You can start testing a business idea without buying ads, building a website or ordering stock. Use the phone, internet access and contacts you already have to find people who recently dealt with the problem. Find out what they tried, what they paid and what still bothers them. Then show a specific offer with a price. Conversations can sharpen that offer; they cannot establish that anyone will buy it.
Start with the free seven-day validation planner. Set the maximum affordable test loss to ₹0 if you have no money available for the experiment. Its numbers are editable planning choices. The week gives you a review date, not a deadline by which a business becomes validated. You can download the plan without signing up.
Begin with a reachable buyer and a recent problem. Use conversations and an honest priced offer to decide what deserves a further test. Record follow-up requests, paid commitments, deliveries and refunds separately. Spend only if the next question requires it and the loss fits your budget. A small successful pilot supports another small decision; it does not establish a profitable business.
Start with buyers you can actually reach
Write down who has the problem in a form that lets you find them. “People who want a tidy home” is difficult to act on. “People who work at a dining table and clear it before dinner” gives you a situation to ask about. If you are still choosing between products, use the product-idea guide to narrow the buyer and use case first.
Ask existing contacts for introductions to people in that situation. In a relevant community, check its rules and ask permission for research before posting. An introduction or a short, relevant invitation is enough; a scraped list and repeated unsolicited messages are not needed. Explain that you are learning about a problem, keep the conversation brief and let people decline.
Record how you found each person. Friends who want to encourage you, buyers introduced by a friend and people who found the offer independently are different groups. Keep them separate when reading the results. If you cannot reach the buyer, your first unresolved question is distribution. A larger market estimate will not solve it.
Ask about the last purchase before presenting your idea
Rob Fitzpatrick's The Mom Test is a useful reference for avoiding biased customer feedback. Start with a recent event the person can describe. For a snack idea, ask what they ate on their last evening commute, where they bought it, what it cost and what they disliked. Leave your proposed product out of that first account.
Follow a vague answer with a request for an example. If someone says they always buy healthy food, ask what they bought most recently. If they describe an annoying problem, ask what they did about it. A problem can be irritating without being worth paying to fix. Record their actual workaround, including doing nothing.
Useful notes include the occasion, the current alternative, the amount paid, who chose it and why the buyer would consider switching. Keep your interpretation in a separate column from what the person actually said. Ask permission before recording a call or following up. Several similar accounts can help you choose an offer to test, but an interview count is not a purchase forecast.
Show one priced offer without pretending it is available
You can describe an offer in a message, a document or a sketch made with tools you already use. A landing page is useful when you need one consistent explanation, but it is not a prerequisite. State the buyer's situation, what you propose to provide, the price and what exists today. If you only have a concept, say so beside the offer.
For a concept, ask whether the person wants a follow-up when a sample is ready. For something you can already deliver, offer a small paid pilot with a clear scope. Keep those actions distinct. Do not label a button “Buy now” if it merely collects an email address, or collect payment details to measure curiosity.
Buffer's founder describes this distinction in his account of its early tests. He first collected interest through a simple page, then added pricing choices. Those clicks informed his decision to build. His first paying customer came after the working product launched. The pricing-page response and the subsequent purchase were separate pieces of evidence.
Keep the price and offer stable for a defined group so you can interpret objections. If you change the bundle or discount halfway through, start a new row in your log. The pricing guide can help you build a cost estimate, but you still need to check whether buyers accept that price.
Keep the denominator beside every result
Here is a fictional example to show how to keep the record. You propose a fold-flat desk organiser at ₹699 for people who clear their work off the dining table each evening. You contact 20 suitable people, speak with eight, and show the same priced sketch to six. Four ask you to tell them when a sample is ready. Nobody pays.
Your record is 20 contacts, eight conversations, six offers shown, four follow-up requests and zero paid commitments. Four out of six people asked for an update. That does not mean four customers, or a two-thirds purchase conversion rate. Note which people knew you beforehand and what they expected the sample to do. Those details matter more than turning a small count into an impressive percentage.
You now have a narrower question: will the organiser fit their space and routine at the proposed price? A later sample might answer it. If your loss limit is still ₹0 and making one requires money you do not have, keep refining the specification with willing participants or pause. The interest remains useful without being relabelled as paid demand. These example counts are not recommended targets.
| Observed action | What you have learned | What remains unknown |
|---|---|---|
| Describes a recent purchase or workaround | The problem mattered in that situation | Whether your offer is a better choice |
| Clicks a price or requests an update | The offer attracted attention | Whether the person will pay |
| Confirms a COD order | The person stated an intention to accept it | Whether delivery and payment will happen |
| Pays for a clearly described pilot | A buyer committed money on those terms | Fulfilment, refunds, satisfaction and repeatable acquisition |
| Keeps the delivered product after the return period | You completed a retained purchase | Whether more buyers will be profitable to serve |
Run a paid pilot only when you can meet the promise
A small pilot tests more than willingness to click. For a physical product, use an existing sample or an honestly described small batch that you can supply. Give the buyer the price, what is included, a realistic delivery date and clear cancellation or refund terms before taking money. Keep enough cash available to meet those obligations. If you cannot fulfil or refund the order, use a non-payment interest test until you can.
