Introducing new innovations in Google Play measurement accuracy

By Elad Mashiach

Over the past few months, we have been working with Google to help improve the accuracy, security and comprehensiveness of data we provide to mobile marketers and developers.

Today, we are proud to reveal the next step forward in mobile data accuracy, reliability, and transparency: Enhancing mobile performance, data accuracy, and security.

Over the last few years, as mobile spend has increased, so have the capabilities to simulate user interactions.

Unfortunately, some fraudsters have found innovative ways to compromise mobile devices, resulting in fraudulent attribution.

Below are a few of the new steps that are now being implemented to secure your data and ensure total data accuracy.

Introducing install start time

Typically, marketers have been able to measure only the click time and the first-launch time. When analyzing their data, marketers and attribution providers used the first-launch time as a proxy for install start time when measuring their click to install time (CTIT) or mean time to install (MTTI).

With the introduction of install start time, marketers gain a powerful new signal that opens a number of new opportunities:

1. Optimize your conversion rates and customer experience

By measuring the time from click to install start, from install start until first launch, and then from first-launch throughout the onboarding experience, marketers, product managers, and developers can improve their conversion rates and their customer experience.

For example, a heavy 600MB app may have a better conversion rate if some of that app weight is moved to a background download during a tutorial.

2. Combat fraud

Traditionally, very short CTIT indicate that malware is likely injecting clicks during the install process, effectively hijacking install attribution by sending the last click before the first-launch. Similarly, very long CTIT times indicate click flooding or click spamming.

However, now that we have access to actual install start times, we may be able to reject clicks that occur after the install start time. Similarly, we may find new insights by benchmarking both CTIST (click to install start time) and CTIT (click to install time – based on first-launch).

Securing referrer data

In the past, attribution provider SDKs used a “broadcast receiver” function to collect referrer information. While the vast majority of the time, the data recorded by the receiver was from Google Play, there was the possibility that a bad actor could send false referrer data to the receiver.

This referrer data often plays a crucial role in attributing installs.

We have been working with Google and provided input as they were building this new Play API. The new data flow allows a direct sync with Google Play improving the integrity and security of the data.

Please be sure to install the com.android.installreferrer in the new Android SDK to enable this security update.

What this means for you

At AppsFlyer, we live and breathe data, obsessing over data accuracy, integrity, transparency and usability. This is what we do all day, every day. This is why the largest brands in the world rely on us for their mobile measurement, and we take this responsibility very seriously.

Over the next few weeks, we will continue to work to better understand and apply these new capabilities across our suite of solutions.

We will also continue to collaborate with our partners to continue to provide best-in-class measurement solutions, and most importantly, to improve your marketing impact.