How to validate a startup idea — the full process

The validation process, step by step

Step 1 — Input parsing

Evaluation begins by parsing your idea description for core signals: who the customer is, what problem you solve, how you solve it, and how you plan to make money. The clearer these signals are, the more accurate your score will be. Vague inputs produce vague verdicts; specific inputs produce sharp ones.

Step 2 — Demand analysis

Is there real market demand or are you solving a problem nobody has? The system checks whether people are actively searching for the solution, whether interest is growing or declining, and whether there is community chatter around the problem space.

Step 3 — Competition and differentiation

Who else is doing this and how are they positioned? Competitive density, differentiation strength, and switching friction all get scored. Competition is not automatically bad — but undifferentiated competition is fatal.

Step 4 — Feasibility and execution

Can you actually build this? Technical complexity, regulatory barriers, platform dependencies, and speed to MVP all feed the execution signal. High feasibility means you can ship something quickly; low feasibility means the idea carries platform or capital risk you may not survive.

Step 5 — Monetisation and distribution

Will people pay, and can you reach them? Pricing benchmarks, competitor revenue models, and distribution channel realism all combine here. An idea that scores well on demand and feasibility but poorly on monetisation or distribution is still a Fix, not a Ship.

Step 6 — Weighted score and verdict

All signals combine into a weighted overall score. The verdict — Skip (0–54), Fix (55–79), or Ship (80–100) — is the output of a structured system designed to be honest when optimism would be easier.

Common validation mistakes

• Asking the wrong people. Friends and family give socially acceptable answers, not market data.
• Counting interest as intent. "I'd use this" is not the same as "here's £20".
• Skipping competition mapping. If you can't name three competitors, you haven't looked.
• Validating the product, not the channel. A working product nobody can find is still a failure.
• Optimising the deck before the offer. Decks don't pay. Offers do.

Last reviewed 26 July 2026