Machine Learning Campaigns

MY ROLE

UX Designer

OVERVIEW

Many businesses see 1:1 personalization, where each customer is delivered a uniquely relevant experience, as a key to their brand’s success. Brands typically optimize their experiences for customers in a one size fits all manner (e.g. addressing a site-wide drop off in completed checkouts) while some invest in peronalizing for different segments (e.g. tailoring the home page to different personas). But each time a new segment is created, new content must be created, tested, and analyzed, which is resource intensive and makes it impossible to personalize in a 1:1 manner. Using machine-learning allows marketers to personalize at a scale that can’t be achieved through manual A/B testing and segmentation. We’ve created two types of machine-learning powered campaigns: one that finds the best experience for the majority of an audience and one that finds the best experience for each individual.

CHALLENGE

Organizations are deeply rooted in testing habits. A key part of this project was tailoring the explanation of how the machine-learning algorithms work (they’re inherently black boxes) in a way that matched people’s expectations and supported current workflows: we needed to help people change their way of thinking while bridging the gap to their familiar A/B testing world.

SOLUTION & IMPACT

A large portion of the client base has adopted the 1:1 personalization platform. We’ve seen results that effectively increase brand engagement by up to 40%. Customers have found the insights generated by these algorithms as validating and fuel for further iterations of personalized experiences.