Skip to content

Industry Perspective

Test demand before launching a new service

Turn early enthusiasm into a clearer decision by investigating a specific buyer, testing a bounded offer and measuring what it takes to deliver.

Illustration of a woman at a market stall holding up a card for the people gathered outside.

A few enthusiastic conversations can make a new service feel ready to launch. The harder question is whether a particular buyer has a problem worth paying you to solve, at a price and delivery effort that can support the business. Testing demand means collecting evidence about that decision before building the full offer.

Research supports a disciplined approach, with limits. A 2024 Strategic Management Journal study covering four randomized trials found that teaching a scientific approach led entrepreneurs to set aside more of the ideas they were testing. In a separate field experiment first published in 2023, early-stage entrepreneurs trained to investigate demand gained more customers and earned more revenue by the end of the program than those trained to focus on resources they already had. Neither establishes a guaranteed launch formula for your service.

Write the decision you need to make

Choose a narrow buyer and problem. “Small businesses need marketing help” is too broad to guide a useful test. A more specific hypothesis might be that independent clinics need help turning an existing patient-information pack into clearer, accessible material before a planned service change. Treat that as an unproven possibility, not a description of established demand.

State the offer you are considering, the proposed price and the main uncertainty. Perhaps buyers value the problem but expect their existing agency to handle it. Perhaps approval sits with someone other than the person who likes the idea. A test should be designed to expose those obstacles.

Set a decision date and a small research budget. Decide what evidence would justify a limited pilot, what would require revision and what would make you stop. These are your planning rules, not statistical confidence thresholds. Writing them in advance makes it harder to reinterpret every response as encouragement.

Investigate the last real occurrence

Speak with people who have recently faced the problem, including some outside your immediate supporters. Ask what happened, how they handled it, who was involved and what the consequences were. Explore the existing workaround and the time or money already spent. Questions about actual experience are more informative than asking whether a hypothetical service sounds useful.

For the clinic example, ask how the last information update was commissioned and approved. If the respondent cannot describe the process, find the person who can. Distinguish the end user, internal advocate and budget holder. Their perspectives may all matter, but they answer different questions.

The NSF’s September 2025 I-Corps solicitation emphasizes firsthand evidence for or against product-market fit and identification of customer segments. That program concerns research-based innovation, rather than ordinary consultancy launches. Its useful principle here is to seek evidence that can challenge the initial idea as well as support it.

Offer the smallest honest paid test

Once the problem is credible, test a bounded version of the service. The clinic offer might be a review of one existing document, with a short findings session and a clearly described output. Agree the scope, price, delivery date and any data-access conditions. Use only material you are authorized to handle.

Make clear what exists and what remains to be developed. Do not imply that a complete service, accreditation or capacity is available when it is not. If you invite reservations or advance payments, ensure the terms, delivery commitments and any refund obligations are appropriate before accepting money.

A paid pilot is stronger evidence of a purchasing decision than a compliment, but it still needs interpretation. A heavily discounted sale to a friend may reveal little about normal pricing. Record who agreed, what they bought, the acquisition effort and why they proceeded or declined.

Measure whether delivery can work

Track preparation, meetings, revisions and follow-up alongside the visible production work. Check whether the client could use the output and whether the same problem is likely to recur. One successful engagement may support another limited test; it does not establish a repeatable market.

Keep a record of refusals and nonresponses too. Separate an irrelevant audience, unclear offer, budget constraint and lack of urgency where you have evidence to do so. Do not invent reasons for people who never replied. A small, convenient sample cannot estimate the size of the wider market.

At the decision date, continue, revise or stop using the evidence you collected. You may discover a narrower buyer, a different purchasing route or an uneconomic delivery burden. Finding that out before investing in a full launch is a useful result. The aim is a better business decision, including permission to set an appealing idea aside.

More to read

Browse insights