In its origins, UX design was traditionally based on designer intuition and empathy, with research and design techniques guiding the way. However, intuition can be biased, and user personas can get outdated fast in a year-long development cycle. As time progressed, astute UX designers began paying attention to big data technology, treating it more than just a buzzword. Machine learning and data analysis techniques can contribute to UX design by balancing on designer biases, and even predicting potential changes within a user group.

But how?

More Efficient Pattern-Findings

Instead of going through results from surveys and workshops, pattern-finding can be implemented with full automation via algorithms, and the findings can be much more comprehensive and insightful. This is achieved by collecting a rich amount of data from various sources which are more accurate than survey and workshop performance, and applying algorithms which are more rigorous than human reasoning.

Revolutionary Persona Building

Persona building is essentially a bundle of pattern-findings. By collecting data and creating models which describe patterns, a user persona can be developed. This persona is a collection of models that classify and predict user preference, behavior, and performance, depending on the application. Once the model is adjusted and optimised, a deep level of UX customisation is made possible.

Fig.1. Traditional UX design flow vs. ML-UX design flow

Fig.2. Use data to represent users; use models to predict users.

Adapting to Changes

Since persona building can be simplified as an input-output process, with a model which is comprehensive enough, it is possible for instant coping to user data changes, which may result in a considerable shift of persona.

Machine learning techniques are powerful tools with the ability to leverage user data. Businesses with the luxury of owning their user data should be adaptive to applying this emerging technology to whatever services they are to provide.