Good practices on mobile monitorization
We in Barkibu, our online petcare platform, are trying to improve our new mobile apps by the usage of monitoring. In this post we are researching about this monitoring topic, also including the reaction to failures.
First of all we should describe the two cases that we’d like to cover:
- Monitoring: We are thinking here about the events that occur inside our apps, so that we can know exactly how’s the customer usage.
- Crashes: This is a very important topic to exactly know when an error happens on production, in order to fix it as soon as possible.
Monitoring
The key for monitoring success is the previous researching work, that we must plan with the other parts of the app as an entry point before even writing a line of code.
First of all we should know which KPIs(values that demonstrates how effectively a company is achieving key business objectives) would be the most important ones to focus on, having a list of them written and described (it would be a great idea to prioritize them if you have a decent number of them, as it would be quite difficult to work with such a big number from the beginning).
When having all the KPIs defined, we move to the declaration of the critical paths for our app. These paths may be done following an event convention, in which every single action you make is an event that we should monitor.
An important thing about these events is retrieving information both from the events and from the user data when the event happens, for example saving an user identifier when the log-in event occurs.
Here you can see a Google example of how they create their event taxonomy.
When this is all done and implemented in your code, it’s time to investigate about the results that the data is giving us. You can do many things with this data as you can tell from these examples:
- Create behavioral cohorts.
- View your critical paths and increase conversion with funnel reports.
- Calculate user retention over time.
- Run experiments, like A/B testing.
- Measure the impact of new feature releases.
Crashes
On the other hand, we must capture every error on our crashing analytics system. To do so, is important to have some quality error reports, focusing especially on:
- Making it as real time as you can.
- Grouping each error on a describing group, to rapidly identify the issue.
- Prioritize each error in order to know which one to handle first.
We are hiring!
In this post we’ve seen the surface of what monitoring is and how a good system enhances the capability of our apps. Are you interested in app monitorization?
