| Idea source or signal | What it tells you | Planning use |
|---|---|---|
| Workarounds and hacks | People with acute pain cobble together shortcuts | Prioritise problems where people already improvise, over problems they only mention |
| Imperfect existing solutions | 10-20 imperfect solutions typically exist long before a strong one | List the current workarounds and name what each one gets wrong |
| Four improvement lenses | Delivery process, location, cost and customer experience | Take a product you admire and test it against each lens |
| Cost-based improvement | Wyze sold cameras for one-tenth the cost of leading competitors | Price cuts only work if testing confirms quality holds |
| Underserved segments | Female solo travelers made up 11 percent of the travel industry | Look for a group the market treats as an afterthought, then confirm the need by talking to its members |
The figures above come from Sequoia’s guide to choosing a startup idea and Harvard Business School Online’s business idea guide.
Here is how to find startup ideas: start with a problem people already work around, then test it through customer interviews before building anything. An idea counts only once real behaviour backs it.
This playbook is one step in the wider path of how to start a startup, and it covers the stage before you commit to a product.
Where startup ideas come from
Startup ideas come most reliably from observed problems, improvements to existing products and models proven in other markets. Each source leads to a different kind of first test.
Problems people already work around
Sequoia’s guidance is that where pain is acute, people will be hacking shortcuts and cobbling together workarounds. A workaround is a stronger signal than a complaint, because it shows the person has already spent time or effort on the problem. A pain point that nobody has tried to fix is often not painful enough to pay for.
The practical habit is to look for frustrations and inefficiencies in everyday behaviour. Ingressive Capital advises founders to watch for the friction that a startup idea could address. Keep a running list of every problem you see and record the current workaround next to it. Judge the list later, once it has some length.
Improving something that exists
Harvard Business School Online suggests thinking of companies you admire and imagining how to improve their products. It frames the improvement around four lenses: delivery process, location, cost and customer experience. This route suits founders who know an industry well but have not yet found an unserved problem, because the starting product already proves that people buy.
Cost is the most tempting lens and the riskiest. The same HBS piece describes Wyze selling cameras for one-tenth the cost of leading competitors, and it adds that testing is needed to ensure product quality isn’t sacrificed for a lower price. If your idea rests on being cheaper, plan the quality test before the pricing page.
Borrowing from elsewhere
A proven idea from another market is a legitimate starting point. Crunchbase notes that there are very few truly original ideas, so the value lies in creating something useful rather than in novelty. It also warns that hunting for ideas inside one local bubble only produces ideas that interest people in that bubble. If you only talk to people like yourself, widen the circle to other cities, industries and customer types.
Customer interviews that surface pain points
A customer interview is a conversation held to learn how someone handles a problem today, not to pitch a solution. Ingressive Capital recommends engaging potential customers and listening to their feedback to understand their challenges and preferences.
Good interview questions follow the pattern of the questions that started well-known companies. Sequoia’s examples show how a founder turns an observation into a question:
- Ask why a behaviour exists. Drew Houston asked why people were always emailing large files to themselves. A repeated, awkward habit points to an unmet need.
- Ask where the friction is. Patrick Collison asked why accepting a credit card on the web was so hard. Phrase your question around the task, and let the interviewee name what makes it difficult.
- Ask what they do today. Ask what they use when the task goes wrong and what they tried first. Their answers reveal the workaround you would be competing with.
- Ask what job they are hiring a solution for. HBS cites Clayton Christensen’s idea of a job to be done, a problem or opportunity somebody is trying to solve. Questions framed this way keep the conversation on the person’s goal and away from your product.
Write down exact phrases, not summaries, and keep your own solution out of the conversation until the end. Interviewing people who differ in experience and context also helps, because it keeps you from hearing only the version of the problem that matches your assumptions. Turn the answers into a one-page problem brief: who has the problem, what the workaround is and how often it happens.
Testing an idea before you build
You test a startup idea by putting evidence between you and the build. Ingressive Capital lists building prototypes, conducting user interviews and gathering data as the core validation tools, and adds market research, surveys or pilot studies to demonstrate demand before pitching investors.
- Choose interviews when the problem itself is unclear. They are cheap and show whether the pain is real before any design work.
- Choose a prototype when the problem is clear but the solution is not. A rough version lets people react to something concrete.
- Choose surveys or pilots when you need a wider read. They show whether the pattern from a few conversations holds across more people.
- Study earlier attempts when the model exists elsewhere. Crunchbase observes that attempts in other markets amount to case studies that produce an immensely valuable kind of data, so you can learn what worked before spending your own money.
One pitfall is stopping at kind words. Compliments in an interview are not validation. A person who describes a workaround in detail, or who agrees to try a pilot, has given you stronger evidence. If an idea fails a test, return to your problem log and move to the next candidate.
Matching an idea to your skills
Match an idea to your skills by scoring each candidate on the same criteria. Ingressive Capital suggests evaluating every idea on market potential, feasibility, passion and expertise. The first two ask whether the problem is worth solving and whether you can build the solution. The last two ask whether you are the right person to do it.
Treat the criteria as a way to compare candidates, not as a pass mark. An idea that scores high on market potential but low on expertise may still be worth pursuing if you can recruit the missing skill. An idea that scores high on passion and expertise but has no visible demand needs more testing before it earns your time.
Sequoia adds two cautions. The first is not to create businesses only to serve yourself, so your own frustration should be confirmed by other people before it becomes an idea. The second is that great founders are described as unwilling to do anything else. Ask yourself honestly whether you would keep working on this problem when easier options appear.
Why starting from the problem works
Starting from the problem works because the problem is the part you can check against real people. HBS defines a job to be done as a problem or opportunity that somebody is trying to solve, and that definition gives you a fixed point to interview around. Ingressive Capital likewise says to start by identifying a problem or pain point your target market is experiencing.
It also keeps the idea in proportion. The UVA Foundry idea library states that the real value isn’t in the idea but in the execution. A problem you have verified, with a customer you can name, gives execution something to aim at.
After you choose an idea
Once an idea survives interviews and tests, two practical questions follow. If you are building with a partner, How To Divide Equity In A Startup: A Practical 2026 Guide To Fair Founder Splits – Startalliance covers how to split ownership. When you want a rough sense of what the company may be worth, Startup Valuation Calculator: Estimate Your Company’s Worth Fast — A Practical 2026 Guide – Startalliance is the place to start.
Frequently asked questions
Where can I browse ready-made startup ideas?
The UVA Foundry idea library offers dozens of free startup ideas browsable by category. Treat them as prompts to test against real customers, not as finished plans.
Should I judge ideas while I am brainstorming?
No. HBS gives a golden rule of brainstorming: there are no bad ideas. Write everything down first and evaluate afterwards, as Ingressive Capital also advises.
Is a small market a bad sign?
Not necessarily. Sequoia says what you really want is a small market today that will be big tomorrow, so a small starting market matters less than whether it is about to grow.
