To validate a business idea before starting, run small tests that ask people to spend something real (money, time, or contact info) instead of just asking for opinions. In a single weekend you can run a landing-page smoke test, take pre-orders, interview 10 potential buyers, scan a marketplace for existing demand, and check competitor gaps. If two or more tests clear a defined threshold, build. If they don't, change the idea before you waste months on it.

The hard part isn't running the tests. It's being honest about the results. Most first-timers don't fail because the idea was bad. They fail because they only looked for "yes" and ignored every "no." This guide is built to stop that.

What "validated" actually means (and what it doesn't)

Validation is not a vibe. It's evidence that strangers will give up something they care about to get what you're selling.

There are three levels of signal, from weakest to strongest:

Signal What it proves How much to trust it
"That's a great idea!" Someone is polite Almost nothing
Email signup / waitlist Mild curiosity A little
Pre-order or paid deposit Real demand A lot

Notice the gap. A survey where 90% say "I'd totally buy this" is nearly worthless, because saying yes costs them nothing. One stranger handing you $20 for something that doesn't exist yet outweighs 100 enthusiastic comments. Design every test so a "yes" actually costs the person something.

There's also a difference between validating a problem and validating a product. The problem is whether the pain is real, frequent, and annoying enough that people already try to solve it. The product is whether your specific solution is the one they'll pay for. Do them in that order: a real problem with the wrong solution is fixable, but a great solution to a problem nobody has is not.

Before you start: set your kill criteria

This is the step everyone skips, and it's the most important one. Before you run a single test, write down the exact numbers that would make you walk away. Put them in a note on your phone right now.

Example kill criteria:

  • Fewer than 3 of 10 interviewees describe this problem unprompted → the problem isn't sharp enough.
  • Landing page converts below 3% of visitors to signups → the message or demand is weak.
  • Zero pre-orders after 50 real conversations → people won't pay.

Why decide this in advance? Once you're emotionally invested, your brain reframes every weak result as encouraging — "only one signup, but it was a really good one." Setting the bar before you see the data keeps your own enthusiasm from lying to you. If you have a co-founder or a sharp friend, ask them to hold you to these numbers. A devil's advocate beats a cheerleader here.

Test 1: The landing-page smoke test (cost: $0-$30)

Build a one-page site that describes your product as if it already exists, with a single clear call to action: "Get early access" or "Notify me at launch."

How to do it in an afternoon:

  1. Use a free or cheap builder (Carrd is about $19/year; Canva and Notion have free options).
  2. Write a headline that names the customer and the outcome: "Bookkeeping software for solo dog groomers who hate spreadsheets."
  3. Add one email capture field. That's it. No menu, no about page.
  4. Drive traffic: post in 2-3 relevant Reddit communities, Facebook Groups, or Discords where your buyer hangs out. Or spend $20-$30 on a tightly targeted ad.

Pass/fail threshold: Aim for a 3-5% visitor-to-signup rate from cold traffic. Below 2% means the offer isn't landing. Above 8% is a strong signal. You need at least 100 visitors before the number means anything; 20 visitors and 1 signup tells you nothing.

Test 2: The pre-sell (the only test that really counts)

A signup says "maybe." A payment says "yes." If your idea allows it, ask for money before you've built anything.

  • Physical product: Take pre-orders or a refundable deposit. "Shipping in 6 weeks, $10 deposit to reserve."
  • Service: Offer a "founding client" discount and book a paid first session.
  • Digital product or course: Sell it on Gumroad before it exists. Be upfront: "Launching next month, early-bird price now, full refund if I don't ship."

Always offer a no-questions refund. It protects your reputation, and a refund request is itself useful data.

Pass/fail threshold: For a higher-ticket service, 2-3 paying clients out of 30 conversations is strong. For a low-cost product, you want a steady trickle. The rule: a sale only counts if it comes from someone who isn't doing you a personal favor.

Test 3: Ten real customer interviews (cost: $0)

Talk to 10 people who have the problem, not 10 people who like you. The goal is to learn, not to pitch. Most first-timers ruin this by selling instead of listening.

How many interviews do you need? Ten is enough to spot a clear pattern. If 7 of 10 describe the same painful workaround, you've found something. If every answer is different, the problem isn't sharp enough yet. Stop counting heads and start counting patterns.

No network? Cold-recruit strangers. Post in a subreddit or Facebook Group: "Doing research on [problem], not selling anything. 15-min call this weekend? Happy to send a $5 coffee gift card for your time." You'll get takers. Discord servers and niche forums work too.

Copy-paste interview script

"Thanks for the time. I'm researching how people handle [task] — I'm not selling anything today, just learning.

