Introducing AI Personalization (AIP)

What if, instead of making 36+ variations of digital ad creative for a single campaign, you could supply just a handful of graphic layers and text variations and let AI take the wheel to assemble and serve the best-performing creative to your audience?

This is AB testing on steroids, and it’s called AI Personalization (AIP).

It’s not personalization like “Hi [first_name], you have [points] to use on your Visa card by [date].” This actually serves ads using AI decisioning for every impression that gets served to your audience.

You may have done AB Testing in the past to see if your audience responds better to creative with a serious versus playful voiceover tone, green versus blue call-to-action buttons, or imagery of different geographies and people. When your target market is full of variation, your creative can now reflect those tiny differences in a scalable way. As AI engagement models have been trained over the last couple of years, it used to take well over a month to learn if slight creative differences created a lift of engagement and only worked for always-on campaigns where you had hundreds of conversions. This made it difficult to use AIP for monthly or promotional campaigns where it could be extremely useful. Thank the speed of AI machine learning for cutting the wait time for results!

For a single display campaign, you would normally create the standard five display ad sizes that basically all look like they’re from the same creative family. It’s pretty taxing on your creative team to create extra sets of display ads just to test a single variable, let alone multiple variables.

When clients are interested in testing AIP, we help them map out a content hierarchy.

Once a concept is set, we help templatize the ad structure by creating a wireframe.

Then we feed the AIP engine with content and creative assets to assemble different test variations in real time.

When ads are served, AIP collects the engagement information and informs us of the impressions that are converting at a higher rate. Then, it programmatically serves more like that in your active campaign period. Results are available in real time, so you don’t have to wait until the campaign concludes to gather useful insights for creating your next campaign. You may learn that not all creative converts at the same rate across different regions of the U.S. AIP allows us to tap into those granular insights that typical AB testing can’t provide. If you’re interested in testing AI personalization in a future campaign, let’s talk.