Where ideas come from, how to research a problem and how to learn whether the market genuinely wants your project.
Good ideas are often discovered, not invented
A useful startup idea often starts with noticing something people already struggle with: a repeated frustration, expensive manual process, awkward workaround or change in technology or regulation. The problem is usually more valuable than the initial solution because the solution can change while the problem remains.
Look for problems with energy around them
A promising problem is painful or frequent, already causes people to spend time or money on workarounds, affects an identifiable audience, can be reached through real channels and has a plausible budget. Timing matters too: a new technology or behaviour can make an old problem suddenly solvable.
Build a problem-finding habit
Keep a problem journal. Observe work processes. Look for spreadsheets, copy-paste routines and services held together manually. Read complaints and reviews rather than only trend reports. Examine niches ignored by broad platforms. Start with a service and notice what could become software. Combine two fields you know unusually well.
Use brainstorming to expand options, not to avoid evidence
Exercises such as “ten problems a day”, “why is this still done this way?”, “what would be ten times better?”, jobs-to-be-done and reversing assumptions can generate directions. Then narrow them using evidence: problem intensity, reachability, existing spend, competition, market change and your ability to learn quickly.
Validation is a ladder of stronger evidence
A compliment is weak. A signup is stronger. A booked call, completed test, deposit, pilot or payment is stronger still. Interviews should focus on past behaviour: when did the problem last happen, what did the person do, what did it cost, and what happened next? Avoid explaining your solution too early because people become polite instead of informative.
Prototype, MVP and final project are different things
A prototype tests understanding or interaction. An MVP is the smallest functioning version that can test a business or behavioural hypothesis. The final project may eventually contain much more. Choose the MVP based on the biggest risk, not the longest feature list.
Find early adopters where the pain is strongest
The first users are rarely a random sample of the whole future market. Look for people who already spend money or effort trying to solve the problem. Contact them directly, visit the communities and platforms they use, and ask for a concrete action rather than “Would you like to test something someday?”
Product-market fit is behaviour, not applause
There is no universal PMF percentage. Useful signs include retention stabilising, customers paying with less persuasion, organic referrals, complaints when the service is unavailable, requests for more and one segment clearly outperforming others.
Traffic, registrations, a good Product Hunt day, investment, one large client or people saying “this is great” are not product-market fit by themselves.
A practical path from idea to PMF
Collect many problems, select a few, speak to real people, identify the riskiest assumption, build the smallest credible test, recruit 5–20 early adopters, ask for action or payment, build a narrow MVP and measure activation and retention. Then focus on the segment where the strongest signals appear.
A real experiment also needs a result that would make you stop or change direction. Otherwise you are not validating the idea; you are searching for confirmation.