When Uptime started charging users for the first time, they did so by using a heavy discount to test the market appetite. We needed to maintain a consistent conversion rate until reaching the full price—a six-fold increase.
Team
This was a team effort. The Growth team consisted of three engineers, a Product Manager, and a Data Analyst.
This is only a brief version of the project. Reach out if you’d like to know the full story.
Goals
Uptime helps people learn faster by summarising books, documentaries, courses, and podcasts into 5-minute ‘Hacks.’ We launched with a discounted £9.99/year price, aiming to reach £59.99/year. The challenge was maintaining a stable Conversion Rate (CR) during the transition to full price.
Baseline Conversion
Our CR was in the neighbourhood of the App Store’s “best in class,” likely thanks to our unbeatable intro price of £9.99/year.
The real challenge was maintaining the CR through the price increases, which together with other metrics would allow us to get a good Series-A valuation. We planned for three price hikes over the upcoming months (£9.99/year → £29.99/year → £39.99/year → £59.99/year). Over this whole period, we ran hundreds of experiments to optimise the CR. Below is a taster of some of them.
First increase
Note that only a fraction of the experiments are shown.
The first price hike lowered the CR by about 89%, which was expected considering the increase from £9.99 to £29.99/year.
We optimised the funnel through a combination of design adjustments and copy optimisation. In this first phase, experiments were based on:
- ideas inspired by other products
- pre-MVP ideas that we left for later
- best practices proposed by our Apple & Google partners.
End of Hack paywall ✅
Looking at data, we noticed gaps in our paywall mapping. We’ve concluded the best time to show a paywall is right after users gain value from the product. One of these moments was finishing a Hack – not a place where we initially added a paywall to not block users from finding something else to consume. Our hypothesis was that people were more likely to pay just after they learned something new. By adding a paywall at the end of a Hack we instantly converted 174% more users.
Gift pop-up ✅
One of the most surprising tests was the gift pop-up. Users would get a 3-day trial on sign-up no matter what. But we tried to reframe this as something that we’re doing as a gesture of goodwill, rather than by default.
We added a paywall with a twist at the end of the sign-up flow: if users dismissed it, a pop-up offered them 3 days free as a ‘gift.’ The product didn’t change – it’s still 3 free days, only now instead of getting it by default, users get it as a gift.
I was skeptical about this experiment. Showing a paywall before users get to experience the product… who would take us up on that? But data argued back. This experiment increased our CR by 125% through a combination of immediate subscriptions and better trial-to-payer conversion.
This made users feel good, since it looked like we were gifting them 3 days for free to check out the product.
Second increase
Note that only a fraction of the experiments are shown.
Bringing the price from £29.99 to £39.99/year decreased our CR by 33%.
During the following weeks, rather than focusing on experiments related directly to conversion, we focused mostly on:
- increasing daily usage and retention
- optimising our onboarding and feature discovery
- stacking experiments on top of previous successful ones.
Autoplay ✅
Our analytics showed the paywall at the end of a Hack performed well, so we focused on getting more users to it. One tactic was AutoPlay, which suggested similar content at the end of a Hack. The hypothesis was that more completed Hacks would lead to more exposure to the high-performing paywall, boosting conversions. This experiment increased End of Hack sessions by 23%, giving us more opportunities to promote Premium.
The next Hack plays automatically after a few seconds.
Premium banner ❌
Not all our experiments were successful. A change that we tried was pushing the Premium banner at the top of the homepage for Free users. We based this on another hypothesis: if we show the upgrade banner at the top of their homepage, more people will subscribe.
The results were flat after two weeks, so we rolled it back since this banner was pushing our most important asset–the content–further down the page. Having the banner at the top, we would later learn in usability testing, created distrust with our users.
The banner was initially in the second position
Adding it to the top was an unsuccessful experiment, so we rolled back to control.
Third increase
Note that only a fraction of the experiments are shown.
After the last price hike to £59.99/year we saw a 45% CR drop. After months of optimising, we got to a stage where we could only squeeze small percentages, rather than leaps of magnitude as before. We focused mostly on work higher up the funnel:
- getting more people to open and finish Hacks
- integrating new features
- increasing retention.
Spotlight ✅
Users saw the most value in Uptime once they consumed a Hack. That’s also where we saw our best conversion. This meant we had to get as many people as possible to consume a Hack as their first action.
We noticed that we were losing about 12% of users after sign-up, but we didn’t understand why. In testing we’ve learned that the next action wasn’t instantly clear to everyone; this would turn people away from the app. So we designed a ‘spotlight’ that appeared immediately after sign up, clarifying the proposition and pushing users to open a Hack.
This increased the number of new users opening a Hack by 3%. It wasn’t a huge win, but brought us closer to 100% engagement post-sign-up.
Users were dropped on the feed and had to figure out what to do on their own.
Two variants of the Spotlight test handholding users through their first visit.
Streak completed celebration ✅
When users completed their daily streak, we celebrated them with an animation. Testing showed us that users were deeply attached to their streaks, so we hypothesised they’d appreciate reminders to keep them going.
We changed the CTA on the celebration screen, prompting users to enable notifications for streak reminders. This experiment increased the amount of users setting reminders by 686%. By getting more users to return daily to maintain their streaks, retention improved by 11.6%—a monumental success that also boosted conversions later on.
Adding this call-to-action here led to one of our most successful experiments.
Conclusion
After six months of optimisation, we brought the CR back to nearly 100% of baseline, despite a six-fold price increase—a massive win.
Behind this success was a huge amount of invisible work. The design work is obvious and easy to talk about, but the real effort lay in endless meetings with the founders and senior stakeholders, framing Design and experiments effectively and even the cheeky “get this experiment approved in any way possible because I’ve got a hunch about it“… all of these were efforts that added pressure on a daily basis.
This project happened during a sensitive time as the company was raising capital. Founders relied on the conversion for VC calls, and even a single day’s drop triggered urgent demands for fixes.
While we later shifted to more strategic, long-term projects, I continued overseeing monetisation efforts throughout my time at the company.