Most advice on how to validate a startup idea stops at "talk to customers." This guide gives you the tools to actually do it: a hypothesis sheet, a 14-day plan, an interview script, a demand test, a commitment ladder and a go/no-go scorecard. It is written for first-time and bootstrapped founders, and for small business owners testing a new offer, not only venture-backed software teams.
Sources were checked in September 2026. The legal and ad-policy notes below are general information, not legal advice, and the rules change, so confirm details on the linked official pages.
What It Means to Validate a Startup Idea
Startup idea validation is evidence that a specific buyer has a frequent, painful problem and will commit money, time or reputation to solve it. Compliments, likes and survey "yes" answers are not validation, because they cost the person nothing.
| Concept | Question it answers | When | Evidence |
|---|---|---|---|
| Market research | How big is the market, who competes and what do people pay now? | Before and alongside validation | Public data, competitor pricing, reviews and search trends |
| Idea validation | Will this specific buyer commit to this specific offer? | Before you build | Interview patterns, sign-ups from real buyers, deposits, letters of intent and paid pilots |
| Product-market fit | Do customers keep using, paying and referring? | After launch | Retention, repeat purchases, referrals and survey results |
Skipping this step is expensive. In CB Insights' March 2026 analysis of 431 venture-backed companies that shut down since 2023, poor product-market fit was cited for 43% of the 385 shutdowns with an identifiable cause, and running out of capital for 70% (many cited more than one reason). Those are VC-backed startups, not all small businesses, but the lesson carries over: build before you have evidence and you spend cash learning what a few weeks of interviews could have told you. For the full failure data, see why startups fail.
Before You Start: Write Your Riskiest Assumptions and Kill Criteria
Every idea rests on a handful of guesses. Write them down on one page before you talk to anyone, so you know what each conversation and test is meant to prove.
| Hypothesis sheet field | What to write |
|---|---|
| Target customer | Role, industry or life situation, and location. "Owners of independent dental practices in Ohio" beats "small businesses." |
| Problem and trigger | What goes wrong, and the event that makes them notice it. |
| Current solution | What they do today, including spreadsheets, a freelancer or doing nothing. |
| Current spend | The money or hours they put into the problem now. |
| Your offer | One sentence: what you do, for whom, and the result. |
| Target price | The price you will test and how often they would pay it. |
| First channel | Where you will reach your first 50 buyers. |
Then rank your assumptions by risk and test the riskiest one first:
- Desirability: do these buyers want the problem solved badly enough to act?
- Viability: will they pay enough, often enough, to cover your costs?
- Feasibility: can you deliver the result at that price with the skills and money you have?
Finally, write pass and fail thresholds, your kill criteria, before the first interview. Once results come in, it is tempting to reinterpret a weak signal as a good one. Numbers written in advance stop that.
Illustrative example
How to Validate a Startup Idea in 14 Days: The Day-by-Day Plan
Here is how to test a business idea in two weeks of focused work. Each block ends with an output you can show someone else.
| Days | Goal | What you do | Output |
|---|---|---|---|
| Days 1-2 | Frame the bet | Fill in the hypothesis sheet, do the desk research in Step 1, write your kill criteria and list 50 people to contact | A one-page hypothesis sheet and a contact list |
| Days 3-8 | Learn the problem | Recruit and run 15 to 20 problem interviews; write up notes the same day | Interview notes and a summary of patterns |
| Days 9-10 | Make an offer | Write a one-page offer in your buyers' words and put it live with a sign-up or request form | A live demand test |
| Days 11-13 | Ask for commitment | Send traffic to the page, follow up with interviewees and ask for deposits, letters of intent or pilots | Sign-ups, booked calls and commitments |
| Day 14 | Decide | Score the evidence on the scorecard and choose go, pivot or stop | A written decision and next step |
Two caveats. Enterprise B2B, regulated, hardware and marketplace ideas usually need longer: buyers take weeks to reply, rules limit what you can promise, and a marketplace has two sides to validate. And 14 days proves demand for the next step, not a whole company. A pass means "worth building the smallest version," nothing more.
