Learning about your static ads, with Zappi’s AI ad predictor
When creating digital ads, understanding what’s going to resonate with your audience is key. Zappi has been testing ads with large sample audiences for almost a decade now; for the biggest of the big global brands. For the first time, we’re helping everyone pre-test their ads — with the Instant Ad Predictor.
Based on a huge amount of historic data, a ton of features from your ad, and some pretty groundbreaking neuroscience-led predictions of attention, Zappi have teamed up with EyeQuant to help you find out how to improve your ads, almost instantly.
How to use the Instant (Free) Digital Ad Predictor
In just a few minutes, Zappi can predict a free, shareable scorecard, helping you diagnose issues with your ad and improve their performance. With a powerful combination of machine learning based on Zappi’s historic data, and EyeQuant’s neuroscience and AI, you get quantitative and qualitative results almost immediately.
Predictive eye tracking; simulate how users will react to your creation.
Modelling and interpreting an audiences gaze and attention has long been present in market research. But, with expensive rigs, or sketchy at-home webcam approaches, it’s not always been present in fast paced decisions. EyeQuant have turned this on its head.
How do they do it? A very abridged version…
- Quantify gaze patterns and eye movements — via laboratory studies — to determine the relative importance of visual features.
- Predict how the designs will be perceived. An artificial neural networks is applied to create a prediction of how users will perceive a design in the first three to five seconds of viewing. Using features such as luminance, edge density, saturation, and structural architecture, EyeQuant’s prediction is nuanced and intricate!
With compelling statistics (how does a 31% improvement in conversion rate sound?), and a list of clients that includes both Google and Meta — we’re pretty excited to bring EyeQuant’s heat maps and scores to you in the free ad predictor!
Predicting how your ad will be perceived
The Scores: Zappi uses a gradient-boosting machine learning algorithm to rapidly deliver a free prediction, based on a huge volume of anonymised and aggregated survey and norms data. What this means in layman’s terms is that a load of tags — both manually added and assigned algorithmically — are collected, cleaned, and fed into a decision tree model that then “boosts” important components, while suppressing less important ones. This is the XGboost regression model. We’re using a machine learning model that predicts key metrics based on the attributes of the tested ad, e.g. branding, celebrity/human presence, length, etc. The model is trained on ads tested on Zappi Video/Digital and the prediction falls on average within 5% of the observed survey scores.
Overall Appeal predicts whether the sentiment to the ad will be positive or negative.
Brand Score is A combination of how people feel about your brand, and how well your brand links into your ad.
Purchase Intent: how likely the ad is to drive behavior change, or make people purchase.
How to use your results
- Instantly get a basic scorecard with a few key measures. Judge how your ad might perform if you were to put it forward into full testing
- Use your EyeQuant heat map to make sure that the right things are getting the right attention. Is your claim getting lost in the noise? Is your brand name really jumping out of the screen? If not, get back to those designs!
- Get a few recommendations on how to improve performance. Making these changes, however simple, should improve your ad.
- If you want to learn more about your ad, carry on to Zappi’s full ad-testing platform, or head on over to EyeQuant!
Can I trust this model?
Yes, you can. It tells you based on history how strong your ad is. That is useful. And we constantly feed and re-train the model to make it better. But, it’s always going to miss some nuance — that’s what the full Zappi products are for. To pick up on the detail, you’re going to want to put your ad in front of real people to understand fully what’s going on!
So, fancy giving it a try? We’d love to hear how you find it.
