What does a bloody third-person shooter game about bank robbery at gunpoint have in common with a hyper-casual game where you have to take a cartoon cartoon bacon out of a frying pan to try and land it on random objects?
The answer: Not much, but both are classified as “action” games, according to Apple and Google in their respective app stores.
(“Bacon: The Game” is real, by the way, and it’s oddly satisfying.)
Privacy changes on iOS and Android, namely Apple’s AppTrackingTransparency framework and the SDK Runtime API coming soon to Android 13 (part of the Android Privacy Sandbox), are making contextual targeting within the app stores increasingly important.
But contextually may not be as effective for targeting or monetization if publisher apps are categorized too broadly, said Puneet Gupta, co-founder and chief product officer of Kayzen, a Berlin-based startup that helps developers make their mobile programmatic purchases. to take. House.
wide strokes
Most game advertisers are developers too, and the purpose of advertising in other games is to encourage their own downloads. User acquisition managers are deeply analytical and data-driven individuals who always have performance in mind.
But the categories offered in Apple’s App Store and Google Play Store haven’t changed much since both stores launched nearly 15 years ago.
And it’s very unlikely that the players of ‘Armed Heist’, the third-person shooter game mentioned above, are the same people who play that very casual bacon game. Yet these audiences are lumped together because the games they play are broadly categorized.
To be fair, the app stores also offer some extra tagging based on subgenre, but this isn’t consistent for iOS and Android.
“They just don’t go deep enough,” Gupta said. “They’re limited and the real context isn’t there.”
Smoothing out the details
Earlier this year, Kayzen launched a new tool that uses a mix of manual tagging and machine learning to categorize apps more accurately and with more nuance.
The feature, called smart app categorization, classifies publisher apps based on a hierarchy of three levels: category, genre, and subgenre.
Take “Armed Robbery”. Instead of choosing “action” as the primary category and calling it a day, Kayzen can label it as “action” or “core.” (Core games are games that generally require a certain level of skill and practice and are often monetized through in-app purchases.)
In this case, the genre can be “shooter” and the subgenre “third-person shooter”, to distinguish it from first-person shooter games.
In comparison, a possible hierarchical division for the bacon game could be “hypercasual” as the category (because bacon flipping really isn’t an action game however you want to divide it), “arcade” for the genre and perhaps “agility” for the subgenre.
After Kayzen manually tagged a large number of apps herself, Kayzen used them [specify what them is here] as seed data combined with data from third-party partners, including Sensor Tower, to feed a machine learning model as a way to automatically tag other apps.
The apps Kayzen has been able to categorize represent anywhere from 70 billion to 75 billion programmatic ad requests per day. Kayzen is also working to expand the number of categories, genres, and subgenres it has to offer.
On a roll
JustDice, a German developer of award-winning mobile game apps, has been working with Kayzen for a little over two years.
Since justDice started targeting campaigns based on smart app categories this year, justDice’s install volume has grown 2.5 times and the number of apps installed per thousand impressions has grown six times.
All of justDice’s apps are rewards-based, meaning discount hunters are one of the most active and lucrative audiences. But to find that particular type of user in the past, you had to rummage through the app stores to find titles that resembled their own titles and to create whitelists, a process that wasn’t automated at all.
Now justDice can target apps that have “rewarded offers” listed as subgenre.
This “really provides a significantly better context in targeting,” said Miika Kenttämies, managing director at justDice.
While programmatic isn’t the largest media investment, justDice controls everything it does in-house, which is why it started working with Kayzen primarily as “an extension of our in-house team,” Kenttämies said.
“With DSPs, we need 100% transparency rather than a managed black box,” he said. “We see running it in-house as a significant added value to our other media purchases because of its transparency and ability for hyper-targeted campaigns across exchanges.”
That’s one way to take the bacon home.
0 Comments