Step 1: Desk Research in an Afternoon (Demand, Competitors, Market Size)
Check search demand
Google Trends is useful for direction: whether interest in a problem is rising, falling or seasonal, and where it is strongest. It does not tell you how many people buy. Google Keyword Planner gives estimates of monthly searches, but it takes some setup: Google Ads Help says you must complete account setup by entering billing information to use features like "Get ideas for new keywords."
Mine competitor complaints
Read one- to three-star reviews of the tools or services your buyers use now, plus Reddit threads, niche forums and, for software, G2 and Capterra reviews. Copy the exact phrases people use. They become interview prompts and, later, your landing page headline.
Size the market from the bottom up
Multiply the buyers you can realistically reach by a realistic price and by how often they buy. That beats a top-down "1% of a billion-dollar market" guess because every number in it can be checked.
Illustrative example
Use free US data
The SBA's market research and competitive analysis guide lists free data sources, including the Census Bureau's Census Business Builder, which offers demographic and economic data by location and type of business. That is handy for local service ideas. One warning: desk research can kill an idea, but it cannot confirm one. Only buyers can do that.
Step 2: Customer Interviews That Don't Lie to You (Script Included)
Rob Fitzpatrick's The Mom Test is built on three rules that keep people from telling you polite lies:
- Talk about their life, not your idea.
- Ask about specifics in the past, not opinions about the future.
- Talk less and listen more.
How many interviews do you need?
Plan on 15 to 20 per customer segment. Qualitative research offers a useful benchmark: in a 2017 health research study by Hennink, Kaiser and Marconi of 25 in-depth interviews, 9 interviews were enough to surface the range of issues, but 16 to 24 were needed to understand them well. Treat 10 to 15 as a floor for a fast first pass, and stop when new conversations stop surfacing new problems.
Where to find interviewees
- Second-degree contacts: ask friends for introductions, not opinions.
- LinkedIn searches by job title, industry and location.
- Niche communities and subreddits, following each community's rules on research requests.
- Industry associations and local business groups.
- People who left detailed reviews of competitors.
Copy-ready outreach message (no pitch)
Customer discovery interview questions: the script
Ten questions in three parts, for a 20- to 30-minute call. Record only with the other person's permission.
- Opening: Walk me through your role and a typical week.
- Where does [problem area] fit into that week?
- Past behavior: Tell me about the last time [problem] happened.
- What did you do about it?
- What did that cost you in time or money?
- What else have you tried, and why did you stop?
- What don't you like about how you handle it now?
- Who else is involved when you decide to fix or buy something for this?
- Close: Who else should I talk to about this?
- If we build something that does [result], would you be open to a follow-up to see pricing, or to a paid pilot?
Afterward, sort what you heard into weak and strong signals. Only the right-hand column counts toward your kill criteria.
| Weak signal (discount it) | Strong signal (count it) |
|---|---|
| "That's a great idea." | They describe the problem before you mention it. |
| "I would definitely use that." | They built a workaround: a spreadsheet, a hire or a patched-together tool. |
| "Let me know when it launches." | They already pay for a partial fix and complain about it. |
| Likes, follows and survey "yes" answers | They ask what it costs or when they can start. |
| Promises about the future | They introduce you to another buyer or commit money, time or reputation. |
Step 3: Run a Demand Test (Landing Page, Waitlist or Fake Door)
Interviews tell you the problem is real. A demand test tells you whether strangers act on your offer. Build a one-page offer with:
- A headline in your buyers' own words, taken from interviews.
- Three benefits tied to the problem, not a feature list.
- A price or a "from" price.
- One call to action: join the waitlist, request early access, book a call or place a deposit.
