Revision as of 20:50, 4 April 2012 by Brad
Funnels in Flurry Analytics
Funnel analyses track users as they execute a defined set of steps, allowing you to see how many of the users that started a given process completed it. For those that did not complete the process, funnels show you at which step they exited. All of this information can be used to adjust aspects of your app to maximize the number of people that reach and complete the last step, commonly referred to as "the goal" or "conversion".
With funnels in Flurry Analytics, any event you have setup in your app can be used to define a step. Questions you can answer with funnels in Flurry Analytics include:
- How many users that view the tutorial end up finishing levels 1, 2 and 3 within 3 days?
- How did our last change registration and confirmation process affect the number of people that successfully registered and started using the app?
- How many users make an in-app purchase within 1-hour, 1-day or 1-week after first launching the app.
- How many of the users that completed an in-app purchase in Fantastic Foragers did the same in Gawking Neanderthals?
- any other question you can think of given your custom event structure...
There are two types of funnels available in Flurry Analytics, "in-app" and "cross-app". As funnels rely on your custom events, funnels that focus on a single app ("in-app") are found within the Events navigation menu under the Applications tab. Funnels that span two or more apps ("cross-app") are found in the Conversions navigation menu under the Company tab.
Creating a new funnel involves configuring two or more steps, each of which leverages a single event you have defined for an app. The order of each step in a funnel is important as all funnels are calculated in the specific order in which they are configured. For steps other than the first, there are controls that allow you to specify both a minimum and a maximum time between steps, thus enabling time-based analyses.
Once created, Funnels will take up to 24 hours to populate with data. While this may seem like a lot of time, it is necessary as funnels within Flurry Analytics are calculated across your entire app history, not just on the events that come in after you create the funnel. The advantage of this approach is that you do not have to know every funnel you want to create at app launch. You can create them over time as different questions arise.
3) A chart showing the Conversion Rate, which is the percentage of users who completed the last step out of those who entered the first step. Also shown on this chart is the overall conversion rate. This is same figure displayed at the bottom of the Funnel Steps table (see #1 above) and allows you to compare periodic conversion rates against the overall.
As with other areas of Flurry Analytics, the Segments, Versions and Time filters are available for funnels. These filters can be used to refine the focus of the funnel and compare the conversion rates for different situations. For example, using the version filter, you could see if that change you made to the tutorial in version 4.6.15 really had the intended improvement in the number of users completing levels 1 and 2.