How to Kill a SaaS Idea in an Afternoon

Most founders do not test an idea. They look for support for it. By the time the support arrives, the decision is already made, and the six months after it are spent discovering that support is not the same thing as a purchase.

The method below is built to do the opposite. Each test is designed to kill the idea in under an hour. If the idea survives all five, you have earned the right to spend a weekend on it, not six months.

Test 1: Name the person who is stuck

Not the market. Not the segment. One person, or a role so specific that you could find ten of them by name this week.

If your answer is "small business owners" or "people who use spreadsheets", you have not finished the sentence. The records are full of requests written at that level of detail. The useful ones read like this: the office manager at a clinic who reconciles insurance payments by hand on Fridays. That version tells you where to look and what to ask.

A useful check: can you describe the person without mentioning your product? If you cannot, you have written a pitch, not a customer.

Test 2: Write down what they do today

Every real problem already has a workaround. Price it.

In the records, workarounds cluster into four types: a spreadsheet the owner maintains by hand, a group chat where requests disappear, a part-time person paid to keep the process moving, and an agency on a retainer. The cost is hours, errors, late payments, or a person who quit because of it.

If you cannot state the cost in a unit the buyer already tracks, the problem is not painful enough to buy against. What a problem statement must contain goes deeper on this section, because it is where most ideas quietly end.

Test 3: Find three people who complained without being asked

Not three people you interviewed. Three people who wrote the complaint in public before you arrived, in their own words, on their own time.

Our library holds 12,631 records collected from ten public founder communities between May 2025 and November 2026. The value of a record is not that someone wanted a tool. It is that someone described a workflow in enough detail to show they live inside it.

If you cannot find three unprompted complaints, you have a hypothesis. That is fine. A hypothesis needs a test, not a build.

Test 4: State the result that would make you stop

Before you build, write one sentence in this shape: "If fewer than X of the first Y people I ask do not do Z, I stop."

The numbers must be decided before you have feelings about the answer. This is the same discipline described in how to validate a SaaS idea before you build it, and it is the step founders skip most often, because the sentence is uncomfortable to write when you already love the idea.

A failure criterion you would not honour is not a criterion. It is decoration.

Test 5: Ask for money, not for an opinion

The last test is the only one that costs anything, and it is the only one that answers the real question.

Ask five of the people from test 3 to pay a small amount now, for a manual version of the outcome. Not a discount on the future product. Not a letter of intent. Money, this week, for a result you can deliver by hand.

You will learn three things in one afternoon. Whether they have the problem, whether they control a budget, and whether they trust you enough to pay before the product exists. Two of five paying within a week is a strong signal. Zero of five, with detailed enthusiasm and no card, is how this usually ends.

Five is not a sample. It is a screen, and it is designed to be wrong in the cheap direction.

Why this order matters

The tests are ordered by cost, and each one exists to protect the next. Naming a person makes the complaint search specific. Finding real complaints makes the failure criterion concrete. The failure criterion makes the money question honest, because you already know what answer ends the project.

Run them backwards and you get the usual outcome: a survey, a waitlist, a landing page with 400 emails, and a build that nobody buys. Every one of those steps feels like progress. None of them costs the buyer anything, so none of them predicts a purchase.

What survives

An idea that passes all five is not proven. It is funded by evidence instead of enthusiasm, and that is the most you can get for an afternoon.

Write the five answers down before you start. The written record is what you read in month two, when the work is hard and you want to change the answer.

The remaining risk is execution. That is the kind you can work on.

If you want to see which problems keep appearing in the records before you pick one, start with the build or do not build topic page. The market view is a useful second read, because it explains why the buyer changed shape even though the demand did not.

Sources

• SaaS Idea Finder idea library, 12,631 records from ten public founder communities, collected May 2025 to November 2026.
• Record count and date range measured from the production database on 2026-11-05.
• The five-test structure is our own method. No external benchmark is cited.