How the research works

Every report gets built in two phases. Discovery comes first: it reads real public discussion to find what people are struggling with, who they are, and what you could build for them. Deep Research comes second, where we pressure-test the most promising idea on demand, pricing, market size, and timing. (See How one idea gets pressure-tested.) This page is about Discovery.

The five Discovery steps

It runs in order, and each step feeds the next:

1. We work out what your niche actually is, and the language that identifies it. (More on this below.)
2. We scan public discussion, searching the places your audience actually hangs out.
3. We pull the recurring problems out of those conversations and back them with evidence. (See How pain points are gathered.)
4. We map who has those problems: the buyer segments, communities, and voices behind them.
5. We sketch a varied, vetted shortlist of things you could build. (See How ideas are generated.)

Understanding your niche

The first step does more than repeat your input back to you. It works out what kind of input you gave in the first place. Did you hand us a market ("home espresso gear") or an audience ("freelance designers")? Those need to be handled differently. If you named an audience, we still map the broader market they sit in and its main buyer segments, so the research doesn't tunnel down onto one slice and miss the demand sitting right next to it.

It also picks out the signature terms, products, and communities that mark your niche out, the names insiders use that the look-alike topics don't. Those act as anchors. They help the next step keep your niche apart from the neighbouring ones that happen to share generic words, so the search stays on target instead of drifting. If your niche is too broad to produce clear anchors, we just skip that check rather than force it.

One more thing: if you came in through the "idea," "audience," or "discovery" entry points, we treat that as a hint, not an instruction. The system still reads your input on its own merits and decides for itself.

What it can't do

We work from public discussion, so if there's very little of it for your niche, a run can stop early rather than make up signal that isn't there. And the niche framing is an interpretation. If we read your input differently than you meant it (a market when you meant an audience, or the other way round), the quickest fix is to rephrase it so it's obvious which one you mean.