4.4
2,100+ reviews on G2
The way to find startup ideas is to look for problems, preferably ones you have yourself, rather than trying to think up ideas, as Y Combinator co-founder Paul Graham puts it. Paul Graham names three traits. Add a demand check before you build:

A problem you or people near you have.
With the skills you or a co-founder have.
Not obvious to everyone.
People would pay, and say so.
Keep a problem log in Evernote. The AI Assistant compares the problems and counts who would pay.
Each time you, a coworker or a customer hits an annoying, slow or expensive task, write it down the same day.
Every few weeks, read the log and mark problems that come up again, or that many people share.
Ask how they deal with it today and what it costs them. Listen for whether they already pay for a workaround.
Who else solves it, what they charge, and how many people have the problem. The SBA's market research questions are a good checklist.
A landing page, a manual version of the service, or a simple tool for a few users. See if anyone signs up or pays before building more.
Idea: appointment reminder texts for small clinics
Manual reminder calls show up in 9 of 15 entries.
Five other clinic managers: three make reminder calls by hand; two pay for add-ons they find hard to use.
Several reminder tools exist, mostly part of large scheduling suites.
Send reminders by hand from a spreadsheet for two clinics for a month, and ask if they would pay to keep it.
Common sources, with the question to ask about each.
| Source | Ask yourself |
|---|---|
| Your own work or hobby | What do I keep wishing a tool did? |
| Other people's complaints | What do people pay to work around? |
| A field you know well | What is missing that insiders would notice first? |
| New technology | What is now possible that was not a year ago? |
| Small software tools (micro SaaS) | What narrow task do a few hundred businesses do by hand? |
Paul Graham, "How to Get Startup Ideas" (2012); SBA, "Market research and competitive analysis".
4.4
2,100+ reviews on G2
4.4
8,200+ reviews on Capterra
4.4
73,000+ reviews on App Store
248M
Registered users
Look for problems rather than ideas, preferably problems you have yourself. Keep a log of them and look for ones that repeat.
A small software-as-a-service product that solves one narrow problem for a specific group of customers, often built and run by one person or a small team.
Look at tasks people in one niche do by hand or with spreadsheets, especially ones they complain about. Talk to them before you build.
Talk to people who have the problem, check what alternatives exist and cost, and run a small test, such as a landing page or a manual version, to see if anyone signs up or pays.
Talking to potential customers is how you learn whether an idea is worth pursuing. Ask them what they would change.
A problem nobody has, or one people do not care enough about to pay to solve. Ideas that only sound plausible, with no real user behind them, are a common trap.
Keep a problem log, your interview notes and your test plan in one notebook.
Write each problem in your log the day you see it. Ask the AI Assistant to compare them in a table: who has it, how often it comes up, and whether people already pay for a workaround.
Ask the Assistant for a repeating task to talk to one potential customer every week. The task reminds you each week.
Save each interview as a note. Ask the Assistant how many people already pay for a workaround, and it reads every interview note and cites them.
Log the problems you see, compare them, and keep every customer conversation in the same notebook.
Start for free