Skip to main content
Book a scoping call
Building & Shipping Products

How to Validate a Startup Idea Before Building

By Obaid Tariq, Founder & Lead Engineer · Published

Disclosure: AI Dev Planner sells validation sprints, so we have a commercial interest in this topic. The method below works the same whoever runs it, and a STOP verdict is a valid outcome even when we are the ones delivering it. See our disclosure policy.

Validation is founder work, and it comes before any build contract. Its job is to answer one question with evidence: is this problem painful, frequent and expensive enough that people will pay to solve it? If the answer is no, or not yet proven, the cheapest thing you can do is find out now.

A good validation phase ends in a written verdict, not a feeling. This guide sets out the five steps, what counts as evidence, and what each verdict means for your next move.

How do I validate a startup idea before building anything?

Start with the problem, not the solution. Interview at least five potential users about how they handle the problem today, test whether the pain is frequent and expensive enough to pay for, and check what competitors and workarounds already exist. Then test demand with the cheapest experiment that gives a real signal, such as a landing page, a clickable prototype or a manually delivered version of the service.

The five validation steps

  1. Write the problem in one paragraph. For a specific customer, doing an important task is difficult because of a current obstacle. If you cannot write it without listing features, you are not ready to validate yet.
  2. Talk to real potential users. Walk them through the problem, not your solution. Ask what they did the last time it happened, what it cost them in time or money, and what they use today.
  3. Test the pain. Ask whether someone would use this weekly, whether someone would pay for it, and whether the pain is strong enough that they would switch from what they use now.
  4. Check competitors and market gaps. Study direct competitors, indirect alternatives and the option of doing nothing. The goal is to find what customers need that existing products do not adequately provide.
  5. Run the cheapest useful experiment. A landing page with a concrete offer, a clickable prototype, a manually delivered version of the service, or a pilot conversation. Pick the experiment that tests your riskiest assumption, not the one that is easiest to run.

What counts as evidence, and what does not

Strong and weak validation evidence
EvidenceStrong signalWeak signal
User interviewsRecent behaviour, costs and workarounds described in detailPraise for the idea with no behaviour behind it
Demand testSign-ups, pre-orders or paid pilots from strangersLikes and encouragement from friends and colleagues
Competitor checkA gap customers complain about in reviewsNo competitors found, with no explanation why
Willingness to payA price discussed and a payment offeredWould pay in theory, someday

Enter agency conversations with evidence, not just enthusiasm. A good agency will challenge your assumptions, so come with assumptions worth challenging. Our validation sprint runs these five steps in one to two weeks and ends in a written verdict.

What does a validation verdict, KEEP, PIVOT, RESEARCH FURTHER or STOP, mean?

KEEP means the evidence supports building the MVP as scoped. PIVOT means the problem is real but the solution, audience or business model must change. RESEARCH FURTHER means the evidence is too thin to decide and one more targeted test is needed. STOP means the idea does not justify a build budget, and stopping now is a success because it saves the full build cost.

The four validation verdicts and what happens next
VerdictWhat it meansWhat happens next
KEEPThe problem is real, painful and underserved, and the proposed solution fitsScope one narrow MVP and price that scope with each shortlisted agency
PIVOTThe problem is real, but the solution, audience or business model is wrongChange one element, re-test it, and get a new verdict
RESEARCH FURTHERThe evidence is too thin to decide either wayRun one more targeted test on the single biggest unknown, then decide
STOPThe idea does not justify a build budgetStop. Keep the research; the money saved is the return on validation

A STOP verdict is not a failure. You spent a discovery fee to save a full build budget, and you keep the research. PIVOT and RESEARCH FURTHER verdicts are equally valuable: they redirect spending to where demand actually is. Our discovery and validation step costs $2,500-$5,000 as a separate paid step and ends in a written verdict; see our pricing for the full terms. Then scope the build with MVP development services.

