Proper App Tracking Setup: Funnels


For the majority of app developers, it is not new news that tracking is a must for any mobile app traffic campaign. Though not everybody knows how to set up this type of tracking properly. In this article I want to talk about how to set up tracking correctly.

If we take the list of mistakes that app owners or developers make, there is one that stands out from the rest. This mistake is when developers judge their app traffic quality by how well it converts to a final event, for instance, a purchase in a short period of time. In real life there can be many events that take place before people proceed with a final action including a purchase, sign up or any other action that app owners consider as a major one. For example, when people download an app for booking hotels they search inside of an app, alter the search criteria, click on different hotels to compare, then end up with a single hotel to book.

The fact that many events may take place inside apps brings us to a funnel notion. To define a funnel for your app, you need to do two things: define a complete list of events you’re going to measure and when you’re going to measure it (i.e on the first day, after 3 days, a week, a month, etc.). As you measure the progress with your app’s funnel, you’ll be getting a graph that shows the number of app users who have taken a specific action on a specific date.

It’s really important to understand that as an app owner or developer you can’t control how much time people will spend in your app and when. You can’t expect that your app users will take action right after interacting with your mobile traffic campaign text, image or video ads. An example below demonstrates how many users have taken a specific action (booked a specific hotel) inside your app on a first, second, etc. day within a week’s period.

Deciding on what day, after your app was downloaded, you expect people to take a specific action is crucial to this analysis. It’s always good to be able to measure any event that demonstrates that you’re getting quality users or people who may purchase later on. You can decide on the traffic quality you’re getting early on as well. Again, that event most likely won’t be a purchase but will show you how likely your app users, who’ve gone through that event, will purchase inside your app. On the other hand you need to wait for a significant amount of accumulated data before making decisions on the traffic quality.

The next important thing to remember is that you need to compare different traffic sources within the same time period.

By looking at the example above you may get an impression that the second source is better, but if you look at the numbers for the same two sources after a week you may discover that the first source is better, but a day period wasn’t enough to judge the traffic source quality.

The next important point is that you need to count a number of unique app users who took an action, not just a total number of events. Consider the following example:


Compare the table above and below. If you don’t count the number of unique users that paid inside your app, you may be mislead and think that the second traffic source has brought you half of what the first one brought. In reality the second one isn’t worse, those $500 were brought by 10 users, while the first source brought you only two paid users. Clearly the second traffic source has better potential to bring you more profit.

Let’s move to the next important question of how much information is enough to judge a traffic source? The simple answer is that it depends on two things, what conversion (ratio between installs / paid users) you have for your app and how many installs you’ve got. The less conversion the more installs you need to judge a traffic source properly.

It’s important to make sure you track events properly. Remember that any tracking system is based on a general approach to events tracking and it may track events inside your app incorrectly. To make sure you’ve set up tracking correctly, launch a small test campaign.

Traffic type and LTV (lifetime value) are the last crucial pieces to factor in. When you compare paid and organic traffic sources you need to remember that the latter has better targeting options and higher potential to bring loyal users. You can think of Twitter and Facebook as two great examples of organic traffic with their superior targeting options. Finally, people don’t download apps for a day or two, they look for a solution for a particular problem and once they find an app that does it for them they stick to it. For example, lifetime value of travel apps can be measured within a year or more.

For more visit ComboApp’s website here.

Art Dogtiev,
Head of Branded Content at ComboApp, Inc.