A niche is a person with a particular job to do
“AI,” “wellness,” and “making money online” describe broad topics. A useful starting niche is more specific: independent consultants who need to turn a sales call into a clear proposal, or local workshop hosts who need a page that answers booking questions. You should be able to name who you would speak to this week.
Start with access and experience. Write down tasks people already ask you to help with, problems you have solved repeatedly, and communities where you understand the language. You do not need to be the world's leading expert. You do need to know where your competence ends and what you can responsibly deliver.
Compare three candidates
Use the niche scorecard to give each candidate a score from 1 to 5 on five questions:
- Problem: Can people describe a recent instance, its consequence, and their current workaround?
- Reach: Can you reach five relevant people without buying a large audience?
- Capability: Can you produce one useful result, with a clear boundary on scope?
- Buying evidence: Is there evidence that this audience spends money or meaningful time solving this problem?
- Repeatability: Does the problem recur, and can you serve another buyer without starting from zero?
Twenty-five is a comparison score, not a prediction of demand. Mark guesses as guesses. A lower-scoring niche with real access to buyers can be a better first experiment than an exciting market you cannot reach.
Turn the best candidate into a test
For an illustrative example, “marketing for businesses” becomes “a one-page workshop launch kit for independent instructors.” The buyer is an instructor with a scheduled workshop. The deliverable is the page copy, a registration checklist, and three launch messages. The initial boundary is one workshop and one revision.
Write your own version: “I help [buyer] complete [job] through [deliverable], with [scope] included.” Then list what is excluded. This stops a small offer becoming an unlimited promise.
Ask about what happened, not what might happen
Talk to five people as an initial research exercise; five is a starting workload, not statistical proof. Ask: When did this last happen? What did you try? What did it cost in time or money? Who decides to buy help? What would have to be true to try a small pilot? Record the actual phrases people use, with permission and without storing unnecessary personal details.
Do not treat “sounds great” as a sale. Stronger signals include sharing relevant material, asking about scope, introducing the decision maker, or agreeing to a paid pilot with clear terms. If people cannot recall the problem, narrow it or choose another candidate.
Set a stop rule
Choose a research window and a maximum amount of time and money before you begin. At the end, decide whether to continue, change the audience or offer, or pause. Keep what you learned even if you stop. Once the offer is clear enough to explain, move to choosing a domain and building a small site.
Your next move
Score three niches, identify your weakest assumption, and arrange one conversation that could challenge it.