As predictive AI continues to define today and tomorrow’s ad industry, marketers still have a lot more questions than answers.
So let’s look at a few of the more technical questions about our Dynamic Intent Predictive Audiences, including:
How does ZeroToOne.AI measure prediction accuracy?
How do we keep predictions from getting outdated?
Which data points are used to point to future action?
How does ZeroToOne.AI create +90% accurate predictions?
If you’ve read our other resources on Dynamic Intent Predictive Audiences, you might’ve seen the “+90% Accuracy” number mentioned a few times. When we call predictions +90% accurate, we mean that the predicted behaviors described in each audience will occur more than 90% of the time for every single consumer in that audience.
Let’s use our QSR Diners – 24 Hours audience as an example. This audience contains consumers predicted to visit a QSR, purchase fast food in-person or via delivery, or exhibit on-the-go dining activity. So, more than 90% of the people in the audience will actually engage in one of these described behaviors within the 24-hour window. Because the audience refreshes every single day, that means a fresh slate of QSR diners every single day, of which at least 9 out of 10 will actually grab quick food.
How do we know the predictions are actually +90% accurate?
The answer is all in ZeroToOne.AI’s Large Behavior Model. Every single consumer you reach using our Dynamic Intent Predictive Audiences goes through a process in which ZeroToOne.AI “runs back the tape” and verifies whether or not the prediction was accurate. For more than 90% of consumers in our predictive audiences, the predictions are observed to have actually occurred, whether that means an in-person purchase, event attendance, online conversion, or other real-world behaviors.
How are audiences refreshed?
This won’t surprise any marketer, but most people don’t follow the exact same routines every day. Most people don’t go to the same restaurant every night, or purchase a new laptop every day, or buy a car every weekend.
That’s why we need audiences that incorporate the complete data on consumer behaviors, both online and offline. Dynamic Intent Predictive Audiences far outshine the competition here, with daily refreshes that reflect consumer behavior changes.
So let’s think about a restaurant visitor that goes out consistently once per week, on a Friday or Saturday night. A Dynamic Intent Predictive Audience with a 24 hour window will automatically include that consumer on Thursday or Friday, showing your ads at the exact moment of intent. Once the location or purchase data indicates that they’ve gone out for their weekly meal, they’re automatically removed from the audience until the next Thursday or Friday.
That’s how the automatic audience refresh works. Every day, Dynamic Intent Predictive Audiences take the latest behaviors into account and dynamically add or remove consumers, keeping the audience fresh, free from overfrequent ads, and reachable when purchase intent is at its peak.
Which behavioral signals actually point to future action?
ZeroToOne.AI’s Large Behavior Model makes decisions about who to include in our predictive audiences using a blend of both real-world and online behavior data, including:
- Location visitation: where people go in the real world.
- Consumer interests: hobbies, habits, and day-to-day behaviors.
- Transaction patterns: what people purchase and when.
- Digital engagement: what people watch, read, research and engage with online.
- Demographics
The model continuously evaluates thousands of data points across these broad categories, ultimately scoring individual consumers on their likelihood to take real action. If that score reaches a high enough threshold, the model automatically adds that consumer to the relevant audience. If it doesn’t, they aren’t included until they meet the necessary threshold.
Try Dynamic Intent Prediction Right Now.
Conventional audiences ask us to depend on what happened yesterday to determine what might happen today, whereas predictive audiences take these signals and transform them, giving marketers a confident idea of what their audiences will actually do.
Our Dynamic Intent Predictive Audiences are available to every single GroundTruth Ads Manager user. As predictive AI audience targeting sends the advertising industry into an entirely new era, the early adopters stand to reap the biggest benefits, so the best time to start is now.
Don’t take our word for it, try Dynamic Intent Prediction for yourself. Create your GroundTruth Ads Manager account today, and start exploring our platform (no minimum spend required). And, if you have any questions, don’t hesitate to contact us.