For traffic, start with direct outreach to buyers you did not interview, then share the page in relevant communities. A small paid test is optional; if you run one, set a hard daily cap. A "fake door" test works the same way inside an existing product or site: a button for an offer that does not exist yet, which leads to an honest "coming soon, join the list" message.
How to read the results
Set your target before the page goes live. For context, Unbounce puts the median landing page conversion rate at 6.6% (Q4 2024 data from 41,000 landing pages). That figure describes live pages of real businesses, not a pass mark for a pre-launch test, and traffic that does not match your buyer makes any rate meaningless. Count only sign-ups from people in your segment.
Guardrails
- Tell sign-ups it is pre-launch on the page and in the confirmation message.
- If you use Google Ads, its Unavailable offers policy (part of Misrepresentation) does not allow ads that promise products, services or offers that are unavailable or not easily found from the destination. Advertise what the page offers, such as a waitlist, early access or a pre-order, not "buy now."
Unsure whether the test needs one page or several? See landing page vs website.
BizTech's own service (disclosure)
Step 4: Ask for a Commitment (Pre-Sale, Deposit, LOI or Paid Pilot)
A commitment costs the buyer something. The higher up this ladder you get, the stronger your evidence.
| Commitment (weakest to strongest) | What it proves | Best for |
|---|---|---|
| 1. Email sign-up | Interest in the message, at no cost to them | Consumer products, SaaS waitlists |
| 2. Booked call | They will give you their time | B2B, professional services |
| 3. Letter of intent | A named buyer puts intent in writing (usually non-binding) | B2B services and software |
| 4. Refundable deposit | They will part with money, at low risk | Consumer products, local services |
| 5. Paid pilot | They pay for a limited version and judge the result | B2B software and services |
| 6. Full prepayment | The strongest signal short of repeat sales | Pre-orders, services, courses |
- B2B: ask for a letter of intent or a small paid pilot with a defined scope and end date.
- Consumer: take a refundable deposit or a pre-order, with the refund policy and delivery estimate in plain words.
- Service businesses: run a concierge test. Deliver the service by hand to your first three to five paying customers before you build systems, software or a team.
Legal note: pre-orders of physical goods (general information, not legal advice)
Step 5: Score the Evidence With a Go/No-Go Scorecard
How do you know if your startup idea is good? Not when people like it, but when the evidence clears the bar you set on day one. Score each criterion 0, 1 or 2.
| Criterion | 0 | 1 | 2 |
|---|---|---|---|
| Problem frequency and pain | Rarely mentioned or mild | Some describe it; it comes up now and then | Most describe it unprompted; it is frequent and costly |
| Money or time already spent | None | Some time, little or no money | They pay for a fix now or built their own workaround |
| Reachable channel | You could not reach buyers | You reached them slowly or at high cost | You reached them cheaply through a channel you can repeat |
| Commitments collected | None | Booked calls or letters of intent only | Deposits, paid pilots or prepayments |
| Tested price vs cost to deliver | The price does not cover delivery | It covers delivery with a thin margin | A healthy margin at a price buyers accepted |
| Gap vs alternatives | No clear difference | Better on one point some buyers care about | Clearly better for this segment |
| Founder fit and access | No access to buyers and no relevant skills | Some access or some relevant skills | Strong access to buyers and relevant skills |
Decision bands (an illustrative rubric, not a validated model)
- Go: 11 to 14 points and at least one real commitment (a deposit, paid pilot or prepayment). Build the smallest version you pre-sold.
- Pivot: 7 to 10 points, or 11+ with no real commitment. Narrow the segment, change the price or change the offer, then rerun the weakest test.
- Stop: 0 to 6 points, or a 0 on problem pain. Keep your notes and move on.
