Rebuilding VyprVPN so it just works
VyprVPN's apps had aged: feature accretion, confusing onboarding, and shaky connections had pulled the store rating to 3.2. We refreshed the brand and rebuilt all four native apps around one expectation, that a VPN has to just work, and the rating climbed to 4.7.
- 3.2 → 4.7 iOS and Android store rating
- 4 Platforms rebuilt end to end
- 10M Users reached during tenure
Context
By the time I took over product, VyprVPN was showing its age. It ran on iOS, Android, macOS, and Windows and had grown to ten million users, but it had never had a real UI or UX refresh. The company had kept building on its own success, stacking on features that worked on their own but didn’t sit well together. I led product, user experience, and customer support globally, with a team that included up to five product managers, so the complaint data and the roadmap lived in the same place. That mattered more than it sounds.
The store rating had slid to 3.2. For a paid consumer subscription that’s a serious number. It sits right above the install button, it feeds the ranking that decides whether anyone finds you, and for a security product a low rating is its own red flag. People don’t route their traffic through something the reviews call broken.
The reviews said the same thing on every platform: connections that dropped or reconnected badly, and an app that had grown confusing enough that people weren’t sure it was even working.
The constraint
A VPN has one expectation above all others: it has to just work. That comes before locations, ease of use, privacy, or protocols. So whatever we changed on the surface, we couldn’t put the reliability of a live connection at risk. If anything, it had to get better at the same time.
The rest of the problem was scope. The apps needed more than a coat of paint. Onboarding confused new users at exactly the moment they were deciding whether the product was worth keeping. The interface had drifted. And we were looking at four native platforms, each with its own conventions, plus a brand that hadn’t kept up. Fixing any one of those alone wouldn’t move the rating. They had to be done together.
The call
We treated the rating as a lagging indicator and went after the causes on two fronts at once.
Underneath, connection quality was the real complaint, so we ran a tech-debt project in the background to make sure the first expectation was always met: the app just works, every time. More reliable connections took the most common complaint off the table at its source.
On the surface, I decided we needed a full refresh, brand and application both, not a patch on the worst screens. I led the design work with our creative team, and we tested it against real customers with live user testing and interviews with actual VyprVPN subscribers, not just internal opinions. Then we rebuilt all four native apps, iOS, Android, Windows, and macOS, from the ground up into one coherent, good-looking product.
Onboarding got treated as its own problem rather than the first screen of the app. It’s the moment a new user either reaches a working connection and sees the point of VyprVPN, or gives up. I focused the team on getting people to that first connection fast, because that’s the habit-forming moment, and a VPN people forget to turn on is a subscription they cancel.
We watched all of it in Mixpanel. We were analytics-heavy by design, so onboarding time, connection outcomes, and habit-forming behavior were measured, not guessed at.
The rebuild moved the metrics on its own, and better reviews followed almost overnight. After that, I built a smarter review-prompt system on top of the same data. We classified how people used the app and how their connections behaved, and ran a set of algorithms to read whether a user was trending positive or negative. Users having a good experience got asked for a review. Users trending negative got a different path: an automatic connection-troubleshooting attempt first, and if that didn’t clear it, a direct line to our support team to make it right fast. The aim wasn’t to keep unhappy users quiet. It was to not abandon anyone. If we couldn’t get someone to a clean connection automatically, a person would.
The tradeoff
The honest answer on cost is that this was a smaller roadmap trade than it looks. We already had several separate projects planned, each chipping at the same problems. Folding them into one combined redesign and connection-quality effort got us to the same place in about the same time, roughly four months, so we weren’t trading away a quarter of new capability to do it. What we did take on was concentration: one big simultaneous rebuild across four platforms for an established user base, rather than a series of smaller, safer steps.
The outcome
The rating went from 3.2 to 4.7 on both iOS and Android. Connection quality improved sharply, onboarding time dropped, and we could see habit-forming behavior in the data, all tracked in Mixpanel. User satisfaction rose, and revenue increased on the back of the redesign. VyprVPN reached ten million users over the period.
What I’d do differently
Honestly, not much. The project worked, and users told us so. The hard part was the one we expected: change aversion. Rebuild an app millions of people are used to and some will push back, however much better it is. We planned for that and prepped users ahead of time with notifications, and it still took a beat to settle. Given where connection quality, satisfaction, and revenue landed, I’d make the same call again.