Thank you!
Your submission has been received!
Your submission has been received!
Oops! Something went wrong while submitting the form.
I look at the job behind the request: what someone was trying to achieve, what they used before, and what would make them switch back. It reframes a feature list into a set of outcomes worth competing on.
Your backlog is full of feature requests and you suspect they're symptoms rather than causes. Customers describe solutions instead of problems. Two teams disagree about who the product is for and both have anecdotes. You've built what people asked for and usage didn't move. The clearest signal: you can list what customers want, but not what they were trying to achieve when they asked.
I interview customers about a specific episode rather than their preferences in general. What were you doing when you first looked for something like this. What did you use before. What nearly stopped you switching.
The framing matters — people are unreliable about what they want and very reliable about what they did. I ask about the last time, not the typical time.
I also mine what you already have: support tickets, sales call notes, churn responses, review sites. AI does the synthesis across hundreds of sources, which means the findings come from your whole customer base rather than the six people there was time to call.
The output reframes a feature list into a set of outcomes worth competing on — and usually kills two or three things on the roadmap.
Two to three weeks for a focused study. Longer if the product serves genuinely different segments, because each one needs its own interviews.
How many interviews is enough?
Usually eight to twelve per segment before patterns repeat. If the tenth interview is still surprising, the segment is wider than assumed — which is itself a finding.
Can you use our existing research?
Yes, and I'd want to. Most companies are sitting on more evidence than they realise. What's usually missing is synthesis, not data.
Isn't this just user research with a different name?
The difference is the unit of analysis. Personas describe who someone is; jobs describe what they were trying to get done. The second is more stable and more useful for deciding what to build.