Is My Startup Idea Good, or Am I Just Biased? A Self-Scoring Framework

You already suspect the answer, which is exactly the problem. You've lived with the idea for weeks. It feels obvious, inevitable, good. But "feels good" is the least reliable signal you have — because you are the single most biased person who will ever evaluate it.

So the honest question isn't is my idea good? It's would anyone without my emotional investment score it the way I do? This guide turns that vague feeling into something you can defend: a score, against explicit criteria, that a stranger could check. Not a number to hide behind — a way to make your own blind spots visible before the market makes them expensive.

Why you can't trust your own read

Four biases quietly manufacture a "yes." Name them so you can catch them:

Confirmation bias. You go looking for reasons to keep building, not reasons to stop. Every polite nod becomes "validation."
Sunk cost. The more time, money, or identity you've put in, the harder it is to score the idea low — quitting feels like admitting the whole thing was a mistake.
Solution love. You fell for your answer, not the problem. So you evaluate how clever the solution is instead of whether the problem is real and painful.
Friendly sample. Your evidence comes from people who like you — friends, your network, your Twitter followers — not from the cold market that has to pay you.

If your read on the idea is mostly a feeling, assume at least one of these is doing the talking.

"Good" is not a feeling — it's a score against criteria

A good idea is one that scores well on a small number of dimensions that actually predict whether a business can exist. Vibes don't predict that. These do — and each is a question with a defensible answer, not a gut rating:

Problem severity. Is this a hair-on-fire problem or a mild annoyance? Painkiller or vitamin? People rearrange budgets for painkillers and ignore vitamins.
Frequency. How often does the pain occur — daily, monthly, once a year? Rare problems are real but build weak habits and weaker retention.
Willingness to pay. Is there an existing budget for this, or would you be creating a new line item from scratch? "They'd love it" is not "they'd pay for it."
Market size and timing. Is the segment big enough to matter, and is there a credible why now — a shift that makes this possible or urgent today that wasn't true two years ago?
Your unfair advantage. Why you, and why now you specifically? Distribution you already own, domain depth, proprietary data — or nothing a competitor can't copy in a weekend?
Reachability. Can you actually get in front of these buyers affordably and repeatably, or are they scattered and expensive to reach?

Notice what's not on the list: how elegant your product is, how much you'd personally use it, how excited your cofounder is. Those feel like signal. They aren't.

A scorecard you'd defend to a stranger

Score each dimension 1–5 with a one-line justification you'd be willing to read aloud to someone with no stake in the outcome. Then resist the one move that ruins the whole exercise: do not average.

A weak idea hides behind a strong average. Here's the same idea, scored honestly:

Dimension	Score	One-line justification
Problem severity	4	Ops leads call it their "most-hated Monday task"
Frequency	5	Happens every single week
Willingness to pay	2	Currently solved with a free spreadsheet; no budget line
Market size & timing	4	~40k reachable teams; remote work made it worse
Unfair advantage	3	Founder ran this exact process for 6 years
Reachability	4	These leads live in 3 specific communities

Average that and you get a respectable 3.7. But the 2 on willingness to pay is the whole story — a frequent, painful problem that people already solve for free is a famous way to build something everyone praises and nobody buys. The rule: a single 1–2 on severity, willingness to pay, or reachability is usually fatal, no matter how strong the rest looks. Score the floor, not the mean.

Four tests that strip out the bias

The stranger test. For each score, ask: would someone with zero emotional investment land on the same number from the same evidence? If your justification is "I just have a feeling," that's a 1 on evidence, whatever you wrote.
Pre-commit the bar. Decide before you score what "good enough to build" looks like (e.g., no fatal dimension, and a real budget exists). Set the threshold while you're still rational, so you can't quietly move it after a low score.
Get three outside scores. Hand the same six dimensions to three people who will tell you you're wrong — ideally one potential buyer. Where your score and theirs diverge by two or more points, that gap is your blind spot.
Two columns, kept honest. Write "Why I'm excited" and "What it actually scores" side by side. When they disagree, trust the score. Excitement is fuel; it is not evidence.

Common mistakes

Averaging away a fatal flaw. A 3.7 with a 2 on willingness-to-pay is not a 3.7 business — it's a no with good lighting.
Scoring the solution, not the problem. You can't out-design a problem nobody has.
Sampling only people who like you. Warm intros validate your charisma, not your market.
Moving the bar after a low score. If the threshold is negotiable, you never had one.
Mistaking one ecstatic user for a market. A single whale, or your most enthusiastic friend, is a sample size of one.

How God of Startups helps

The reason "is my idea good?" stays a feeling is that the evidence lives in your head, unstructured. God of Startups turns it into a legible, scored read instead of a gut call. From a short brief, its agents work the idea across exactly the dimensions that matter — a sharpened pain point, how often it bites (frequency), whether there's a real budget for it, the market scale and why now, and the competitive landscape — and pull every one of your underlying bets into an assumptions registry and a risk map.

The result is the opposite of confirmation theater: you can see where the idea is genuinely strong and where it's a bet you haven't tested — a readable assessment you can act on, share with a cofounder, or hand to an advisor, rather than a number you talked yourself into. It won't tell you your idea is good to make you feel better; it makes the weak spots impossible to look away from. The verdict is still yours — but now it's grounded.

FAQ

I'm emotionally attached to this idea. How do I score it honestly? Outsource the scoring you can't trust. Pre-commit the criteria, get three outside scores, and weight a potential buyer's view above your own. Your job is to fight for the idea; their job is to score it. Don't do both.

What score is "good enough" to start building? There's no magic average. The practical bar is: no fatal dimension (nothing scoring 1–2 on severity, willingness to pay, or reachability), at least one genuine strength, and a credible why now. A "good enough" idea is one with no dealbreaker — not one with a high mean.

Isn't all of this just subjective? The scores are judgments, but the evidence behind each score is not. "4 because ops leads called it their most-hated task" is checkable; "4 because I feel good about it" is not. The exercise forces you to attach evidence — and the absence of evidence is itself the most useful result.

My idea scores low but I still believe in it. Now what? Find the specific low dimension and test that, cheaply, this week. If willingness to pay is the 2, run a pre-sell, not a survey. Belief is permission to test the weak spot — not permission to skip it.