3 Levels of Understanding your Mobile Customers
Mobile analytics is about understanding your mobile customers. It seems that every day a new analytics vendor releases a “must have” analytics product for your app. Instead of trying or adopting every new product, I recommend mobile app publishers focus on a “mobile analytics maturity model” - a model that promotes building up your analytics capabilities along different axes.
This blog post describes the three levels of analytics in our maturity model, who in the organization needs to be involved, and which analytics vendors can help you get the metrics to enable better decision making for your apps to support your users.
Level 1: Marketing Analytics
Marketing analytics is the awareness of who is using your mobile app, and what they are using in terms of devices, operating systems and features. You should track basic information, such as number of users per month, average time on site, and basic conversion funnels. There are a lot of established tools for this: Google and Flurry provide excellent free tools that are loved and trusted by many. Omniture also provides compelling marketing metrics. MixPanel is a new player in town, and they have some really cool analytics too. Don’t forget customer reviews: they are a great place to get real user feedback on your app.
It is critical for the person who is in charge of ensuring the business success of an app (usually marketing or product management) to look at your marketing analytics on a weekly basis to ensure that the right app features are implemented to drive business success.
Level 2: Operational Analytics
Now that you figured out who is using your app, and on what devices, the next thing is to figure out if your app is working as expected. As the basic features get implemented, it will be good operations that separate apps that people love from apps that people hate. In a recent survey by Apigee, 98% of respondents said they would delete an app if it freezes, crashes, or is consistently slow. The problem is that most of these symptoms can be triggered by a network failure (spotty coverage, improper settings, optimistic latency handling in your code) and can go undetected despite thorough testing by your app team. To make matters worse, because app deployment is controlled by the app store, deploying fixes requires time. To overcome the latter hurdle, I recommend that major new code changes should be activated by configuration that is controlled on the servers’ side. That way, if something goes wrong, it is easy for the operations team to revert to the old functionality.
Even though popular packages such as Google Analytics and Flurry provide some level of crash reporting, they do not give the same level of insight as they do for marketing analytics. Tools such as Crashalytics, Bug Sense, Critterism, and Apigee Mobile Analytics have much deeper capabilities to make sense of application errors and crashes. In addition, AT&T’s Application Resource Optimizer and Apigee Mobile Analytics each provide network analytics to help you determine if your app is slow because of network issues or if your API dependencies are causing issues. Apigee Mobile Analytics also has configuration management to enable you to roll back your code.
Today, mobile operations is often the responsibility of app developers — but as apps become mission critical, it is imperative that IT & ops teams tak
e over this task to uphold service guarantees.
Level 3: Customer Experience Analytics
Customer Experience Analytics — the exact tracing of what a user is doing with the mobile app and when — is a great way to gain insight into what needs improvement and determine how to optimize existing features.
One of the biggest mistakes that mobile companies make is the false assumption that an app is not attracting users because it doesn’t have enough features. Unlike desktops, mobile screen real-estate if very limited, so the number of features that are used is also extremely limited. In addition, people’s attention span is short (usually less than 10 minutes per usage), so it is vital that they are able to complete their action within a small time window. Therefore, it is important to optimize every user interaction.
It is a mistake is to use custom events from Google Analytics or Flurry to instrument every single button and interaction. This creates an unnecessary flood of data that no one looks at. Tools such as MixPanel, Critterism, and Apigee Mobile Analytics have more elegant “tracing” capabilities to zoom in on the order of events that a user is taking. Heatma.ps is also another hot tech startup that provides “user interaction heat maps” to help you make sense of all this analytics data. Knowing exactly what your users are doing gives you the ability to take corrective action.
Giving access to interaction analytics to both your designers and developers is critical to make sure that somebody is aware of problems and acts upon them. Sometimes all that is required is for a designer to change the size of a button slightly to prevent “fat finger” taps, or for a developer to squash a bug that causes the app to slow down.
What about API Analytics?
Of course, collecting analytics on mobile devices only gives you part of the picture. Because virtually all mobile apps use APIs, collecting API-side analytics gives you additional insight into your marketing, operations, and customer experience. Collecting analytics only in the app tier will be increasingly incomplete as apps become more sophisticated.
Last thoughts on the human aspect of analytics: mobile analytics only matters if there is someone to look at the data and make sense of it. Finding passionate analytics users (whether marketing, ops, designers, developers, others) is probably even more important than finding the analytics vendor.
If you have any questions about mobile & API analytics, feel free to contact Alan & Kumar — we are happy to answer technical and strategic questions about your analytics programs.