Illustrative example: bookkeeping for independent dental practices
This is a hypothetical founder with made-up numbers, scored after 15 interviews and a two-week demand test.
| Criterion | Evidence (hypothetical) | Score |
|---|---|---|
| Problem frequency and pain | 10 of 15 practice owners described month-end bookkeeping problems unprompted | 2 |
| Money or time already spent | 7 already pay a part-time bookkeeper or spend weekends on the books | 2 |
| Reachable channel | Reached through a state dental association newsletter and LinkedIn, but slowly | 1 |
| Commitments collected | 2 paid pilots at $300 for one month's close | 2 |
| Tested price vs cost to deliver | Pilots sold at $300 for about 6 hours of work; the planned $500 monthly price is not yet tested | 1 |
| Gap vs alternatives | Dental-specific reports, but generalist bookkeepers are easy to find | 1 |
| Founder fit and access | The founder spent five years managing a dental office | 2 |
Total: 11 of 14, with two paid pilots, so the rubric says go, narrowly. The weak spots, channel and differentiation, become the first things to fix while building the smallest version.
Startup idea validation checklist
- Hypothesis sheet written and riskiest assumption named
- Kill criteria written before the first interview
- 15 to 20 problem interviews in one segment, with notes logged
- Most interviewees described the problem unprompted
- One-page offer live with a price and one call to action
- Demand test shown to buyers who match your segment
- A real commitment requested: deposit, LOI, pilot or prepayment
- Scorecard filled in and a go, pivot or stop decision written down
How to Validate a Business Idea by Type: Local Service, SaaS, Product and B2B
The steps stay the same, but the best test and the strongest signal change with what you sell.
| Business type | Best first test | Strongest commitment signal | Biggest trap |
|---|---|---|---|
| Local service | A one-page site or Google Business Profile, plus direct outreach to nearby buyers | Pre-booked first jobs, ideally with a deposit | Friends and neighbors booking as a favor |
| SaaS | Problem interviews plus a clickable mockup | A pre-sale or paid pilot at a founding price | A waitlist that never converts to paying users |
| Physical product | A pre-order or crowdfunding page (the FTC rule in Step 4 applies) | Paid pre-orders | Underestimating landed cost: manufacturing, shipping, duties and returns |
| B2B services and software | Separate interviews with the budget holder and the day-to-day user | A signed letter of intent or a paid pilot | Users love it, but the budget holder will not pay |
| Marketplace | Test each side separately, often by matching people by hand | Suppliers commit time or listings and buyers pay | Validating only one side |
If you are working out how to validate a SaaS idea, watch the workflow, not the wish list. Buyers who export data into spreadsheets, stitch tools together or pay someone to do the task by hand have a problem worth solving. Ask them for a paid pilot or an annual pre-sale at a founding price, not for a spot on a waitlist.
7 Validation Mistakes That Create False Positives
- Asking friends and family. They want to support you, so their answers skew positive. Interview strangers who match your segment.
- Pitching before listening. Once you describe your solution, people react to it instead of telling you about their problem.
- Asking hypothetical "would you use it" questions. Opinions about the future are cheap. Ask what they did the last time the problem came up.
- Running surveys before interviews. A survey only measures the options you thought to include. Interview first, then survey to size what you heard.
- Counting sign-ups or likes as sales. A free email address is the lowest rung on the commitment ladder. Ask for something that costs the buyer money, time or reputation.
- Moving the goalposts after the results. If you lower the bar once the numbers come in, you have stopped testing. Write kill criteria first and honor them.
- Trusting an AI validator score instead of buyers. An AI idea score or AI-generated persona can help you prepare, but it is not evidence that anyone will pay.
For mistakes beyond validation, such as hiring too early or ignoring cash flow, see our list of common startup mistakes. If you want AI help with the research and note-taking around validation, our guide to AI tools for startups covers that.
What to Do After Your Idea Passes (or Fails)
If it passes
- Build the smallest version that delivers what you pre-sold, and keep the price you tested.
- Register and set up the business: our guide on how to start a business in 30 days covers the setup. Founding in Texas? See how to start a business in Houston for state and city filings.
- Plan the stages ahead with a startup roadmap and the launch with a go-to-market strategy for startups.
