According to Attention Insight’s CEO and co-founder, Kamile Jokubaite, marketers are increasingly seeking ways to make their campaigns more efficient. As it is stated in a press release, recent developments in AI and deep learning have opened doors to advancements such as predictive eye tracking that provide an accurate and affordable alternative to tests with real people. However, this was initially limited to desktop and mobile website testing.
At the same time, brands still invest significantly in outdoor advertising, including posters. Also, many advertisers still widely consider ad posters to be the foundation stones for other campaign designs, including images for social media and digital advertising.
“For these reasons, our R&D team trained a predictive eye-tracking algorithm that is based on Deep Convolutional Neural Network technology, specifically on print poster images. For training, we used a wide variety of different posters – more than 24,000 image data sets used, including attention data on poster ads from campaigns by both small and big brands, movie posters, social awareness posters, event posters, etc. We also included specially designed posters, made by our designers in order to train the algorithm to take into account small changes in poster design. This allows our users to make incremental changes in the design of their posters and to see how these changes affect the attention distribution on their poster design,” says Jokubaite.
For algorithm training, Attention Insight used only proprietary eye-tracking data. The company reached 90% accuracy in predicting visual attention on poster designs.
Attention Insight can now provide customers a heatmap that shows the attention distribution on a poster’s design, a focus map that shows which design elements are noticed and which are missed in the first 3–5 seconds, and an exact percentage of attention for the selected design element or elements in a poster.
“We have seen the biggest interest in the new product from the FMCG segment. However, our company aims to democratise eye-tracking in all steps of the brand journey,” Jokubaite says.
So far, Attention Insight has performed more than 60,000 tests on websites and other designs. In 2021, the company doubled its user base and built Photoshop and Figma plugins, plus a Google Chrome extension. In 2022, Attention Insight plans to launch shelf and package design testing, as well as new features on its current product portfolio.
Attention Insight is headquartered in Vilnius (Lithuania), with a branch in Warsaw (Poland) and an office in Hamburg (Germany).