If the verdict is KEEP, the next question is what to build first. Read MVP vs prototype vs proof of concept before you commit to a build, and see the full decision in our guide to the best software agency for your idea.

How much does validation cost, and how long does it take?

A focused validation sprint takes one to two weeks. Our discovery and validation step costs $2,500-$5,000 as a separate paid step. A design-first clickable prototype for user testing can cost around $16,000 as a market figure from the four AI reports this cluster synthesises, so test with interviews and a landing page before you pay for a prototype.

Validation is deliberately cheap next to a build. A one-to-two-week sprint with user interviews, competitor research and one demand experiment is usually enough for a verdict on a single product idea. Spending more than that before a verdict usually means the scope of validation has drifted into building, which is the mistake validation exists to prevent.

If the riskiest assumption is technical rather than commercial, a small technical spike or proof of concept can be the right validation spend instead. See MVP vs prototype vs proof of concept for how to choose between them.

What if potential users will not talk to you?

Talk to fewer people, better. One detailed conversation with a person who has the problem beats twenty polite chats with people who do not. If nobody will talk, that is itself a signal: either you are reaching the wrong people, or the problem is not painful enough to discuss. Try a landing page with a concrete offer, a concierge version of the service delivered by hand, or a pre-sale conversation instead.

A concierge test is often the fastest way forward: deliver the service manually to a handful of real customers, using spreadsheets, email and phone calls instead of software. If people will not even accept a free manual version, they will not adopt a built product either. If they do use it, you learn exactly which parts matter before you pay to automate them. Our guide to first SaaS customers covers how to find those first users, and the next step after a KEEP verdict is in our guide to the best software agency for your idea.

Do not buy validation by surveying a large audience with a vague question. A hundred responses to do you like this idea tell you less than five conversations about what someone did last week, what it cost them, and what they would pay to fix it.

Frequently asked questions

How many users should I talk to before I decide?
At least five potential users, and keep going until you stop hearing new problems. Five is a minimum for a first verdict, not a sample that proves a market.
What if users say they like the idea?
Liking is not evidence. Ask what they did the last time the problem happened, what it cost them, and whether they would pay or switch now. Behaviour and money are evidence; praise is not.
Can an agency validate the idea for me?
An agency can run discovery, competitor research and prototypes, but the founder still owns the customer understanding. Come to discovery with user evidence, and treat an agency that never challenges your assumptions as a warning sign.
What happens after a KEEP verdict?
Scope one narrow MVP around the core journey, price that scope with each shortlisted agency, and build in weekly sprints. A KEEP verdict justifies the next step, not the full original wish list.

Sources

  1. ChatGPT: Best Software Agency to Turn Your Idea Into a Live Product (22 November 2026) — Validation activities and deliverables, and the go or no-go recommendation that ends discovery. Accessed 22 November 2026.
  2. Gemini: The Founder's Guide to Choosing the Best Software Agency (22 November 2026) — Idea validation and product discovery as the first phase, with a validated value proposition as the deliverable. Accessed 22 November 2026.
  3. DeepSeek: The Founder's Guide to Choosing the Best Software Agency (22 November 2026) — Validation as founder work before hiring anyone, and the paid discovery phase as the key safeguard. Accessed 22 November 2026.
  4. Perplexity: Best Software Agency to Turn Your Idea Into a Live Product (22 November 2026) — Validation experiments, evidence log, and the decision gate before funding a build. Accessed 22 November 2026.
  5. GOV.UK Service Manual: How the discovery phase works — Understand the problem before committing to build; do not start building in discovery; stopping after discovery is not a failure. Accessed 22 November 2026.

Recommendations without a source are AI Dev Planner’s own engineering opinion. Platform features and prices change; check the linked pages for the current position. Found an error? See our corrections policy.

Not sure your idea is ready to build?

A short validation sprint gives you a written verdict and a scoped MVP, or tells you to stop before you spend a build budget.