- After launch, check product-market fit. One common test is Sean Ellis's survey, described by First Round Review: companies with strong traction almost always had more than 40% of users answer "very disappointed" to losing the product. It is a post-launch check, not a pre-launch test.
If it fails
Change one variable and rerun the weakest test: a narrower segment, a different problem you heard in interviews, a new price or a new channel. If nothing clears the bar, stop. Keep your interview notes: problems people mentioned in passing can point to your next idea.
Get a second opinion
Free help exists. The SBA says SCORE mentors offer advice at no cost, and America's SBDC says local Small Business Development Centers offer no-cost business consulting and low-cost training.
If you want a structured outside review, our Basic Advisory ($400 one-time), which includes startup idea validation, is a structured review of your concept, target market and initial business model assumptions, with documented output; final scope is confirmed at kickoff. It does not replace talking to your own buyers, but it can pressure-test your hypothesis sheet and scorecard. Learn more about our startup advisory services.
How to Validate a Startup Idea: FAQ
How long does it take to validate a startup idea?
The plan in this guide is built to reach a go, pivot or stop decision in 14 days of focused work, which suits most service and software ideas. Part-time founders can spread the same steps over a few more weeks. Enterprise B2B, regulated, hardware and marketplace ideas usually take longer, because buyers move slowly, rules apply or both sides of a market need testing.
How many customer interviews do you need to validate an idea?
Aim for 15 to 20 interviews per customer segment, with 10 to 15 as a floor for a fast first pass. In a 2017 study, Hennink, Kaiser and Marconi found that 9 interviews surfaced the range of issues, but 16 to 24 were needed to understand them well. Stop when new conversations stop surfacing new problems.
Can you validate a business idea for free?
Mostly, yes. Customer interviews, desk research with free Census Bureau and SBA resources, and a waitlist built with a free form tool cost only your time. Google Keyword Planner requires a Google Ads account with billing information entered. A small paid traffic test and a one-page website are optional extras, not requirements.
What is the difference between idea validation and product-market fit?
Idea validation happens before you build: it shows that a specific buyer has the problem and will commit money or time to solve it. Product-market fit comes after launch, when customers keep using, paying and referring. A common post-launch check is Sean Ellis's survey, where companies with strong traction usually see more than 40% of users say they would be very disappointed without the product.
Can AI validate my startup idea?
AI can speed up the work around validation: summarizing competitor reviews, drafting an interview script, grouping interview notes and writing landing page variants. It cannot validate the idea itself, because a model cannot commit money on your buyers' behalf, and an AI-generated persona is not a customer. Use AI to prepare and analyze, and real buyers for evidence.
Should I ask people to sign an NDA before I share my startup idea?
Usually not. Many potential customers and investors will not sign one for an early conversation, and asking can end it. Share the problem, not the secret sauce. If the idea is a patentable invention, see a patent attorney before any public disclosure: US law gives inventors a one-year grace period after their own disclosure, while Europe has no general grace period.
Patent sources: USPTO, MPEP 2152 (35 U.S.C. 102(b)(1)) and the European IP Helpdesk on the EPO's grace period survey. This is general information, not legal advice.
The Bottom Line
How to validate a startup idea comes down to three habits: evidence over opinions, commitments over compliments, and thresholds set before you look at the results. Two weeks of interviews, one demand test and a direct ask for money will tell you more than months of building.
It is also how to validate a business idea before starting anything expensive: run the 14-day plan, fill in the scorecard and let the numbers decide. For a second set of eyes, Basic Advisory ($400 one-time, part of our startup advisory services) reviews your concept, target market and business model assumptions, and a $50 one- or two-page same-day site can host your demand test.
Ruhan Rafiq Bhaleshah
Chief Technology Officer
Ruhan Rafiq Bhaleshah leads the technical vision at BizTech, overseeing full-stack development, AI-powered workflows, and ensuring engineering teams collaborate with design and SEO specialists.