Am I the only one noticing this?

UX researchers and designers, who were once separate from data analysts, are now working together. What great news! When organisations ask for a UX designer now, one of the key requirements is that they be data-driven.

Studies show that data-driven organisations make better strategic decisions, experience higher operational efficiency, improved customer satisfaction, and stronger profit and revenue levels. Recent research even reveals that data-driven organisations are twenty-three times more likely to acquire customers, six times more likely to retain them, and nineteen times more likely to be profitable.

This highlights an important truth: at the intersection of two or more seemingly unrelated disciplines, far more valuable insights emerge than from within a single field alone.

I’ve always been a data-driven designer, though not officially. No one had ever requested that I take usage data into account when conducting research—until now.

What I see happening is a shift away from the reliance on surveys, as we incorporate analytics into our mix of quantitative methods and desk research.

The triangulation of information in user research could look like this:

Triangulation isn’t a miracle method; it has its critics. Sure, it increases the likelihood of making informed decisions, but there’s a catch. Surveys have long been criticised for being inaccurate or biased—whether because respondents don’t answer truthfully or because the analysis is flawed. We only need to look at election polls: when are they ever spot-on? And what’s true for surveys also applies to usage data. While there aren’t respondents tailoring their answers to what they think we want to hear, the data is still analysed by humans. Even with machine learning, the process is indirectly biased (take a look at the debate on “math washing”).

Data is math, and math is binary—it proves something right or wrong. There’s no disputing a correct mathematical result. And that’s where things can go wrong.

 Machine learning is math, and its outcomes leave no room for debate or doubt. It presents hard evidence.

This places a lot of power in the hands of data analysts.