Market research to start a business: what to check before you build

Market research before starting a business should answer three questions in order: does a specific group of people have a problem they already spend time or money on, can you reach enough of them, and will they pay you instead of their current workaround. Answer them with conversations and small tests first, then use public data to check that the market is large enough. Reports and surveys come later, if at all.

This page is for founders starting something new, with no customers yet. If you already run a business and are weighing a new service or location, market research for small business covers that situation.

A worked example: software for independent veterinary clinics

Suppose you are thinking about building software that helps independent veterinary clinics track vaccine and medication stock. You have heard from a friend who works at a clinic that stock runs out without warning. This is a hypothetical example used throughout the page.

At this point you have an idea and one secondhand story. The goal of the research is to replace that story with evidence you can act on, or to find out quickly that the idea does not hold up.

Write down the assumptions that could kill the idea

List what must be true for the business to work. For the clinic example:

Stockouts happen often enough in independent clinics to cost them money or time.
A particular person in the clinic owns the problem and can choose a tool.
The clinic's existing practice software does not already handle it well enough.
Clinics will pay enough to support a business, and you can reach them affordably.

Mark the one you are least sure about and would be most damaging if wrong. Research that one first. The customer discovery research plan walks through turning assumptions like these into a first round of interviews.

Talk to people who have the problem

For a new business, conversations are usually the fastest way to test the first two assumptions. NSF's I-Corps program, which trains scientists and engineers to judge a technology's commercial potential, builds its curriculum around hands-on customer discovery: talking to potential customers, partners and other industry stakeholders.

Ask about the last time the problem happened, not whether they would like your product:

Tell me about the last time you ran short of a vaccine or medication. How did you notice, who dealt with it, and what did it cost you?

Listen for who handled it. If the practice manager, not the veterinarian, orders stock, that changes who you sell to and who you should interview next. Find customers to interview covers where to look before you have any customers, and how many customer interviews covers when to stop.

Talk to people around the problem

The people with the problem are not the only useful sources. Distributors, people who build existing clinic software and consultants who advise practices see many clinics at once. They can tell you how clinics buy, which tools they already pay for, and why earlier attempts at your idea failed.

Treat what they say as leads to check with clinics directly. A supplier's view of what clinics want is not the same as a clinic telling you. Research a market with customer, supplier and operator interviews explains how to combine these viewpoints without mixing them up.

The directory has pages for veterinary operations professionals and veterinary product professionals if you need a starting list.

Size the market from the bottom up

Once interviews suggest a real problem, check whether enough buyers exist. Count the buyers you could plausibly reach, then multiply by what one would realistically pay. This is more useful than quoting a large industry total from a report.

For a US business, County Business Patterns gives annual counts of establishments with paid employees by industry and county, and the Census Business Builder presents similar data in maps and reports. Find your industry's classification code, pull the establishment count, then narrow it: clinics that are independent, above a certain size, in regions you can serve.

Write the estimate as a chain you can defend: establishments in the industry, the share that fit your first segment, the price your interviews support. Each link should point to a source or a conversation. If one link is a guess, label it as a guess. Choosing your first customer segment helps with the narrowing step.

Check the alternatives and the price

The SBA's market research guide lists questions to answer about your market, including demand, market size, saturation and what customers pay for alternatives. For a startup, the alternative is often a spreadsheet, a habit or a feature inside software the buyer already uses. Find out what that alternative costs them in time and money.

Then test price with something closer to a commitment than an opinion: a paid pilot, a preorder, or a letter of intent. How to research willingness to pay covers the question wording and what counts as evidence.

Decide what you learned

After each round, write down which assumptions are now supported, which were contradicted and which remain open. Pick the next step based on the riskiest open assumption. Stop researching a question once the answer would no longer change what you do next.

If you need to reach people outside your own network for these conversations, Instant Expert can help. Describe who you need, such as "practice managers at independent veterinary clinics who handle ordering", review the people it finds, and send invitations. You pay only for calls that get booked.