Case Study: Laanebanken
44Pixel is a Danish-based agency run by passionate people specialized in CRO and data analysis. They want to make the internet a better place by creating better experiences for their users. That’s what CRO is about. The way we see it, this is done by understanding user behavior and having a broad knowledge of how the human mind works. When this knowledge of human behavior is combined with CRO testing, businesses are able to create more effective experiments and provide their users with better experiences, resulting in higher profits for the customer’s website.
Spar Nord is a major Nordic bank based in Denmark that can trace its history back nearly 200 years. Spar Nord engaged 44Pixel to help them improve conversion rates of loan applications on their Laanebanken site, one of Spar Nord’s fastest-growing business segments.
Laanebanken’s visitors are typically shopping for loans online and are driven to their site after they have been pre-qualified through an affiliate they may have visited. Hence, converting on-site visits into filled-out applications in the moment is key. Working with 44Pixel, Laanebanken hypothesized that they needed to test the loan application flow and identify where they could remove or address obstacles in the process to make it easier for the applicant to identify and complete the required steps.
44Pixel used a combination of cognitive research, user recordings, and heatmaps to develop several hypotheses for what changes would increase conversions. With Ascend, they were able to test multiple experiments for each hypothesis simultaneously, thereby increasing their testing velocity. Through these testing efforts, 44Pixel was looking for ways to remove distractions and find designs that would improve the experience on Laanebanken’s site and therefore increase the number of completed loan applications.
While 44Pixel has deep experience in testing and optimization, they wanted to find a way to more efficiently test all their different ideas and Ascend gave them this ability. One challenge for optimization in countries the size of Denmark is the relatively small size of the online audiences, which makes A/B testing a very slow, one-idea-at-a-time process.
Ascend is able to test more ideas because it uses a powerful form of AI called evolutionary algorithms, which are modeled on the workings of Darwinian natural selection. The AI tests sets of designs, identifies the top-performing elements among the set, and automatically generates successive pools (or “generations”) of designs that incorporate those top-performing elements (or “genes”) from previous generations. This process continues over multiple generations, constantly learning and evolving to achieve higher and higher conversion rates. One benefit of this evolutionary approach is its ability to search through very large numbers of possible designs to find the best design, while only testing a fraction of the actual possible designs to get there.
44Pixel used Ascend’s to search through over 750 different possible design combinations for the Laanebanken funnel. In six weeks, Ascend’s AI evolved a design that was converting 5.8% better than the control design.
By The Numbers
In just six weeks, 44Pixel helped Spar Nord achieve dramatic results with Ascend:
week test duration
possible designs evaluated
"Having worked with both google optimize, vwo and optimizely, Ascend is something different, its more than just a testing tool. It is like having an extra employee doing the follow up on the initial tests. I’m a fan."
– Simon Bak
CEO & Co-Founder, 44Pixel