*Artigo publicado originalmente no Medium. Acesse aqui.
Now more than ever, businesses of all kinds need to learn from their customers and use this knowledge as the feedback that drives innovation. “Big data” rings some bells when analytics is mentioned and everyone wants to go all in. However many of those methods should be used only when there is a need for it and not just because we are hype-maniacs. These tools should aid us drive valu to our customers and businesses.
It corresponds to a state in the development of a product that its creators want to add all tracking services available on Earth. It sounds like a joke but it is closer to reality that one might think. Reasons for that may vary from top-down decisions, poor judgment, misconceptions and so on… To illustrate this a bit, suppose you’re creating a mobile application. First you will integrate an analytics sdk to track crashes in your app PLUS one to monitor your network and because you’ll certainly have push notifications so lets add another one to do that. Google Analytics? Sure… why not? But using it directly is old school. Let’s bring Google Tag Manager with it. If you find yourself on the huge competitive ecommerce market where margins year after year are shrinking you cannot let any space for uncertainty. That means: you must make sure that the marketing budget is being spent right. So a marketing analytics SDK. Fair enough! But that is not all. None of them can be trusted to track your R.O.I. during marketing campaigns. So far we have 6 different analytics in our application, but wait there is more. Wouldn’t it be nice to do some sort of experimentation and test different versions of a feature to a segmented user base? To find out what works best for a different target? Another analytics sdk, this time for A/B Testing, congratulations! You have seven different sdks inside your app a magic number, things are for sure on the right track.
The big secret
Funny as it seems, is not difficult to find cases of apps doing a lot of tracking without any learning from it. This may be surprising to a lot of people but the reality is crystal clear: data only matters if it is being used to drive change and consequently generates value. If it is not the case your are just wasting time and money .
What are you saying ?
So is this just a waste of time? Should I do other things instead of tracking all this information? Of course not, this is valuable information that may be used to really improve your services and products. However, the way things are being tracked is fundamentally wrong. Suppose that push sdk would need information about your users that must be provided by you on your application, in order to, lets say, create different targets on the user base. Doing so is extremely important to effectively create push campaigns and re-engage users on your app. The A/B Testing framework needs this segmentation too, and many of the other analytics that you have also need this information . So your doing the same thing again and again to provide data for those analytics services that claim to be distinct. So what? A bit of repetition won’t kill, right? Wrong! Checking data across N different dashboards on different services is not the funniest thing to do. Am i just crazy? Am i the only one who thinks some of those different analytics sdks are just mere features wearing a product Suit ?
Fortunately for us, different sdks are reaching the market. Approaches like liquid or leanplum, just to name a few, are driving down the logic road. Push, Retargeting, App usage and similar things should not be separated in several different SDKs, increasing app development complexity. Analytics should help us to make better choices not going crazy. This is not to advocate the use of only one tool but to be aware that seven and counting for sure is not helping that much.
So you are saying …
The future about analytics usage is changing fast, and we should expect good things to come, hopefully we won’t have to handle and evolve 10+ different analytics dashboards at the same time .