How to pressure-test a business idea
The judge on our homepage gives you a snap verdict in a few seconds. This is the slow version you can do yourself in an afternoon. It follows the same questions the judge asks, because they are where ideas usually break: is there a real problem, is there a real customer, will anyone pay, does it already exist, and can you actually build it?
None of this needs a business degree. It needs honesty, a search engine and a few awkward conversations.
1. Say it in one sentence
Fill in the blanks:
For [a specific kind of person] who [has this problem], [my idea] does [this], instead of [what they do today].
If you can't fill a blank, you've found your first job. “For everyone” is not a person. “It's like Uber, but for…” describes a business model, not a problem. And if the sentence needs a second sentence to make sense, a stranger reading your landing page will give up sooner than you think. That is what the judge's Easy to understand score is reacting to.
2. Find the problem, not the feature
Most weak ideas start with a cool feature and go looking for a problem. Turn it around. Ask:
- Who has this problem, and how often? Daily pain beats yearly mild annoyance.
- How do they deal with it today? Look for workarounds: a messy spreadsheet, a group chat, a paid assistant, a sticky note on the fridge. Workarounds are evidence that the problem is real.
- What does it cost them now, in money, time or embarrassment?
If the honest answer to “how do they deal with it today?” is “they don't really think about it”, the need is probably low, however clever the solution. In our examples, a restaurant charging $50 for a glass of tap water scored 0/10 on need. It's perfectly clear and fairly original. Nobody needs it.
3. Name one real customer
Describe one actual person, or one actual kind of business, who would use this next week. Then split two roles that are easy to blur: the person who uses it and the person who pays for it. Parents pay for kids' apps. Managers pay for tools their teams use. If the user loves it but the payer doesn't care, you have a problem with a nice interface.
Then go and talk to five of them. Don't pitch, and don't ask “would you use this?”, because people are polite and hypothetical answers are cheap. Ask about the last time the problem actually happened: what did they do, what did it cost, what did they try? If nobody can remember a last time, that tells you something too.
4. Follow the money
Write down who pays, how much, how often, and what they stop spending money on to pay you. Then do the napkin math: a realistic price, times the number of customers you could realistically reach in a year. Not the size of the whole market, the part you could actually get to.
Sometimes the answer is “a few hundred a month”. That's a fine hobby or side project, as long as you know it's one. The judge's favourite roast for this situation is “Great hobby. Tough business.” It's a description, not an insult.
5. Search before you build
Give it twenty honest minutes. Search for the problem, not just your product name. Check app stores and the places your customers already talk to each other. You'll usually find one of three things:
- Nothing. Either you've found a gap, or there's a reason nobody does it. Try hard to find the reason before assuming the gap.
- Lots of competitors. Good news about demand, bad news about easy wins. Now you need a real difference: cheaper, simpler, built for a group others ignore, or noticeably better at one thing.
- Dead competitors. The most useful kind. Find out what killed them.
“It already exists” isn't fatal. “It already exists, and mine is the same thing but later” usually is. That's the Originality score.
6. List the hard parts
Every idea has a part that is easy to describe and hard to do. Write yours down honestly. Common ones:
- Regulation: anything touching health, money, food, children, transport or personal data.
- Two-sided markets: a marketplace needs buyers and sellers at the same time, and neither shows up without the other.
- Trust and safety: if strangers meet, pay or share things through your product, someone will misuse it.
- Atoms: hardware, inventory, shipping and returns cost more and move slower than software.
Hard doesn't mean no. It means you should know the price before you start. That's the Execution difficulty score, the one score where higher is worse.
7. Run the smallest real test
Before building the whole thing, find the cheapest way to see whether anyone actually wants it:
- A one-page site that explains the idea and asks for an email address or a pre-order.
- A manual version: do the service yourself, by hand, for a handful of people, and only automate what keeps repeating.
- An honest post in a community where your customer already spends time, asking about the problem rather than selling the solution.
A test counts when someone has to do something slightly costly: give an email, give time, give money. Compliments don't count.
8. Decide what would change your mind
Before the test, write down the result that would make you stop or change direction. “If fewer than ten people sign up after two weeks of honest effort, I'll rethink it.” Ideas are easy to fall in love with, and it's much harder to set the bar after you've seen the results.
Where the judge fits
Use Is My Idea Stupid? at the start and again after you've sharpened the one-sentence version. It's quick at spotting the obvious hole: no clear need, no obvious customer, no way to make money. It doesn't know your market, your skills or your timing, and it can't replace step 3. Treat the “biggest problem” line as the first question to answer, not the last word. Here's how the verdicts are calculated.