VictorOps is now Splunk On-Call! Learn More.
More often than ever before, people are adopting mobile solutions for everyday problems. Businesses are being built completely around mobile apps and end-users are expecting the same seamless experiences they’ve come to expect through web applications. So, the question becomes, how do DevOps and IT teams consistently monitor mobile app performance and report on the health of the business?
Most end-users are looking to use both mobile apps and web apps and switch seamlessly between the two. Tracking the right mobile app metrics, building useful dashboards and establishing benchmarks for system health can lead to better end-user experiences. And, ultimately, leads to lower costs of downtime and more revenue for the business.
So, let’s dive into the basics of mobile app metrics, how they’re categorized, and what you need to know to manage more robust mobile applications and make incident management suck less.
Because there are so many mobile app metrics to track, you need to break them down into four categories: technical performance metrics, end-user experience metrics, engagement metrics, and business-focused metrics. In this post, we’ll mainly focus on mobile app performance metrics you should be tracking. But, it’s also important to see how these metrics flow into the other mobile app KPIs and benchmarks you should establish across the business.
A measurement of the technical performance of the mobile application and its supporting network and infrastructure. Typical mobile app performance metrics include CPU usage, latency, page load speed, etc. Performance metrics are tactical KPIs based on the way technical services interact with one another and respond to external forces such as traffic.
End-user metrics show exactly who’s using your mobile app, when they’re using it, and how they’re experiencing your service. You can track thinks such as device types and OS versions, physical geography of end-users and the overall activity of people on your apps. App stores provide a lot of basic download and usage data but APM tools such as AppDynamics, New Relic or SignalFX provide much more granular end-user data.
Understanding how engaged people are with the mobile app can help the team make business decisions. How long are users using the app? How long is it between the user’s regular sessions? How often are users deleting the app from their phone? The most effective mobile apps will be engrained in a user’s daily life – adding value to their life.
App metrics need to translate into real business-centric metrics such as customer acquisition cost, customer lifetime value, revenue and, particularly for e-commerce businesses, abandonment metrics. And, of course, a business revolving around their mobile app will live and die by their app store ratings. The better the DevOps and IT team are at identifying problems with mobile app performance, the better the business will be at turning that performance into revenue.
Mobile app performance metrics need defined KPIs that show the engineering and IT team exactly what’s happening with a team’s mobile app. Which parts of the app are “healthy” and which parts aren’t? Every mobile application is different, so the most important KPIs and benchmarks will differ from app to app. But, there are a few basic analytics that every app developer should monitor to ensure they’re maintaining the most robust mobile app experience possible.
A mobile app is considered “crashed” when the launch of the app fails. In the industry, a healthy crash rate is typically between 1-2%. But, if you can gain more information than the simple fact that the app has crashed, you can better fix problems with the app. By setting up powerful APM tools like AppDynamics, New Relic or SignalFX, you can gain actionable insights from your crash reports. Not only can you assess the overall crash rate of your mobile app, but you can see why the app is crashing – helping you optimize it for the future and make it more reliable.
Latency across the entire mobile app is important to keep track of. For both APIs and end-to-end applications, you need to monitor the round-trip time from a request to a response. If the requests are taking longer than one second to receive a response, it’s likely you need to take action to improve the speed of that process. According to a white paper from AppDynamics, if any response takes longer than 3-4 total seconds of response time, the majority of users (60% or greater) are likely to abandon a transaction and may even delete your mobile app altogether. Latency seems like a small performance issue, but if it’s ignored, it can lead to major business impacts.
Over time, you need to be sure your mobile app is scaling with the growth of regular traffic and users. Can the mobile app handle the load? Is there performance degradation as more end-users use your application? Tracking page load speed over time can help you maintain stable user experiences and continuously optimize for more storage and users as your mobile app scales. Depending on when your app sees traffic, you need to be sure the system is able to handle the load when it changes.
Poor connections to networked services, typically via HTTP, can cause apps to crash, create latency and decrease page load speed. Knowing when network errors are happening and what that means to the overall mobile app is important to maintaining uptime. A lot of mobile app developers will focus only on the application, not the underlying infrastructure and network. It’s important to ensure the connection between app and network continues to operate seamlessly.
Now that you’re monitoring the right metrics for mobile app performance, you need to create actionable processes based on that data. It’s one thing to simply know when an error occurs in the mobile app, it’s another thing to know how to respond to it. So, the first thing you need to do is create visualizations and dashboards of important mobile app KPIs and incident management metrics. This way, the team can quickly see how the mobile app is behaving and what end-users are experiencing.
Once you’re ingesting the proper mobile app metrics and putting them into dashboards, you need to assign “healthy” benchmarks to these KPIs. What’s a level of latency that’s acceptable to the business and the end-user? How often should the app crash before sending out an alert to the appropriate on-call responder? The benchmarks you decide upon will rely heavily upon the type of app you’re maintaining and the level of impact that certain metrics have on the business. Associating mobile app performance metrics with appropriate thresholds and automatic alerting processes will help DevOps and IT teams fix mobile app issues faster.
End-users expect constant uptime and performance across all of their technical applications and infrastructure. Many developers and IT professionals rely on mobile apps and web applications to perform their jobs effectively. And, they’re constantly using mobile apps outside of the workplace for personal reasons, from entertainment to navigation to finances, etc. So, mobile app developers need a flexible, on-call incident management system just as much as web application developers or IT infrastructure admins. The faster they can detect issues, respond to problems and remediate them, the more often end-users will walk away with a positive experience.
Incident management for mobile apps takes two things – detailed monitoring of all app interactions and integrated alerting processes based on that information. Then, with mobile app metrics ingested by a collaborative communication tool, engineering and IT teams can work cross-functionally to fix problems with mobile applications.
On-call incident management with mobile apps isn’t much different than with web applications or IT networks and infrastructure. It’s about digesting information about service health and app performance in a quick and easy way, then getting that data to the right person at the right time. In a world of highly integrated systems and intertwined mobile app and web app ecosystems, consistent tracking and alerting on performance metrics is the only surefire way to ensure uptime.
Turn mobile app metrics into a successful process for on-call incident management with VictorOps. Sign up for a 14-day free trial or request a personalized demo to learn more about an integrated approach to DevOps and IT monitoring, alerting and real-time collaboration.