Validate a micro-SaaS idea in 7 days (without building it)
Most micro-SaaS ideas die from no demand, not bad code.
Building first is the expensive way to learn nobody wanted it. This is a 7-day sequence that costs under $100 and ends in a number you can act on: go, pivot, or drop it.
- 7 days
- Start to decision
- Under $100
- Total cost
- 5
- Conversations
- 0
- Lines of code
Day 1 · Write the one-sentence bet
Fill this in and you have a testable claim: "[Who] loses [how much time or money] doing [job] with [current tool], and would pay [price] a month to stop."
Example of a testable bet: "Freelance bookkeepers lose 4 hours a month chasing receipts in email, and would pay $19/month to stop."
Example of an untestable one: "Small businesses need better finance tools." There is nobody to ask and nothing to measure.
If you cannot name the person, you do not have an idea yet. You have a category. Go narrower until a specific human comes to mind.
Day 2 · Find where they complain
You are not looking for an audience to build. You are looking for one that already exists and already talks.
Search for your job-to-be-done plus a complaint word on: Reddit, X, Facebook groups, Slack and Discord communities, and G2/Capterra reviews of the incumbent. Reviews of a competitor are the highest-signal source on the internet: people describing exactly what they hate about a tool they already pay for.
Collect 20 verbatim quotes into a doc. Do not paraphrase. The exact words become your landing-page copy on Day 4.
Rule: if you cannot find 100 people complaining about this in public, the market is too small or the pain is too mild.
Day 3 · Talk to five of them
Five conversations, 20 minutes each. DM the people whose quotes you collected.
Ask only about the past, never the future. "Would you use this?" is worthless: people are polite. "What did you do the last time this happened?" is worth everything.
The four questions:
- Walk me through the last time you did [job].
- How long did it take?
- What did you try to fix it?
- What did you pay for, if anything?
The signal is question 4. Someone who has already spent money on this problem, even on a bad tool, even on a VA, has proven the budget exists. Someone who has only complained has proven nothing.
Day 4 · A landing page with a price
One page. Their words from Day 2 as the headline. A concrete outcome, not a feature list. A price on it.
Put the price on it. A page without a price measures curiosity; a page with a price measures intent. "$19/month" and a button that says Get access.
What happens on click: a form asking for their email and one question, "What are you using today?" That answer tells you who you are really competing with, which is usually a spreadsheet.
Build it in an afternoon. Do not design it. Design signals nothing at this stage.
Day 5-6 · Buy 200 visitors
Spend $50-100 to put the page in front of the exact people from Day 2. The cheapest honest traffic is usually the communities themselves, then narrow paid social.
Do not post it as a launch. Post it as a question in the places you found them: "I am building X for Y, does this match your problem?" Link it. That earns real feedback and real clicks at once.
You want roughly 200 targeted visitors. Fewer than that and the numbers on Day 7 are noise.
Day 7 · Read the number and decide
The only metric: of the visitors who were genuinely your target, how many gave you an email at a stated price?
One number, four verdicts. Find your rate and read across.
| Conversion | The verdict |
|---|---|
| Under 2% | No. Not a copy problem: there is no pain here at this price. Kill it, or go back to Day 1 with a different bet. |
| 2-5% | Maybe. There is something, but not enough to build on. Change ONE variable, the price or the audience, not both, then re-run Days 4-6. |
| Over 5% | Go. But not to code yet: the next section is the last test. |
| Over 10% | Go, and raise the price. Conversion that high on a cold page usually means you underpriced. |
Sell it once before you build
The strongest validation is not an email. It is a charge.
Go back to the people who signed up and offer to do the job manually for a month, at the price on the page. No product. You in a spreadsheet, doing it by hand.
If five people pay you to do it manually, you have a business and a specification: you now know exactly which steps to automate first, because you did them yourself and hated the same ones twice.
If nobody pays once the card is out, you just saved three months. That is the whole point of the seven days.
The costs, honestly
Domain: about $12. Landing page: free on any of the current builders, or an afternoon in code if you already write it. Traffic: $50-100. Total: under $120 and one week.
Compare that to the alternative most founders pick: three months of building, then discovering the same thing.
The uncomfortable part is not the cost. It is that the answer is usually no, and finding out in seven days feels like failure when it is actually the return on the exercise.
The whole point of the week is to spend $100 and seven days finding out what most founders pay six months and a product to learn.
Browse ready-made micro-SaaS ideasMore guides
- How to choose the tech stack for your productChoose your stack once. Live with it for years.9 min read
- The 5-section landing page that converts visitorsStop hoping visitors convert. Design it precisely.8 min read
- Web app or mobile app? A decision frameworkChoose the right platform before you start building.7 min read
- When to redesign your site (and when not to)A redesign is expensive. Make sure you actually need it.7 min read
- Stripe integration: from setup to first paymentGetting paid is the hard-easy part.8 min read
- What is an MVP? (And what it isn't)An MVP is the smallest experiment that answers your biggest question.6 min read