In short: Scoring a startup idea means rating it on a fixed set of independent criteria so that ideas can be compared against each other rather than judged in isolation. The value comes from the comparison, not the number: a score of 71 means nothing on its own, but 71 against a median of 59 across a hundred alternatives is a real signal.
Score to rank, not to decide
A score cannot tell you whether to build something. It can tell you which three of your twenty ideas deserve a week of validation each. That is a genuinely useful narrowing function, and it is the only job a score should be asked to do.
Treating a score as a verdict is how people end up building the thing that scored 84 instead of the thing they actually understand. Founder-market fit is not in any framework and beats every framework.
The criteria worth using
| Criterion | Question it answers | Common mistake |
|---|---|---|
| Demand | Are people already spending money or hours on this? | Counting enthusiasm as demand |
| Moat | What stops the second mover from copying it in a month? | Assuming being first is a moat |
| Monetisation | Who pays, how much, and how often? | Pricing against cost rather than value |
| Feasibility | How small can version one be? | Scoping version three |
| Timing | What changed recently that makes this possible now? | 'AI got better' as a why-now |
| Open field | How well is this already served? | Mistaking an empty market for an opportunity |
Weight the criteria to your situation
The same idea is not equally good for everyone. A solo developer should weight feasibility far above moat, because an unfinished defensible product is worth zero. A venture-backed team should do the reverse, because a feasible undefendable product gets copied and margin-crushed.
This is why fixed frameworks mislead: they encode one situation's priorities as universal truth. If you use a scoring system, adjust the weights before you adjust the ideas.
Four traps
- Scoring your favourite idea last, after calibrating on ideas you do not care about. Score it first, then compare.
- Letting one criterion leak into another. If 'timing' and 'demand' always move together in your scoring, you have one criterion, not two.
- Using a score to justify a decision you already made. If no score would change your mind, skip the exercise and admit the decision.
- Comparing scores across different scoring systems. A 78 from one framework and an 78 from another are unrelated numbers.