When most people hear the word “data,” they picture spreadsheets. Percentages, charts, a dashboard with a satisfying number of green arrows.
That kind of data is useful. But it only tells you what is happening. It almost never tells you why. And in design, the why is where all the good stuff lives.
Say your analytics show that 60% of people abandon a sign-up form on step three. That's a real, valuable finding. It tells you exactly where the problem is.
What it can't tell you is whether people are confused, annoyed, distracted, worried about privacy, or just realised they needed a document they don't have handy. Each of those needs a completely different fix. Guess wrong, and you can spend months carefully solving a problem nobody had.
Sit down with five people and watch them fill in that form, and you'll usually know the answer inside an afternoon.
Surveys feel rigorous. Hundreds of responses, tidy percentages, a chart for the slide deck. But they have a quiet flaw: they can only answer the questions you already thought to ask.
If you didn't know the real problem existed, it won't be on the survey. People will dutifully pick from your options, and you'll get a confident-looking answer to the wrong question.
A conversation is different. You can follow the interesting tangent. You can ask “tell me more about that.” You can notice the pause before someone answers, or the thing they mention in passing that turns out to change everything. The best insights in my work have almost always come from something I didn't know to ask about.
Ask someone “would you use this feature?” and most will politely say yes. Ask “how often do you exercise?” and you'll hear a number that's a little more aspirational than accurate.
That's not dishonesty. It's just being human. We're generous with our intentions and forgetful about our habits. That's why I put far more weight on watching someone actually try something than on asking whether they would.
One of the most freeing things in design research is that you don't need a huge sample to find the big problems. Usability researchers have long pointed out that a handful of test sessions will surface most of the major issues, because people tend to trip over the same things.
So instead of one giant study at the end, I'd much rather run small rounds often. Test with a few people, fix what they found, test again with a few more. Each round is cheap, quick, and teaches you something you can act on by Monday.
It's less formal than it sounds. A few of my favourite approaches:
None of these need a lab or a big budget. They need curiosity, a bit of time, and the discipline to listen more than you talk.
This isn't an argument against quantitative data. Numbers are great for spotting patterns, measuring change, and knowing whether a fix worked at scale. The strongest projects use both: numbers to find where to look, and people to explain what's really going on.
But if I had to choose only one at the start of a project, I'd choose the conversations every time. You can always count things later. It's much harder to understand them after the fact.
Data isn't just numbers. It's people, and the reasons behind what they do. Fun, by design.
