Here is the uncomfortable part of evaluating a business idea: enthusiasm feels exactly like evidence. The idea you keep turning over in the shower produces the same warm certainty as the idea three strangers have already asked you to build. From the inside, they're indistinguishable.
The takeaway up front: liking an idea and the idea being good are two separate measurements, and only one of them lives inside your head. Liking is about you — your taste, your skills, the story you'd get to tell. Goodness is about someone else — a problem they already have, already spend something on, and would notice if you solved. You can't reason your way from the first to the second, but you can test for the gap in an afternoon.
None of this is about killing the ideas you love. Genuine interest is a real advantage; it's why you'll still be at this in month five. The point is knowing which kind of confidence you're holding before you spend a season on it.
Liking is a measurement of you, not of the market
When you feel drawn to an idea, you're usually detecting one of three things — and all three are about your own position, not the world's.
Identity fit. The idea would make you a certain kind of person: the one with the shop, the newsletter, the neat little app. That's a real motivation and worth naming honestly, but it says nothing about whether anyone else wants the thing.
Fluency. You can picture it easily — the logo, the first page, the packaging — and your mind reads that ease as plausibility. Ideas that are simple to imagine feel more likely to work. Often the reverse is true: the easiest ideas to picture are the ones everybody pictures, which is why so many of them already exist in a crowded form.
Imagined effort already spent. You've been turning it over for weeks, and dropping it now feels like wasting that. It isn't — thinking is how you found the flaw, and finding it early is the whole return.
None of the three touches demand. That's the gap: an idea can fit your identity, be easy to picture, and be long-considered, and still have nobody on the other end of it.
Six tests that separate the two
Run these on one idea at a time, on paper, in about forty minutes. They're deliberately unglamorous — each one asks for something outside your own head.
1. The source test: where did this actually come from?
Write down, in one sentence, the moment the idea arrived. Then sort it into one of two piles.
Inside sources: "I thought it would be cool," "there's nothing like it," "I'd use this myself," "I saw a video about the market." Outside sources: "three different people asked me to do this," "I watched someone pay badly for a bad version," "I did this manually for a client twice and they asked me to keep doing it."
Inside sources aren't disqualifying — plenty of good businesses started as a personal itch. But an inside-source idea has no evidence attached yet, and should be treated that way rather than letting your enthusiasm get counted twice. An outside-source idea arrives with a small piece of proof already stapled to it.
2. The restatement test: say it as their problem, in their words
Rewrite the idea as a sentence spoken by the person who'd pay, using vocabulary they'd actually use. Not "an AI-powered scheduling platform for wellness professionals" but "I lose an hour every Sunday texting people to move their appointments."
Two things go wrong here, and both are informative. If you can't write the sentence at all, you have a product concept without a person — the most common failure in a liked idea. If the sentence you write is one you've never heard a real person say, you've invented the customer along with the product. A good idea usually lets you quote someone almost verbatim.
3. The substitute test: what do they do today instead?
Nobody is waiting in a vacuum. Whatever your idea solves, people are handling it right now — with a spreadsheet, a competitor, a friend who does it as a favour, a WhatsApp group, or by tolerating the annoyance.
Find that substitute and describe it plainly. It tells you the real bar you have to clear — you're not competing with nothing, you're competing with a spreadsheet that's free and already set up. It also confirms the problem exists: if you genuinely can't find any substitute behaviour, the usual explanation isn't untouched territory, it's that the problem isn't annoying enough for anyone to have bothered solving it.
4. The costly-signal test: separate approval from interest
Ask ten people what they think of your idea and most will be encouraging, because encouragement is free and disagreeing with a friend at a barbecue is awkward. Approval is close to worthless as evidence, and it's exactly what a liked idea attracts most.
What counts is anything that costs the other person something:
- Money — a deposit, a pre-order, a first paid job even at an embarrassing price.
- Time — a booked call, a ten-minute form, a session where they show you their current workaround.
- Reputation — an intro to a colleague, or agreeing to be named as a first user.
"That's a great idea, you should build it" is a data point about your friendship. "Can you do mine next Tuesday — what do you charge?" is a data point about the idea.
5. The kill-criteria test: what would make you drop this?
Before you go looking for evidence, write down what would make you stop — in advance and in specifics. "If I talk to eight people in this trade and none of them describes this as a recurring annoyance, I stop." "If nobody will pay for the manual version, I stop."
If you can't finish that sentence, you're not evaluating any more; you're gathering support. That's the clearest tell that you're holding a liked idea rather than a good one — a liked idea has no conditions under which it loses, because your commitment came first and the reasons are being fetched afterwards.
Written kill criteria also protect the idea from the opposite failure: dropping it the first time someone is rude about it at dinner.
