At cape we constantly innovate to make digital marketing faster, better and easier. In this series we will take you through our latest innovations: what we are working on, why we do it and how we do that. This is Cape Labs. Today we will discuss our research in the field of creative analysis. We all strive at achieving the best possible results with our ads, where in reality we hardly touch different variations or truly optimise visuals, although everyone knows that this can really impact the performance of your ads. At Cape we already solved that problem by allowing our users to create different A/B tests. However, it will be even more powerful when you could give hints to designers and creators during the design process and that's why we started experiments with creative analysis. A visual can gain much better results based upon different creative aspects. So, this can for instance be the brightness, the colours, contrast, rule of thirds, copy and many other aspects. The effects of those different visual aspects on ad performance have been researched quite a lot in the past. So, what if we could use that data and analyse ad-performance before even publishing them? We started experimenting with automating this process. So, we use different analysis tools for fetching all the data related to the visuals. This is done, for instance, via Google Vision API and various Javascript libraries. Based on available research, we score images on various scales. So, this is based upon colour, composition, sharpness, brightness, logo positioning scale, facial expression, text, etc. etc. This is only our first step of visual analysis. Our next goal is to combine this with actual ad-performance and create a self-learning system to optimise your ads.