knowing when to stop spending a small team’s time, without writing off the opportunity, is its own skill. The idea is still in the backlog, and the groundwork is waiting for whoever picks it up.