Google Analytics Adds Predictive Features for Advertisers

Google Analytics has announced their new feature for Predictive Audiences and Predictive Metrics. These are part of their Beta for App + Web properties, and not available in regular Analytics situations.

Predictive Audiences

Located within the Audience Builder, users will now see new audiences, based off two audience types: likely buyers and likely churning purchasers.

These spawn specifications for things like “Likely first-time 7-day purchasers” or “Likely 7-day churning purchasers.”

This new audience feels a lot like Facebook’s conversion goal, which similarly targets users’ likely behavior.

Like other Google Analytics audiences, linked Ads and Analytics accounts will be able to import these audiences for bidding and targeting.

Predictive Metrics

Also launching are predictive metrics, which can help you analyze data in ways that may better help you shape your marketing campaigns in the future.

Located within the Analysis module, its aim is to show you things like which marketing campaign gave you users with the highest purchase probability. In other words, what are the things contributing to building those Predictive Audiences.

Requirements for Predictive Metrics:

• Equal populations for users who bought and users who did not. The minimum size is 1,000 users in each bucket.
• Sustained audience quality over a period of time, versus a situation that fluctuates wildly.

The metrics for each model will be generated by user, once time per day. If the model quality doesn’t meet the above, the predictions will become unavailable.

Why This is Notable

One of the divides that media buyers account for as they run on multiple platforms is how “ready” the users are to buy.

In a medium like search, the user is giving indicators as to where they are on their buying path. Are they researching? Are they close to buying? While some searches can reveal that, in many cases, there are search phrases that don’t really say either way.

The fact the user is searching means they are at some level of awareness, though.

Facebook ads, on the other hand, are the opposite. Beyond interests and lookalike audiences, there isn’t much to know where the user might be in relation to buying. One of the most powerful features of Facebook ads has long been the ability to optimize for conversion.

You tell Facebook what type of action you’re looking for, who you’re targeting, and off it goes to drive as many of those actions as possible.

Google has become much  more robust over the years with its targeting, giving search marketers the ability to go beyond just keywords and actually layer in things like interests. This provided additional insights and control for media managers who want to bid more specifically on different user personas.

“You can use these metrics to help drive growth for your business by reaching the people most likely to purchase and retaining the people who might not return to your app or site via Google Ads.” – Steve Ganem, Sr Product Manager, Google Analytics

This addition will help media buyers perhaps not feel so reliant on Facebook to deliver this type of targeting in the future. With skyrockets ad costs and longer purchase cycles, this will provide an interesting test for marketers looking to test their ad dollars elsewhere.

Google’s announcement is here, and they also provide step-by-steps for audience creation, predictive metrics creation, and information for signing up on the App + Web property beta.

Images courtesy of Google