Applications of machine learning have brought us to the era of predictive analytics and marketing in which bids, targeting and ads are optimized based on predicted results. The most recent sign of the in Google Analytics is the inclusion of predictive metrics and viewers in Program + Web properties.
You’ll see hints for crowds using either of these metrics, when building audiences in an App + Web land in Google Analytics:
- Purchase Probability: Users marked as most likely to make a purchase sometime within the next seven days. The notion here is to have the ability to identify and target more than simply cart users and reach those people that didn’t put an item in their own carts but are seen as likely to generate a prospective purchase.
- Churn Probability: Active users deemed unlikely to see your website or app in another seven days. These are website visitors who are not very likely to come back within the week to your site on their own.
You will find two predictive audience options based on these metrics that are new:
Based on these new metrics, there are two predictive audience options:
- Purchase Probability crowds: Probably 7-day purchasers and arguably first-time 7-day purchasers.
- Churn Probability audiences: Probably 7-day churning purchasers and Likely 7-day churning users. Give site visitors that are past reason.
Assess this using predictive metrics. Using the Analysis module, you may use these new predictive metrics to construct reports such as understanding which advertising campaigns brought in consumers deemed to have the maximum Purchase Probability.
Why we care. Machine learning-driven audiences are not new in Google Analytics — Smart Lists debuted back in 2014 — but those updates offer more options and flexibility. Expect more to come. Google states the new metrics and viewers will soon be available in the forthcoming weeks. You will have to have purchase events set up or automatically measuring purchases in your App + Web land.