AppLovin Data Desk 4: Moneyball for Mobile Advertisers: Using Data to Maximize Returns

by John Krystynak on Mar 31, 2015

Remember Moneyball, the story of how Oakland A’s general manager Billy Beane transformed his team through a new approach to statistics? Beane took a fresh, more analytical approach to the entrenched statistical indicators that for decades had driven investment in players, and by doing so he found players that were undervalued in the market and won more games than teams with double the payroll.

Moneyball is a lesson in the use of advanced metrics to uncover opportunities in the market and capitalize on them accordingly. It showcases just how valuable data and gleaning the right metrics can be in finding hidden gems that give you an edge over the competition. 

At AppLovin, we know that lesson is highly relevant in mobile marketing. Here’s the data analysis that we did to back up our theory that applying the Moneyball philosophy can have a dramatic impact on your bottom line.

Moneyball Method

When our partners share engagement data, it gives us a better view of the “full marketing funnel.” We can see sales, bookings, in-app-purchases (IAP), etc., and that allows our platform to better optimize for the highest returns. One way we do that, like the Moneyball approach, is to uncover “sleeper” variables that offer an advantage by finding the most valuable users and then re-engaging with them.

Our findings: the undervalued players

Our sample size of thousands of completed transactions garnered from over 37MM impressions during the month of January uncovered a few interesting trends that marketers can take advantage of to better target valuable users for less. Here are a few of the top examples:

For iOS casual gaming apps, there are app categories, such as photography, that may not intuitively be a good match for casual gaming ads, but that the data indicate are inexpensive and actually perform well. Therefore the cost of a re-engagement, in this case of a user completing an in-app purchase (IAP), is relatively low on a photography app for a casual gaming advertiser. For example, a casual player retargeted in a photography app purchased 64 percent more in a casual game, while it only cost 10 percent more to serve the ad to that player than to a player retargeted in another casual app.


On iOS, eCommerce ads served on Brain and Puzzle apps outperformed, in terms of CPS (in this case a product sale), those served in any other category, including other eCommerce apps. Data showed that ROAS was 23 percent higher when eCommerce ads were served on Brain and Puzzle apps. The users on eCommerce apps were more expensive to target because they were in higher demand, yet they didn’t purchase enough to compensate for their cost.


It’s not just categories that provide opportunity — geography, ad format type, gender and other variables can yield valuable optimization points that might not always be intuitive. For example, in the Travel vertical on Android, U.S.-based travel apps often target U.S.-based users. However, when re-engaging consumers (getting them to come back to the app and book), the data indicate a huge opportunity to focus on users in Canada. While U.S.-based customers re-engaged more, they were also more expensive — about double the cost. In Canada, the Travel apps re-engaged with more consumers for less, yielding an advantage of 49.5 percent.


How to spend less and win more in mobile marketing 

There’s no doubt that intuition has its place when it comes to marketing, but there’s incredible value to be had in technology that can uncover patterns that you’d never put together on your own. If you ignore what technology can do for you, you risk entirely missing re-engaging with valuable customers on mobile.

We have one customer, a large eCommerce company, that improved its ROAS by 250 percent over 1.5 months after it shared its data with us. It shared data on product sales that were driven by the ads served. That meant we had the information needed to see where it was more relevant and cost effective to advertise, which then  allowed for better optimization for user re-engagement campaigns across all categories, geographies, ad formats, etc. (not just the ones we and the eCommerce company assumed would be profitable). This eCommerce company understood the importance the data played in fueling its decisions, and ultimately it reaped a successful outcome.


The Moneyball of mobile marketing is constantly evolving. The variables change on a millisecond by millisecond basis. It’s impossible to make calls based on gut reaction in a millisecond. The best way to optimize a campaign’s learnings and engage quality users for less is to first recognize how important the data are. Understanding the full funnel of data allows for the technology to better optimize, and that means better results. Never forget the secret to the eCommerce company’s 250 percent win: Share the data and let technology do its job.

John Krystynak is AppLovin’s Chief Technology Officer.