Analytics will tell you what your users did. They will never tell you why.
That gap is where most products go wrong, and it is the reason I keep coming back to one uncomfortable discipline: before you build anything to solve a problem, go and watch a real person live inside that problem.
Not a survey. Not a call. Watching. In their setting, on their device, with their distractions, on their bad day.
I picked this up in an unlikely place. Long before design, I trained in microbiology, and the first rule you learn in a lab is that you do not decide what the sample is doing. You observe what it is doing. The system holds the truth and your only job is to watch closely enough to catch it. I have never found a better description of what real user research is.
Here is the tension. Watching feels like the opposite of progress. Building is visible and satisfying and easy to show off. Observing is slow, quiet and a little humbling, because you usually discover that the thing you were proud of does not survive contact with a real person in a real kitchen. So we skip it. We tell ourselves we already understand the user, because we are users too.
We are not. We know where every button is, because we put it there.
When I build services used by millions of people for the UK government, the first design principle is not speed or polish. It is start with user needs. Do the research. Talk to users. Do not assume. The order is deliberate, and it is the order almost everyone gets wrong.
What this should change is simple. Before your next build, spend one hour watching five real people use what you have, or use whatever they use today instead of you. Not a thousand. Five. Proximity beats volume every time.
The problem you actually have is almost never the problem you assumed you had. You just have to be in the room to find out.
I wrote the full version of this on the blog this week, including what sitting with someone on a call based in a country I have never been to taught me that a year of testing did not. Link below.
https://blog.samuelakintunde.com/the-first-thing-to-do-is-not-build-it-is-watch/