Learning 001: Sometimes you shouldn’t force user login, even if everyone says you should.
I’m blogging about my learnings founding and growing Circuit, my Route Optimizing SaaS startup. I’m doing this partly so I can refer to them later, partly to promote Circuit, and partly so others can learn from my mistakes.
These posts will probably be short and badly written, but here’s 001.
The accepted practice for a SaaS Android/iOS app, or pretty much any app that does more than sell a simple In-App Product, is to force the user to login before they get access to the app itself. The theory behind this practice is simple and makes perfect sense:
- You get the user’s email address, and can use that to keep them engaged with your product, get feedback and sell to them.
- If user commits more “effort” in the login process, they’re more likely to spend that extra bit of time to understand the value proposition you’re offering before uninstalling (If they don’t get it immediately).
- You can track analytics for that user across device and platform, and use this data to help make better decisions.
- Probably some other stuff I’m unaware of.
Firebase Auth UI
The Firebase team have built libraries for a bunch of platforms to help streamline the login process. They’ve been built to handle the entire flow, with minimal to zero effort required on the developers part.
It makes sure everything is secure, automatically logs the user in with Google’s Smart Lock, and according the firebase team, reduces login friction hugely by allowing sign in with multiple providers like Facebook, Google and Twitter.
It also means you can avoid implementing this yourself:
The fact that these libraries exist is great for apps that benefit from forcing the user to login, but it can also tempt you to force user login even if it’s not the right option for your app, like we did originally with Circuit.
Forced Login for Circuit
Here’s what Circuit’s Upgrade Funnel looked like when forcing login, and using Firebase’s Auth UI library.
As you can immediately see, only 58% of users who saw the Login Activity actually proceeded to login. 58% is absolutely abysmal, I was expecting more like 80% at the very worst.
By the time users make it through the rest of the funnel, the install -> subscriber conversion rate was 1.7%.
Here’s the funnel once the login was removed (It’s slightly different, but it shows the same stuff). It shows the conversion rate at 2.9%.
This marks a 62% improvement, which is extremely close to the amount of users we were losing to the forced login. Being so close removes the possibility that those who were not logging in were unlikely to commit the time to learn the product and upgrade anyway, something that I had considered.
Conclusion
Although it’s a shame to lose the email addresses of users that a registering, In Circuit’s case, It’s simply not worth the tradeoff to force users to login.
I think login works for larger companies as users have far greater trust for a company they’ve heard of before, so maybe I’ll revisit the forced login in the future. For now my plan to retrieve email addresses from those who actually subscribe, or what to sync data between their devices.
Takeaway
If you’re forcing login for your app (or are considering it), you should definitely track the metrics and make sure it’s the right decision for your app.
Check out Circuit on the Play Store
Follow me on Medium and Twitter for Learning 002+
