Tracking our fears
Tracking (or analytics) hasn’t always had a good press; more knowledgable users are wary of mobile apps or websites that are 'too chatty', and seem to be constantly pinging messages back home to their servers.
Concerns can range from; 'how much bandwidth is being used up?' or; 'whether the app will slow down?' Through to suspicions about what the messages contain - 'what do they say about me?'
The fear is that your privacy is being violated in some way, that your data or online behaviour is being harvested and stored in the cloud somewhere, to be later used against you (cue sound of faint ominous cackle).
To Track or Not to Track
Some of these concerns can be legitimate; unscrupulous app or website owners might be trying to flout the law and have criminal intentions, but the vast majority of tracking is done with better motives.
Tracking user behaviour can be done anonymously (they don’t know it's you) and can be used to inform future app design - by seeing which parts of the app are used and how often, developers and designers can redesign areas that aren’t working and plan which new features to introduce.
How much tracking?
One of our apps already used a couple of tracking frameworks and so when we were asked to include yet another one, it got me thinking: How much tracking can be done without impacting the user? Is it better to write the app to use one framework and ask the client to produce their own reports from the tracking data?
My initial assumption was to standardise our tracking across the Android and iOS versions of the app and that it would be better to use just one tracking framework, probably Google Analytics, and try and export the data to the other frameworks the client was using.
I did some quick research to see what 'best practice' for app tracking was and whether it would be possible to integrate our tracking frameworks. I found the following VentureBeat insight report with three key findings that surprised me:
Less than a third of developers use Google Analytics
Despite Google’s massive market share, fewer than a third of mobile developers consider it their primary app analytics solution.
Most mobile developers and publishers use more than one analytics solution
Interestingly, the more successful developers are, the more likely they are to use multiple solutions. Big Fish Casino, for instance — the fourth-highest grossing app on iOS — uses no fewer than five: Flurry and TestFlight and BugSense and Tune and Kochava.
Racing Rivals, the 13th-highest grossing app on Google Play, uses five as well: Google Analytics and Flurry and BugSense and Tune and Amazon Insights.
Different platforms require different solutions
Mobile publishers who are hoping to have the same app analytics suites on both iOS and Google Play should think twice, the study makes clear. The platforms are different and require different solutions. Just one example is Apsalar, one of the outstanding analytics solutions on iOS and a VB Insight recommended pick on that platform.
On Android, however, the solution has been uninstalled 14,000 times for every 1,000 current installs — a huge red flag. Clearly, analytics solutions need to be evaluated on their own merits on each platform individually.
The report recommendation was:
“First: use one of the major platforms (Google or Flurry) as your default solution. And second, pick additional app analytics partners carefully, depending on what mobile development or marketing factor matters most to you right now: user experience, user acquisition, engagement, monetization, or pure analytical power.”
"Despite Google’s massive market share, fewer than a third of mobile developers consider it their primary app analytics solution. "
It was encouraging to see that other app developers were facing similar challenges to us and based on the design of the most popular apps, it wasn’t necessarily a bad thing to be using more than one tracking mechanism.
As we are using the same tracking solutions across our different app platforms we will need to keep an eye out to make sure we continue to use the best solution for each platform.
I suspect we will pick Google as our default solution and it has subsequently been suggested that Google Tag Manager could be used as a way of standardising our tracking, so I’ll investigate that and it might the subject of a future post.