7 business idea generation techniques that actually work
The techniques worth using all do the same thing: they force you to look where you would not naturally look. Problem interviews find funded pain. Mind mapping gets you past the first three obvious ideas. SCAMPER and constraint flipping produce variations on things that already work. Jobs-to-be-done shows you the real competition. Trend riding finds what is newly possible. And founder-market fit tells you which of them is yours to win.
1. Problem interviews
Best for: finding real, funded problems instead of guessing.
How to run it
- 1Pick a group you already have access to, for example the industry you last worked in.
- 2Ask ten of them about their last bad week, not about your idea. Never pitch.
- 3Listen for two signals: the problem recurs, and they already spend money or hours on a workaround.
- 4Write down the workaround. The workaround is your competitor, and its price is your ceiling.
Example
Talking to five dental practices and hearing all five describe the same no-show problem, and all five paying a receptionist to chase patients by phone, is a stronger signal than any brainstorm.
2. Mind mapping
Best for: escaping the first three obvious ideas.
How to run it
- 1Put a domain, not an idea, in the middle: 'independent restaurants' rather than 'restaurant app'.
- 2Branch outwards into the actors (owners, staff, suppliers, customers), then their jobs, then the pain in each job.
- 3Keep branching until you reach leaves that are specific enough to be uncomfortable, for example 'chasing suppliers for missing invoices'.
- 4Harvest ideas from the leaves, never from the trunk. The trunk is where the obvious ideas live.
Example
A map of 'local gyms' branches to owners, then to admin, then to 'reconciling class bookings with billing', which is a far better starting point than 'a gym app'.
3. SCAMPER
Best for: generating variations on something that already works.
How to run it
- 1Take an existing business or product and run it through seven prompts: Substitute, Combine, Adapt, Modify, Put to another use, Eliminate, Reverse.
- 2Eliminate is the most underrated. Ask what happens if you strip out the most expensive or most hated part.
- 3Reverse is the second most useful. Ask what happens if the customer becomes the supplier, or the price model is inverted.
Example
Applying Eliminate to a full-service agency gives you a productized fixed-price package with no account managers, which is a real and repeatedly successful business shape.
4. Jobs-to-be-done
Best for: seeing the real competition and the real reason people buy.
How to run it
- 1Frame the customer's goal as a job: 'when I ___, I want to ___, so I can ___'.
- 2List everything they currently hire to do that job, including spreadsheets, an assistant, and doing nothing.
- 3Your opportunity is where the current hires do the job badly, expensively, or slowly.
Example
'When a client signs, I want to get them onboarded fast, so I can start billing.' Today they hire a checklist and three emails. That gap is a product.
5. Constraint flipping
Best for: finding non-obvious models in a crowded market.
How to run it
- 1Write down the assumptions everyone in the category treats as fixed: how it is priced, who delivers it, how long it takes.
- 2Flip each one and ask what business would exist if the opposite were true.
- 3Most flips are nonsense. One or two are a real wedge.
Example
Bookkeeping is assumed to be billed hourly. Flip it to a fixed monthly price and you get a productized bookkeeping service, which is a genuinely different and often better business.
6. Trend riding (the why-now)
Best for: finding ideas that are newly possible rather than newly imagined.
How to run it
- 1Ask what changed in the last 12 to 24 months: a cost collapse, a new capability, a regulation, a behaviour shift, a new platform.
- 2Then ask which existing, boring business that change makes dramatically cheaper or faster to run.
- 3Do not chase the trend itself. Chase the boring business the trend unlocks.
Example
The interesting play is rarely 'an AI startup'. It is the established service business, like claims appeals or lead-pack preparation, where AI removes most of the cost of delivery.
7. Start from your unfair advantage
Best for: picking between several good ideas.
How to run it
- 1List what you have that most people do not: a skill, a network, credibility in a niche, an insight from a past job.
- 2Score each candidate idea on how much it uses that advantage.
- 3An average idea you have unfair access to usually beats a brilliant idea you do not.
Example
If you spent six years in insurance claims, an unglamorous claims-adjacent service is a far better bet for you than a consumer app you have no way to reach.
One rule that matters more than any technique
Separate generating from judging. In the generating phase, produce quantity and write everything down, including the options that feel silly. The moment you start evaluating, you stop generating, and the unusual ideas (which are often the good ones) never make it onto the page. Judge afterwards, deliberately, against structure: who pays, what it costs to start, how long until revenue, and whether that revenue recurs.
Would rather skip to a vetted idea?
We already ran this process on hundreds of ideas. Each survivor comes with a full playbook.