  1. Walk me through the last time you dealt with [problem]. What happened?
  2. What was the most frustrating part?
  3. What do you use to handle it now? What's annoying about that?
  4. Have you ever paid for anything to fix this? What, and how much?
  5. If a magic wand fixed this, what would change for you?"

The gold is question 4. "Have you paid to solve this before" predicts future spending better than any hypothetical. Ask about the past, never the future. "Would you buy X?" gets you a polite lie. "When did you last buy something like X?" gets you the truth.

Test 4: The marketplace demand check (cost: $0)

People are already searching for, and buying, solutions to real problems. Go find that traffic.

  • Search your idea on Etsy, Amazon, or Udemy. Listings with hundreds of reviews mean proven demand — and the review complaints are your roadmap for doing it better.
  • Use a free keyword tool (Google's Keyword Planner, or just Google's autocomplete and "People also ask") to see if anyone searches for this monthly.
  • Browse subreddits and forums for the phrase "is there a tool that..." or "how do I..." Repeated questions = unmet demand.

Pass/fail threshold: You want evidence that money already moves here. Zero competitors is usually a red flag, not a green one — it often means there's no market. A crowded marketplace with frustrated reviewers is a much better sign. If you're worried the space is too full, read all my ideas already exist and what to do when your business idea already exists — competition is usually proof of demand, not a reason to quit.

Test 5: The competitor-gap scan (cost: $0)

Pick the top 3-5 competitors and read their 1- and 2-star reviews. This is the cheapest market research on earth. Each angry review is a customer telling you exactly where the gap is: too expensive, too slow, bad support, missing a feature.

Write down the three complaints you see most often. If you can credibly fix even one, you have a wedge. If every competitor has glowing reviews and no real gap, that's worth knowing before you spend money trying to out-compete them. Stuck for the wedge itself? How to come up with a business idea covers turning gaps like these into a focused offer.

B2B vs. consumer: the tests are not the same

Validation thresholds differ sharply depending on who you're selling to. Don't apply consumer logic to a B2B idea.

Consumer / B2C Business / B2B
What proves demand Volume — many small signals Depth — a few serious buyers
Strong signal 100+ signups, steady pre-orders 3-5 letters of intent or paid pilots
Best test Landing page + marketplace scan Direct interviews + pre-sell
Watch out for Hype that never converts One enthusiastic contact ≠ a market

For B2B, five letters of intent from real decision-makers beat 500 survey responses. For consumer products, you need volume — one fan can't carry the idea.

Want the weekend playbook and a new validation teardown each week? Subscribe to the howtostart.biz newsletter and we'll send the next one straight to your inbox.

Your weekend go/no-go scorecard

Run the tests, then score honestly. Give each a clear pass (1) or fail (0):

  • [ ] Problem is real: 7+ of 10 interviewees describe the pain unprompted
  • [ ] People pay for this: at least one stranger pre-ordered or paid a deposit
  • [ ] Demand exists: landing page cleared 3%+, or the marketplace shows active buyers
  • [ ] There's a gap: you found a fixable, repeated complaint about competitors
  • [ ] It clears your kill criteria: none of your pre-set deal-breakers were hit

Score of 4-5: Strong signal. Start building a minimum version this week. Score of 2-3: Promising but unproven. Adjust the offer or the audience and re-test one weak area. Score of 0-1: Pivot or drop it. That's not failure — that's a weekend that just saved you six months.

When do you stop validating and start building? When at least two independent tests pass and someone has paid you real money. Past that point, more validation is just procrastination in disguise. Build the smallest real version, put it in front of buyers, and let the market keep grading you.

Frequently Asked Questions

How do I validate a business idea with no money or audience?

Use the free tests. Cold-recruit 10 interviewees from Reddit, Discord, or Facebook Groups (offer a $5 gift card for their time). Scan Etsy, Amazon, and forums for existing demand, and read competitor reviews for gaps. A free landing page plus organic posts in 2-3 communities validates demand for $0. You don't need an audience — just go where your buyers already gather.

How many customer interviews do I actually need?

Ten is usually enough to see a pattern. You're counting how many describe the same problem and the same workaround, not counting heads. If 7 of 10 do, that's signal. If answers scatter, the problem isn't focused enough, and more interviews won't fix that.

What counts as proof someone will pay — a survey, a waitlist, or a payment?

A payment, every time. Surveys and waitlists cost the person nothing, so "yes" is cheap and unreliable. A pre-order, deposit, or paid pilot is the only signal where someone risks real money.

How do I know when to stop validating and start building?

When at least two tests pass your thresholds and one stranger has paid you real money. Past that point, more research is procrastination. Build the smallest usable version, ship it to those early buyers, and learn from real usage instead of more hypotheticals.