A polished landing page, a clever name and encouraging comments from friends are not validation. They are signals of effort, not proof of demand. If you want to know how to validate business ideas properly, start with one hard question: will a defined group of people give up money, time or attention to solve this problem now?
That question protects scarce cash, stops teams building features nobody needs and gives you a stronger basis for every next decision. For lean businesses, validation is not a drawn-out research project. It is a disciplined way to replace assumptions with evidence before you commit months of work.
Start with a painful, specific problem
Weak ideas are usually broad: “small businesses need better marketing” or “people want to save time”. Strong ideas identify a customer, a situation and a cost of doing nothing. For example: “Independent e-commerce brands losing margin to rising ad costs need a faster way to identify which products deserve paid spend.”
Specificity matters because customers do not buy vague improvement. They buy relief from an expensive, frustrating or risky problem. The more clearly you can describe the current workaround, the closer you are to a commercially useful idea.
Speak to people who have experienced the problem recently. Ask what happened, what they tried, how much the issue cost and who approved the spend. Avoid asking whether they like your concept. Most people will be polite. Their past behaviour is more valuable than their opinion of your future product.
A productive conversation sounds like this: “Talk me through the last time this happened.” Then ask, “What did you do instead?”, “What did that cost?” and “What would have made you change course?” You are looking for detail, urgency and evidence that they have already tried to solve it.
Define the customer and the buying trigger
Not every frustrated person is a customer. A founder may feel the pain, but a finance lead may control the budget. A marketing manager may use the tool, while the agency owner decides whether to buy. Validation improves when you separate the user, buyer and influencer.
Create a narrow first segment rather than trying to serve every possible customer. You might start with UK agency owners managing five to 20 staff, Shopify operators turning over more than £500,000, or B2B SaaS founders preparing for their next funding round. A narrow segment makes outreach, messaging and pricing tests far more credible.
Then identify the trigger that turns a persistent problem into a buying decision. It could be a missed sales target, a new compliance requirement, a staff departure, a cash-flow squeeze or an upcoming product launch. Without a trigger, even a real problem may stay low priority.
Test willingness to pay, not just interest
The most common validation mistake is treating attention as revenue. A large survey response rate, social media likes or a long list of email sign-ups can be useful, but none proves a viable business.
The strongest early evidence is a financial or operational commitment. That might be a pre-order, a paid pilot, a deposit, a letter of intent, a booked implementation call or access to the customer's data and team. The right test depends on what you sell. A £20 consumer product can be tested with a simple checkout page. A £10,000 B2B service may require a scoped pilot and a direct sales conversation.
Do not give the offer away for free unless free access is part of a deliberate strategy. Free users often praise products they will never purchase. If you do run a free trial, ask for a meaningful commitment in return, such as a decision meeting, regular feedback and permission to measure usage.
Price early. It feels uncomfortable because pricing forces a real response, but that is precisely why it works. If prospects say the offer is valuable but reject every plausible price point, you may have a positioning issue, a weak pain point or the wrong customer segment.
Build the smallest credible test
You do not need the finished product to test demand. You need a believable promise and a way to deliver enough of the outcome to learn whether customers care.
For a service, sell a tightly defined pilot before building a full delivery operation. For software, use a clickable prototype, a landing page with a clear call to action, or a concierge version delivered manually behind the scenes. For a physical product, test the proposition, price and purchasing journey before investing in a large production run.
The test must be honest. Do not imply that a feature already exists when it does not. You can say that you are onboarding a limited group for an early version, explain what is included and set a realistic delivery timeline. Early customers can tolerate rough edges if the value is clear and you communicate professionally.
A manual test also reveals the work your future product must remove. If you repeatedly need to chase missing information, translate messy inputs or make judgement calls for customers, that is useful intelligence. It may expose a product opportunity, but it may also show that the model will be expensive to scale.
Set decision thresholds before you collect data
Validation becomes unreliable when you decide what the data means after seeing it. Set your thresholds in advance.
For instance, you may decide to continue only if you secure five paid pilots from 30 qualified conversations, achieve a conversion rate above 8 per cent from a targeted landing page, or retain at least 60 per cent of pilot customers after the first month. The numbers will differ by business model, price and sales cycle. What matters is that they reflect an economic reality you can build on.
Your thresholds should cover three areas: demand, delivery and economics. Demand asks whether customers respond. Delivery asks whether you can produce the promised result consistently. Economics asks whether the revenue can support customer acquisition, fulfilment and a reasonable margin.
A product can pass one test and fail another. You may find enthusiastic customers but discover that onboarding takes 15 hours per account. That does not always mean stop. It may mean raise the price, narrow the offer, automate a step or target customers with a higher-value use case. Validation is designed to guide a better decision, not force a simplistic yes or no.
Run interviews like an operator, not a pollster
Customer interviews work best when they are part of a commercial process. Recruit people who fit your first segment through your network, direct outreach, industry groups and existing customers. Aim for conversations with enough qualified people to see patterns, then test those patterns through an offer.
During interviews, listen for the language customers use to describe the problem. Those words should shape your sales page, outreach and product positioning. If customers call their problem “wasted reporting time”, do not lead with “data workflow optimisation” unless they already use that language.
Watch for false positives. “I would probably use that” is weak. “We budgeted for this last quarter”, “Can you send me a proposal?” or “If this works, I can introduce you to our operations director” is stronger. Your job is not to win approval. It is to find the boundary between curiosity and commitment.
Measure behaviour after the sale
A first sale is a milestone, not final proof. Customers sometimes buy because they trust the founder, want to be helpful or hope a new tool will solve a problem. Retention and usage reveal whether the value holds up.
Track what customers do after they start. Are they completing the key action? Are they returning without being chased? Are they referring colleagues? Are they asking for expansion, or quietly disengaging? For a subscription offer, renewal intent and active use matter more than a flattering testimonial.
This is where structured executive thinking can save time. A platform such as RPM ReallyDoing.It can help founders pressure-test positioning, shape pilot offers, model basic unit economics and turn customer feedback into a focused execution plan. The aim is not to outsource judgement. It is to make better decisions faster, with fewer blind spots.
Know when to refine, pause or proceed
Validation should give you permission to act decisively. Proceed when you see a repeatable pattern: the same type of customer recognises the problem, responds to a clear offer, accepts a workable price and receives enough value to continue.
Refine when the problem is real but the offer is not landing. You may need a sharper segment, a different promise, a simpler delivery model or a higher-value trigger. Pause when conversations produce politeness but no commitments, when acquisition costs make the model unattractive, or when customers cannot describe a meaningful consequence of doing nothing.
The goal is not certainty. No founder gets that before launch. The goal is a level of evidence that makes your next investment sensible. Build the next step that customers have earned through their actions, and let real demand set the pace.