A service or software idea may allow a manual first version using skills and tools you already have. Paul Graham's essay on doing things that do not scale describes founders recruiting early users directly and doing work manually before automating it. Tell the customer what you will do, how much is manual and what result you are committing to deliver. A paid manual service does not yet prove that a future automated product will work.
For a D2C pilot, count confirmed COD orders separately from delivered and paid orders. Track prepaid refunds too. The RTO guide explains why a refusal changes the costs of an order. Use the first-customer guide for outreach ideas, with a target you choose for your test. No particular number of orders removes the need to examine delivery costs, buyer relationships and the conditions under which people paid.
Use a week to make a smaller decision
A short schedule helps stop research drifting into endless preparation. On the first day, name the buyer, the current alternative, one offer, a proposed price and the maximum loss you can absorb. On the next two days, arrange conversations and reconstruct recent purchases. Use the fourth day to put the offer in front of suitable buyers, changing any wording they cannot understand.
Over the next two days, present that version consistently and record the response. Ask for a paid commitment only if you can deliver the offer within your constraints. Otherwise record the interest and the missing capability. At the end of the week, review what actually happened. Buyers may need more time, particularly when another person approves a purchase. A calendar deadline does not turn a pending decision into a rejection.
Validation Sprint™: set aside a short period to examine one assumption and decide on the next limited commitment. Write the buyer, offer, target and affordable loss before starting. A completed plan or a met target does not certify demand. Review who paid, what you owe them and what the result leaves unanswered.
If people describe the problem but decline the offer, ask what stopped them: urgency, trust, price, the proposed product or the buying process. If nobody saw the offer, improve access to the buyer before drawing a conclusion about the product. If some people pay, fulfil those commitments and review the costs before increasing the batch.
When a paid traffic test is worth considering
Ads are an optional way to test access to buyers beyond your immediate network. Consider them when you have a clear offer, can measure the action you care about and have a budget you can afford to lose. A click campaign answers a different question from a purchase test. The Meta ads guide covers the mechanics if you reach that stage.
Choose a spending ceiling and a review date before launching. For example, a founder might choose ₹2,000 as an affordable limit for one question. That is an illustration of a limit, not a claim that ₹2,000 buys enough evidence. Traffic may be expensive or the response too small to interpret. Stop at the limit and record the uncertainty; raising it simply to reach a preferred result changes the experiment.
A zero-spend first test and a later paid test have different costs and purposes. There is no single ₹3,000 or ₹15,000 budget that validates every business. Samples, delivery, tools and refund obligations can cost more than the ads. Use the launch-cost guide when estimating the next stage, and list actual quotes rather than assuming the first week's cost applies to a launch.
Read paid results alongside unit economics. Separate offers shown, paid orders, deliveries, refunds and retained orders. Include product cost, packaging, shipping, applicable payment charges and acquisition spend. A positive margin before advertising does not show what remains after acquiring the buyer. A first sale also leaves repeat demand and future customer acquisition untested.
In the fictional organiser example, treating four follow-up requests as four buyers would be the error. Neither the price nor the product has survived a purchase decision. Ordering a batch at that point adds a stock commitment to an unanswered question. Keep the four people in the interest column and ask what they need to see before they could decide.
- Name a buyer you can reach with permission.
- Record a recent purchase or workaround before pitching.
- Show the same clear offer and price to a defined group.
- Separate follow-up requests, paid commitments, deliveries and refunds.
- Set an affordable loss limit, including ₹0 when necessary.
- Record what you still cannot conclude before approving the next expense.
Put the next expense beside the evidence meant to justify it. A supplier deposit needs a different answer from a decision to arrange another conversation. A useful first test makes the next decision smaller and clearer, even when the answer is to hold the money.
Next action: write the test before spending
Open the free validation planner and write one buyer, one problem and one offer. Use ₹0 as the loss limit if that is your constraint. Choose reachable prospects and a paid-commitment target as planning assumptions, then record unpaid interest separately. If you cannot offer something fulfilable yet, the target remains unmet while you investigate what is missing. Review the plan in a week without treating its completion as a verdict on the business.
For an Indian consumer-product business, the complete execution system continues into suppliers, costs, store launch and operations inside D2C Acquisition.Lab. Your immediate step is to build the free test plan and find the first person whose recent buying decision you need to understand.
