Whats worse for app growth? Crashes or a few API Errors? Crashes are worse right? Nope. API Errors are worse. In fact, MUCH worse. API errors result almost immediately in a 17.5% permanent loss of users.
Developers at Rejourney analyzed 1.6 million mobile app sessions across 43 apps with end-users in 169 countries to find interesting patterns and tips to build better products. Among the many finds, this was by far the most interesting and initially counterintuitive. One would assume crashes are worse because they are much more destructive and result in a complete, sudden termination of the app. Interestingly enough, many users immediately come back after a crash. 1,187 out of 1,231 users came back after a crash (96.4%). A majority of these users came back within 6 minutes.
As for API errors, 126,212 out of 152,986 users came back after 1-9 API errors (82.5%). API errors are also highly correlated with other issues. Rage taps, for example, were 8.4x more likley to occur given API errors had occurred.
Further more supporting data, the average future session count of a user that experiences a crash is 171.1 sessions, while the future session count for users with API Errors is 121.7 sessions. For baseline reference, a clean session (no issues) had an avg of 198.1 future opens. Every API error session costs you roughly 76 future opens per user compared to a clean session.
Why? How does this make sense? A crash is the most destructive action that can happen to an app, so why is API Errors a much detrimental growth stunt?
The reason is ambiguity.
When an app crashes, users have sight on what happened. The app closed, and something went wrong one time. The path forward is obvious; just reopen it. There is no confusion about whether the action they were taking worked, or whether their data was saved. It is a clear, recoverable event, and users treat it accordingly.
Inversely, API errors offer users nothing. A button that fails to respond, a feed that stops loading, and a form that appears to submit with no confirmation. The user does not know if the problem is the app, their connection, their account, or something they did. That uncertainty is far more corrosive than a crash. This uncertainty resolves into choosing something else.
From a personal prescriptive, remember the last time a button refused to work, or a sign up failed. Unless the service is critical, you probably abandoned it.
So what should developers do? Monitor and fix quickly. On Rejourney, every session replay comes with all the monitoring tools installed. Every API failure needs to close the ambiguity loop for the user not with an error code, but with a next step or clear path. For example, a banner that says the data is from an hour ago because the refresh failed. Anything that tells the user what happened and what they can do.
Users forgive what they can see and abandon what they can't. Ambiguity is a silent killer in app growth.