6. The boring-middle test: do you want the work or the outcome?
Every idea has a glamorous tenth and a repetitive rest. Write down what the rest actually is — for a small product: packing, listings, emails about shipping. For a service: quoting, chasing invoices, the fourth revision. For content: publishing on a Tuesday when nobody read Monday's.
Then ask honestly whether you'd tolerate that specific work for a year. This test doesn't measure the idea's quality; it measures whether you are the right owner of it. A strong idea attached to a middle you'll quietly avoid is worse, for you, than a modest idea whose daily work you don't mind.
A worked example: two ideas, side by side
Say you're choosing between two candidates.
Idea A — a subscription box of beautiful stationery. You love stationery and can already see the packaging. Source: inside. Restatement: the best you manage is "I wish more people appreciated good notebooks" — your feeling, not their problem. Substitute: people who want nice stationery already buy it easily from many shops, so the substitute is excellent and cheap. Costly signals: friends saying "I'd totally subscribe," which costs nothing. Boring middle: packing boxes and answering shipping emails every week.
Idea B — monthly bookkeeping tidy-ups for tattoo artists. It came from doing this twice for a friend, who then referred a colleague. Source: outside. Restatement is close to verbatim: "I do my books at midnight the day before the deadline and I'm never sure I've claimed the right things." Substitute: a shoebox of receipts and a panicky weekend, or a general accountant who doesn't know the trade — real, visible, unsatisfying. Costly signals: one person already paid you, another asked your rate. Boring middle: reconciling transactions and chasing people for receipts.
Idea A is the one you like. Idea B is the one with evidence. The conclusion isn't automatically "do B" — if B's boring middle genuinely repels you, it isn't your idea to run. But you now know which is which, and that A is currently a preference wearing the costume of a plan. The useful move for A is a cheap test rather than a promotion: sell one small batch by hand and let it earn its place.
Put it on a scorecard before you reread the list
Evaluation usually goes wrong not for lack of tests but because each idea gets judged against whichever criterion it happens to win on. That's the same trap that makes a shortlist unchooseable, covered in our guide to why you can't think of ideas: with no criteria fixed in advance, the goalposts move each time you look. So fix them first, then score:
| Column | What you're scoring | Scale |
|---|---|---|
| Evidence | Outside source, quotable problem, real substitute, any costly signal | 0–3 |
| Reachability | Can you find twenty of these people this month without a budget? | 0–3 |
| Effort to first test | Could you deliver it manually, once, within two weeks? | 0–3 |
| Middle tolerance | Would you accept the repetitive ninety percent for a year? | 0–3 |
Two rules make it work. Set the columns before you reread your list — otherwise you'll design the scorecard around the answer you already want. And let enthusiasm break ties only: when two ideas score the same, take the one you like. That's the right place for liking, and it isn't a small one.
What to do with the idea you like anyway
Don't bin it. Demote it to an experiment with a cheap, defined shape: one batch, one client, four posts, one weekend. Write the kill criteria before you start, and set a review date so the idea stops living rent-free in your attention.
Plenty of liked ideas earn their evidence later. The mistake is skipping the earning step because the feeling was strong — and the fix is small: make the idea produce one costly signal from someone who isn't you.
FAQ
How do I evaluate a business idea without spending money? Spend attention instead. Write the customer's problem in their words, find the substitute they use today, and try to collect one costly signal — a booked call, a paid manual version, a deposit. Most of the evaluation that matters happens in conversations and one delivered-by-hand attempt, long before there's anything to buy.
How do I know if my business idea is any good? Look for evidence outside your own head: it came from someone else's repeated complaint, you can quote the problem almost verbatim, there's a visible substitute people already tolerate, and at least one person has spent money, time, or reputation on it. Your own interest is a reason to persist, not a reason to believe.
Is it a bad sign if nobody else is doing my idea? More often than founders expect, yes. An empty field usually means the problem is mild, the buyers are hard to reach, or someone tried and quietly stopped. Treat "no competitors" as a question to answer rather than a moat, and find out what happened to whoever tried.
How many ideas should I compare before choosing one? Enough that you're choosing rather than settling — twenty or thirty candidates generated before any judging, then three to five scored properly. Generating and evaluating in one sitting corrupts both jobs: the judging kills the list early, and the enthusiasm survives the scoring.
What if I like an idea that scores badly? Give it a small, cheap test with kill criteria written first, and let the result decide instead of arguing with yourself. Liking an idea is a real advantage over a year of unglamorous work — it just isn't evidence that anyone wants it.
Still building the shortlist?
Every test above works better with more candidates, because comparison is what stops an idea being judged against itself. If your list is short enough that your favourite wins by default, widen it first — get idea starting points from the Mellow Ideas generators for business names, taglines, content topics, or side-hustles. They're our generators, and we'll say plainly what they do: hand you a batch to react to. The evaluating stays yours.
