Founded in 2015, GeoZilla is a family safety solution that connects wearable devices to a single platform and strives to make families all over the world safer. With over 4 million registered users, the California-based company uses real-time location tracking and emergency prediction algorithms to notify users if a family member, friend, or pet deviates from their day-to-day routes. Unlike other location tools, GeoZilla has a unique combination of geolocation technology and machine learning to indicate when loved ones have arrived at their destinations or might be in need of help.
GeoZilla’s app, Family GPS Locator, uses subscriptions as its primary revenue source. However, as is common with most subscription apps, GeoZilla saw that non-paying users had low engagement with their app. To solve this, GeoZilla used a mixture of AdMob smart segmentation and ad placement A/B testing to show ads only to non-paying users, improving their day-one retention by 60%. The amount of users who created an account also grew by 120%, with 131% more users going on to complete the full onboarding, and no negative impact to their total revenue.
Balancing their business model
GeoZilla initially relied on their premium, in-app features to increase retention and engagement, and drive more subscriptions — such as “Driver Protection,” which sends alerts to parents if their kids get into trouble while driving. As with other premium features, “Driver Protection” was initially only available to subscribers. The team wanted a way to let non-subscribers gain access to these features in hopes of driving them to subscribe, but without compromising their subscription-focused business model or showing all their users ads.
The team decided to invest in AdMob smart segmentation to supplement their premium features, and personalize their users’ ad experiences. With smart segmentation, GeoZilla leveraged Google’s machine learning to automatically segment users into predicted subscribers and non-subscribers, showing rewarded ads only to predicted non-subscribers while leaving the predicted subscribers experience unchanged.
A/B testing plus Smart Segmentation
GeoZilla’s first step was to create a new ad unit, and turn on the smart segmentation toggle within the AdMob UI. Next, GeoZilla tested three different user flows to find the best rewarded ad placement.
First, the team split their users into two test groups — planning to show both groups the same ad unit and ad prompt, but at different times in the onboarding process. The moment GeoZilla received a new user, the smart segmentation tool predicted if this user would subscribe or not. If they were a predicted subscriber, they went through the onboarding process as normal. If they were a predicted non-subscriber, they would be presented with the offer to watch an ad in exchange for access to GeoZilla’s premium features for 24 hours.
Altogether, the experiment was made up of three cohorts:
- Predicted non-purchasers who saw the ad after closing GeoZilla’s paywall screen.
- Predicted non-purchasers who saw the ad on the paywall, next to a subscribe option.
- A control group of users without smart segmentation who experienced GeoZilla’s normal onboarding experience — ending with the paywall and no ads.
GeoZilla tested the behavior of the first two groups against the behavior of the third group for a week, and measured retention, accounts created, and percentage of users who completed onboarding. The team also recorded how many users performed key actions on day one of app installation — such as sending an invitation to share location — in order to gauge user stickiness. “Throughout the test we received a lot of support from AdMob that helped us make improvements, and showed us which group was performing the best.” said Pavel Shikhov, Senior Product Manager at GeoZilla.
Boosting retention by 60%
With smart segmentation, GeoZilla had a new way of growing their engagement and retention through increased subscriptions, without affecting the app experience of paying users. After a week of tests, GeoZilla saw significant improvements in group #2, with 120% more users creating an account and a 131% increase in onboarding completion. They also saw a 60% uptick in day-one retention, with day-seven and day-15 retention also doubling.
The team also saw no negative impact to their overall revenue.
The GeoZilla team is thrilled to see that AdMob can provide the solution they were looking for without compromising their original subscription-focused business model. “We really liked the expertise and support of working with AdMob. We were able to dive deep into our problems and improve our metrics, without impacting our main source of revenue” said Pavel.
As an added bonus, GeoZilla was able to understand their users from a geographical perspective. “We learned that users in some areas prefer to watch ads in exchange for access to all our features; whereas users in other areas prefer to watch ads for access to only a few specific features. This all helps us understand our user base further” said Pavel. With GeoZilla now able to provide personalized value to their global user base, their next plan is to continue leveraging machine learning tools to grow their subscriber base.
