The part I like best

The part I enjoy most about consumer research is not gathering the data. It's the moment when the mess starts to make sense.

I taught myself to build a scraper this spring. For about a week it felt like a superpower. Then I looked at what it actually gave me: hundreds of Reddit comments, reviews, trend curves. A pile. Some of it useful, most of it noise.

AI is genuinely good at sorting the pile. It finds patterns faster than I ever could, and it lets me ask questions I wouldn't have had time to ask before.

But a pattern is not an insight. Some patterns sit on thin data. Some only exist because I assumed them into the analysis. Some are real and still don't matter. So now I audit my own research the way I would audit someone else's: where did this claim come from, and would it survive a skeptical colleague?

The strange thing is that the more I use AI, the less I believe it has the answer. It does more of the work every month. Deciding what to trust and what it means still comes down to judgment and taste.

Which is the part I like best.