App Based User Acquisition (UA)

Precision Targeting with Apps

Try the Appfluencer® Targeting Demo

Change the Filter:

Select Hay Day in UA Campaign and Set Category to Health & Fitness.

Reading the Chart:

The vertical axis measures Lift (see below for a detailed explanation), while the horizontal axis lists the apps being compared. The chart contains all the apps that appear on all the devices with ‘X’ app. Because we have adjusted the filters to Hay Day and Health & Fitness, we see how the apps along the horizontal axis impact the presence of Hay Day on the device.

Measurement and Output:

On a daily basis, we measure the probability that an app is on the device because another app is on the device first. Lift is used to measure the impact one app has on the presence of another. A Lift value close to ‘0’ means that the apps appearance is unexplainable and the apps appearance is nothing more than a random event.

Apps with the farthest value away from ‘0’ (zero) have the highest impact on probability.  Apps with the highest positive values are the best UA targets while Apps with the highest negative values are to be avoided.

Interpreting the Chart:

Identify the app with the highest Lift value closest  to the vertical axis (furthest left in the view) – ‘FitnessClass’.  It has a lift value of 0.3, in this instance, if you were targeting new users for Hay Day, you would target devices that contained ‘FitnessClass’.  Conversely, two apps to the left, ‘Fitbit’ has a value of -0.18.  This isn’t to say that some Hay Day users do not have Fitbit, more accurately it states that Hay Day users probabilistically do not have Fitbit. Some do, but most don’t. This is an example of an outlier.

[wpdatachart id=”113″]

What’s Behind the Appfluencer® targeting and Scoring Algorithm

On a daily basis we perform over 24 Trillion measures, comparisons and calculations.

At a very high level – this is the process used to determine target values for the Appfluencer®

  1. Calculate the probability that an apps <app x> appearance on the device is a random event
  2. Calculate the probability that an app <app y> appears on the device because another app <app x> was on the device first
  3. The next step is to calculate the result of step 2 among every pair of apps that exist to come up with a set of ratios
  4. After more calculations, a final comparison is made, another ratio results and final values are determined – one of these values, the lift value, is the key value used in targeting.
  5. The Lift value is used to measure the impact one app has on the presence of another

On a daily basis, we determine the relationship between every pair of apps and the influence each app has on all the apps that appear together on the devices.

We utilize BigData processing but more importantly Smart Data Analysis to identify actionable intelligence used in Appfluencer® Targeting & Scoring Algorithm, Appfluencer® Market Research and Appfluencer® Actionable Intelligence.

Chart Style

[wpdatatable id=”127″]
Contact Us

We're not around right now. But you can send us an email and we'll get back to you, asap.

Not readable? Change text. captcha txt