How to find a startup idea: start with who pays and why now

How is a startup idea different from an app idea?

An app idea can be good because you want it to exist. A startup idea has to be good because someone else will pay for it repeatedly, and because you can reach more of those people than you can talk to personally. If you are looking for a side project to finish and enjoy, the guide on how to find app ideas fits better. This page is for the case where the goal is revenue.

That changes the starting question. For an app, ask “what would I use?” For a startup, ask three questions in this order:

1. Who pays? A specific role at a specific kind of organization, or a specific kind of person, with budget authority.
2. For what outcome? A result they can describe in their own words, such as fewer missed appointments or invoices paid sooner.
3. Why now? Something changed in the last few years that makes the old way worse or a new way possible.

If you cannot answer the first question in a sentence, you do not yet have a startup idea. You have a topic.

Where do good startup ideas come from?

Ideas that hold up tend to come from contact with a real workflow. The table lists sources you can mine in a few days, how to use each, and what to watch for.

Whatever you pick, write down the source and the date. If your idea came from one person’s complaint, you have one data point, and you need nine more before it counts as a pattern.

How do you find out who actually pays?

The user and the payer are often different people. A teacher uses a classroom tool, but a school or a district buys it. An employee uses an expense tool, but the finance lead chooses it. A startup idea needs the payer’s motive, not only the user’s annoyance.

• Ask who chose the current tool. The answer names the payer and tells you how long the decision took.
• Ask what they pay today. Even a rough range tells you where your price must land. If they pay nothing, because the work is done by hand, ask what the hand work costs in hours.
• Ask what would make them switch. Switching costs are often higher than the price difference.
• Look for a budget line. Software that replaces a labor cost, a legal risk or a lost sale is easier to buy than software that makes something nicer.

Consumer ideas can work too, but the price is lower and the audience must be reached through content or word of mouth. Many solo founders find business customers easier to reach and to charge, because a business can approve a monthly fee without a long debate. This is a tendency, not a rule, and you should test it against the buyer you actually have access to.

What does “why now” mean for a startup idea?

“Why now” is the change that makes the idea possible or urgent today and unremarkable five years ago. Without one, the idea likely already exists, or failed for a reason you have not found. Reasons are specific and checkable:

• A capability became cheap. Language models can now read messy documents, draft replies or classify requests at a cost a small product can pay for. Name the task the model does and how you would check its output.
• A rule or deadline changed. A new reporting duty, a platform policy, or a standard that businesses must adopt. Link to the rule itself and read it before building.
• A platform opened or closed. A marketplace, API or app store changed what third parties can do, leaving a group of customers underserved.
• Behavior shifted. A group started working in a new way, such as remotely or through a messaging app, and the old tools do not fit.
• Building got cheaper. AI coding tools let one person serve a niche that was too small for a team. A market that could not pay for a team may pay for one founder.

Be honest about the last reason. It makes it easier for you, and equally easy for everyone else. A small niche protected only by its size is a weak moat, so pair it with something harder to copy: access to the buyers, domain knowledge, data you collect, or integrations the buyers need.

How do you turn a rough idea into something testable?

Write a one-page memo before you do anything else. If you cannot fill in a line, that line is your next piece of research. Copy this template.

Buyer: [role, type of organization or person, size]
User (if different): [who actually uses it]
Problem in their words: [one quote or paraphrase from a real person]
What they do today: [tool, spreadsheet, person, or nothing]
Cost of the problem: [hours, errors, lost sales, risk]
Outcome I would deliver: [one measurable result]
Why now: [the specific change]
How I reach ten of them: [communities, lists, introductions]
First version: [one workflow, no more]
Price guess and why: [per month, per use, or per seat]
What would prove me wrong: [a specific result in a specific test]

The last line matters most. Decide what result would make you drop the idea before you start, so you cannot move the goalposts later. The app idea validation planner can turn the memo into interview questions and a first experiment, and how to validate an app idea covers the evidence log.

How do you compare several ideas before you commit?

Collect three to five memos, then compare them on the same few criteria. Be strict about evidence, not enthusiasm. A short list of criteria that works for a founder with limited time:

• Access. Can you reach ten buyers within a week?
• Pain. Did at least three people describe the problem unprompted?
• Budget. Does money already move around this problem?
• Buildable. Can you ship a first version alone in a month or less?
• Interest. Will you still care about this in a year of selling it?

If you are weighing software-as-a-service ideas, how to pick your first SaaS idea gives a weighted scoring table. For a list of current options to react to, see startup ideas for 2026. Use lists to provoke your own thinking, then check each against your own buyer access.

Which startup ideas should you avoid?

• Ideas with no payer. “Everyone could use it” means nobody has a budget for it.
• Ideas that need a crowd to work. Marketplaces and social products need two sides or a network before they are useful. They are hard for a first company.
• Ideas that depend on one platform’s goodwill. If a single company can switch off your access, plan for that on day one.
• Ideas where your edge is “it’s AI”. The customer buys an outcome. Describe the result, and mention the tool only if it matters.
• Ideas in areas with heavy rules you have not studied. Health, finance and children’s products carry obligations that should be understood before you build.

If you have an idea in one of these groups that you do not want to pursue, you can post it on the board. Someone with the right access or background might.