Today, it is common knowledge that choice of audience and media channel effects the performance of an ad. Our previous research in this area has led to significant improvements for ads in different channels, but we wanted to take it even further. We wanted to go beyond segmentation and channel selection to examine the effect of the ad itself. How much does it matter what picture you use or what words you chose for the headline? Where in the ad should the CTA button be, and does it matter what color it is?
We wanted to go beyond segmentation and channel selection to examine the effect of the ad itself.
A blueprint for the perfect ad
The main goal of this research project is being able to give live suggestions for ad images, text and font. To do this, we need to determine the relationship between these factors and CTR. So far, most of the work has been done focusing on the images. With data from the real estate industry, we use picture recognition to identify and tag ad images. The tag includes image content, but also other features, like brightness, colors and fonts. As of now, over 500 000 pictures is classified with more than 30 million combinations of pictures and tags. We continue to reveal new insight as we use this data to test different algorithms.
New insight leads to new products
Even though this R&D project is far from completed, it has already led to new product ideas. One of them is the Ad Layout optimization module, which is being developed right now. With this module, your ads will always be optimized to get the most traction among the targeted audience. Maybe the image will change from outdoor to indoor or the CTA button will disappear or change color. No matter what, it will all happen automatically and throughout the life span of the campaign. We are convinced that we have just scratched the surface of the learning and development potential that lay in this research and are excited about the further work.