How iCabbi made observability central to their app redesign

The problem: high Datadog costs, lack of comprehensive observability

Before Bitdrift, iCabbi didn’t have a clear way to do observability.

The team was relying on Datadog, but high costs kept them from seeing the data they really needed. Specifically, the team relied on heavy filtering to keep costs maintainable, but this meant the data flowing through wasn’t always useful. This lack of visibility into what was really happening in their Driver app meant users were left to report issues to support – and the time it took to verify then debug those issues was taking precious time away from the team.

With Bitdrift, the iCabbi team has been able to expand their focus on mobile observability and claim back some of that time.

“I was not very aware of what mobile observability needed to do or why we needed it” says Andy George, Senior Product Manager at iCabbi, “it's been a great learning experience for me…. now it's really front and center”

Building reliability into an app rewrite

When Renault purchased a majority stake in iCabbi, the team had the chance to reimagine the way their app was built.

“We've rewritten both of our apps from scratch” says Andy. This rewrite afforded the team an opportunity to make observability a first class citizen: “I've taken the stance that, we spent so much money on this rewrite, to go out and to let it fall apart in, you know, 10 years doesn’t make sense.” He goes on, “we had a chance to throw out the startup code and now use code that's more structured and meaningful, [and] that leverages the latest technologies and observability to help us keep an eye on performance.”

For instance, the team was running into an issue where taxi fares would jump unexpectedly, causing confusion among their users. Now, in Bitdrift, the team has a workflow setup that will trigger automatically if an unexpected fare jump occurs – giving the team the precise data they need to help solve the issue. “It's like having the guards at the front door” says Andy, “we're ready to go and we're just waiting for the thing to happen. And that's very reassuring. If something does happen, we're going to catch it.”

Using Bitdrift to preempt issues

With Bitdrift, the iCabbi team is able to find and address issues before they get a user complaint. Rather than waiting for a customer to reach out to say “this app is crashing” or “there’s an issue with load time”, the team can identify those potential problems - and why they might be happening - in Bitdrift.

For example, Andy shares: “I think it was over Christmas we spotted an area in the UK that all of a sudden was getting all these network issues coming in. [We were able to] ring support and say, ‘you're going to get calls from these fleets because their Driver apps are not going to work properly and it's because of the connectivity issues.”

Over the last few months, the team has been able to build out more and more workflows in Bitdrift to catch relevant data for recurring issues.

Arming support with the right information

Similarly, Bitdrift equips iCabbi’s support team with the exact information they need to help troubleshoot issues.

With session replay, the team can see exactly what happened before, during, and after an issue has occurred. Rather than relying on customer testimonials alone, the team can dig into what happened, frame-by-frame, in their app via Bitdrift.

Automated workflows also help streamline the support process. When a driver contacts support, Bitdrift automatically triggers a session collection when the “contact support” button is pressed. This enables iCabbi’s support and engineering teams to look at detailed session information - giving them a full view into the experience of the user and reducing overall MTTD.

Conclusion

With iCabbi running in over 80,000 devices, the team relies on Bitdrift to identify potential issues, debug problems as they occur, equip support with precise information, and to keep their observability costs predictable.

To learn more about Bitdrift, check out our website, or try it yourself in our sandbox.