Google added a slew of new solutions, ad formats and targeting options to its automated app ad platform and to AdMob, the company said on Wednesday.
And there’s a theme: Automation and machine learning are becoming increasingly central to Google’s vision for app advertising.
Advertisers can use automation to be more efficient with their user acquisition, while publishers can rely on it to help encourage better customer lifetime value, said Sissie Hsiao, VP of Google’s mobile app advertising division.
“Apps want to acquire users, but they also want to retain them, to see them come back to the app and be loyal for as long as it’s ROI positive for the business,” Hsiao said. “How can we use machine learning to effectively spend a marketing budget, and not just to encourage a single action – that is something I’m asking my team [at Google] to really think about.”
For buyers, Google is automatically extending its app campaigns promotion product to three new inventory sources: Discover, YouTube search and in-stream display via YouTube.
Discover (previously known as the Google feed), displays topics and news items of interest within the Google Search app and Chrome. Google first introduced ads into Discover, which reached 800 million monthly active users, in May.
App campaigns will also appear amidst YouTube search results on mobile. And for advertisers with video assets, it will appear as a skippable ad in-stream either before, during or after a video plays.
“It’s important to help developers and marketers find valuable users, and we’re always looking for ways to expand that product,” Hsiao said. “But generating effective monetization from those users is also very important.”
For the sell side, Google is rolling out a new ad format in AdMob that displays briefly as a splash screen while a user is waiting for an app to load.
“This is an area that has not previously been monetized by a lot of developers,” Hsiao said, “but it gives developers a brief moment to introduce an ad unit without being disruptive to the user experience, because the user is waiting anyway.”
The unit is now available in beta and eventually all major demand sources, including Google’s own app campaigns platform, will be able to bid for it.
On top of that, Google is including new reporting and measurement capabilities within AdMob, including anomaly detection that automatically alerts publishers to any weird fluctuations in their impressions, revenue or eCPMs, as well as an extension of its smart segmentation offering.
Smart segmentation uses a machine learning algorithm to split off users who are unlikely to make in-app purchases (IAP) so that developers can monetize them through ads without sacrificing IAP revenue.
Google is expanding smart segmentation, which was already available for interstitial units, to rewarded video, an especially popular unit in games, Hsiao said.
“This way, you don’t offer rewards to people who are predicted to pay,” she said. “That improves the experience for non-paying users by allowing them to engage in rewarded ads, and it helps developers not cannibalize users who do choose to pay.”