Google test surfaces user data in DoubleClick for Publishers as part of new Insights Engine Project
Several new initiatives are aimed at bringing machine learning into publisher products and offering solutions for driving subscriptions.
Google is laying the groundwork for turning the love-hate relationship many publishers have with the company into a love-love relationship. In an event for publishers hosted at the Google offices in Chicago Tuesday, executives laid out several new initiatives aimed at extending a helping hand to publishers across multiple aspects of their businesses.
“This is not some secret disintermediation strategy,” said Philipp Schindler, SVP & chief business officer, at the outset of the event. ”We are all in this together … and we are all invested in seeing you succeed.”
Among the announcements was the launch of a new effort to share user data with publishers and introduce more machine learning into Google’s publisher products.
Insights Engine Project
Jonathan Bellack, director of product management, introduced a new effort called the Insights Engine Project.
[blockquote cite=”Jonathan Bellack, Google director of product management”]Great user experiences today means mobile. And increasingly, mobile advertising means native advertising.[/blockquote]
The new Insights Engine Project “is not one thing,” Bellack explained, but an ongoing effort and commitment to bring machine learning into Google’s products for publishers. Google is kicking off the new initiative with seven new features:
- For the first time, Google is making user (aggregated and anonymized) data available to publishers with insight cards. Publishers will be able to see and create audience segments based on demographics and interests, including in-market behaviors of their readers. Audience insights can then be used in programmatic guaranteed deals made through DoubleClick Bid Manager. These efforts are currently ongoing but involve very limited tests.
- New insight cards in DoubleClick for Publishers are powered by machine learning to show ads insights and benchmarks relative to competitor publications . The initial benchmarks will be for latency and viewability metrics.
- The ability to import DoubleClick data into BigQuery, the data warehouse product that’s part of the Google Cloud offering, is currently in beta now and will be available generally next year.
- Publishers will be able to import DoubleClick data into Google Data Studio to create and share visual reports. This is also currently in beta and will roll out early next year.
- The company has updated its forecasting engine with more machine learning capabilities to automatically detect what’s happening with ad sales growth, including seasonal expectations, and adjust forecasts automatically. A new visualization in the DoubleClick UI will show publishers past and future ad revenue estimates on a continuum.
New test to support subscriptions
Yesterday, Google announced it was ending the First Click Free program, and replacing it with Flexible Sampling. Google announced Tuesday that it is in an early testing phase of applying machine learning and audience insights to see how it can drive subscriptions for publishers.
The test uses machine learning to detect readers who are likely to be candidates for subscribing. Rather than serving a standard ad (or ads) to those users, DFP would instead serve them house subscription ads.