# How Do I Know If My Idea Is Good? The Honest Answer

> Source: [https://botensten.com/articles/how-do-i-know-if-my-idea-is-good](https://botensten.com/articles/how-do-i-know-if-my-idea-is-good) (canonical)
> Author: Botensten — Botensten, https://botensten.com
> Published: 2026-08-09

## TL;DR

Your idea is good if real customers already feel the problem and spend money or time on a worse fix today. Prove it before building: interview 8-10 people about their past behavior using The Mom Test method, then ask for one real commitment — a deposit, pre-order, or signed intent to buy. If two or more commit, build the smallest version. If everyone says 'great idea' but nobody pays, the idea is weak. CB Insights found 35% of startups fail from no market need; validation is how you avoid joining them.

You know your idea is good when at least five potential customers describe the same problem in their own words, admit what they already spend to work around it, and agree to a real next step — all before you write a line of code. Good ideas survive contact with paying buyers. Weak ones only survive contact with friends. Everything below is how to tell the difference fast and cheap.

## How do you know if your idea is good?

A good idea targets a specific group who already feel a painful problem and pay — in time, money, or clumsy workarounds — to fix it today. You confirm that by studying their past behavior, not by pitching. Rob Fitzpatrick's [The Mom Test](https://momtestbook.com) gives the core discipline: ask what someone did the last time the problem hit, and never ask "would you buy this?" Past actions are facts. Future intentions are flattery. If nobody has even tried to solve the problem yet, that is usually a sign the pain is too small to fund a business.

## What signals prove real demand?

Real demand shows up as behavior, not compliments. The strongest signal is a person already spending money, time, or reputation on a worse solution — a paid tool they hate, a manual spreadsheet, a freelancer on retainer. Rank the evidence you collect from weakest to strongest:

- "Great idea, you should build it" — near-zero value; people are just being nice.
- A specific pain story with dates and dollars attached — moderate; it's real but not yet a purchase.
- A pre-order, deposit, or signed letter of intent — strong; money changed hands or was promised.
- Budget already spent on a workaround they'd love to drop — strongest; the market is proven.

Chase the bottom two. Discount the top one, no matter how good it feels to hear.

## How do we validate ideas before building at Botensten?

We don't write production code until an idea clears three gates: a named customer, a problem they describe unprompted, and a committed next step. Because we ship software with AI every day, building a prototype is cheap — which is exactly why we distrust it. Cheap building tempts you to skip the conversation and fall in love with your own demo.

Here's a real trade-off we hit. We wanted to add a client-portal module to our stack and were sure customers wanted it. We ran eight interviews first. Six described the pain clearly, but only two had ever paid to solve it, and both wanted something smaller — a shared status link, not a portal. We shipped the small thing in a week. The portal we'd planned would have taken a month and served almost nobody.

What burned us earlier was the opposite habit. We once shipped a feature on a hunch with zero calls. It worked perfectly and nobody used it. That dead feature is the reason the three-gate rule exists now.

## Which validation method should you use?

Match the method to your riskiest assumption and your budget. Interviews test whether the problem is real; pre-sales test whether people will actually pay. Don't run all five at once — pick the one assumption that would kill the business if it's wrong, test that, and escalate only when it passes.

| Method | What it tests | Time | Cost | Signal strength |
|---|---|---|---|---|
| Customer interviews | Is the problem real? | 3-5 days | ~$0 | Moderate |
| Landing page + small ad | Will people click and sign up? | 3-7 days | $50-200 | Moderate |
| Concierge MVP (manual delivery) | Will they use the solution? | 1-2 weeks | Low | Strong |
| Pre-sale / letter of intent | Will they pay? | 1-2 weeks | ~$0 | Strongest |
| Fake-door button | Is there interest in a feature? | 1-3 days | Low | Weak-moderate |

David Bland and Alex Osterwalder's [Testing Business Ideas](https://www.strategyzer.com/library/testing-business-ideas) catalogs dozens more experiments, each ranked by cost and evidence strength.

## What are the signs your idea is bad?

The clearest sign is that nobody currently pays anything — in money, time, or effort — to solve the problem. [CB Insights' analysis of startup post-mortems](https://www.cbinsights.com/research/startup-failure-reasons-top/) found "no market need" was the top reason startups fail, cited in 35% of cases. Other red flags: everyone loves the idea but nobody will pre-order; you have to explain the problem before people recognize it; the only interested buyers are other founders; or the market is real but each customer is worth too little to reach profitably. A bad idea isn't always a bad concept — often it's a real problem attached to people who won't or can't pay.

## How fast can you validate an idea?

You can get a strong signal in five working days for under $100. You don't need a finished product — you need conversations plus one small commitment test. Run this sprint:

1. Day 1 — Write your single riskiest assumption as one sentence ("busy solo founders will pay $X to automate Y").
2. Day 2 — List 15 real people who have the problem and message them for a 15-minute call.
3. Days 3-4 — Run 8-10 Mom Test interviews about past behavior; count how many already pay to work around the problem.
4. Day 5 — Ask the three warmest for one commitment: a deposit, a pre-order, or a signed intent to buy.

If two or more say yes with money or a calendar hold, build the smallest version. If everyone says "love it" but nobody commits, change the idea, the audience, or the price — then run the sprint again.

## Related reading

- [How to Launch a SaaS Startup: the Honest Playbook](/articles/how-to-launch-a-saas-startup)

## Frequently asked questions

**How do I know if my idea is good?**

Your idea is good if a specific group already pays — in money, time, or workarounds — to solve the problem today. Confirm it by interviewing 8-10 of those people about their past behavior, then asking for one real commitment before you build.

**How many people should I talk to before building?**

Aim for 8-10 focused customer interviews. That's usually enough to spot whether the pain is real and repeated. If patterns are still unclear after 10, your target customer or problem is probably too vague.

**What is The Mom Test method?**

It's Rob Fitzpatrick's rule for customer interviews: ask about specific past behavior instead of pitching your idea. Questions like 'what did you do the last time this happened?' get honest facts; 'would you buy this?' only gets politeness.

**Should I build an MVP or validate first?**

Validate first. Interviews and a pre-sale cost almost nothing and can kill a bad idea in a week. Build the smallest MVP only after real people commit money or time to the problem.

**What's the difference between a good idea and a good market?**

A good idea can still fail in a bad market. What matters is whether enough reachable customers feel the pain and will pay enough to make the business profitable. Test the market's willingness to pay, not just the concept.

**How much does it cost to validate an idea?**

Often under $100 and five working days. Interviews are free, a simple landing page with a small ad runs $50-200, and pre-sales cost nothing but effort.

**What if people love my idea but won't pay?**

That means the idea is weak as a business, even if the concept is nice. Enthusiasm without payment is the most common false signal. Change the audience, the problem, or the price and re-test.