How to Conduct Market Research Before Launching Your Business

The week before Webvan launched in 1999, its founder Louis Borders told reporters the company would fundamentally transform how Americans bought groceries. He had raised over $375 million in venture capital — at the time, one of the largest pre-revenue funding rounds in history. He had built enormous automated fulfillment warehouses across multiple metropolitan areas. He had hired thousands of employees and invested in a fleet of refrigerated delivery vans. What he had not done, as subsequent post-mortems would reveal in painful detail, was validate whether American consumers were actually ready — in the numbers necessary to make the unit economics work — to buy their groceries online and pay for home delivery.

Webvan burned through $1.2 billion in capital and filed for bankruptcy in July 2001. Less than two years later, Amazon acquired its warehouse assets for $5.9 million. The idea was not wrong — today, online grocery is a multi-hundred-billion-dollar global market. The timing, the cost structure, and the depth of customer validation were wrong. And wrong by enough to end the company before the market caught up to the vision.

This is an extreme version of a mistake that repeats at every scale, in every industry, on every continent: a business built on what the founder believes the market wants, rather than what the market has actually demonstrated it will do.

What Market Research Is Actually For

There is a persistent misconception about the purpose of market research. Many founders approach it as a process of proving that their idea is good — gathering evidence that validates a decision already made. This is the wrong frame entirely, and it produces research that is worse than useless because it creates false confidence.

Real market research is a structured process of learning, which includes learning that your idea, in its current form, doesn't work — and understanding the specific reasons why in enough detail to make a better decision. The founder who goes into research with genuine curiosity and emerges with the discovery that her target customer won't pay what she assumed is lucky. She learned this before spending her savings and two years of her life building. The founder who treats research as confirmation-gathering and discovers the same truth twelve months and $200,000 later is in a categorically different situation.

Think of market research as systematic risk reduction. Every assumption embedded in your business plan — about who your customer is, what problem they urgently need solved, what they currently use, what they'd pay, how they make buying decisions — is a risk. Good research converts those assumptions into knowledge or, where knowledge isn't yet available, into better-informed hypotheses with explicit uncertainty ranges attached. You build from a foundation rather than from a guess.

Secondary Research: Start Here

Secondary research means studying information that already exists — reports, studies, competitor data, government statistics, industry analyses. It is the fastest and cheapest way to get oriented in a new market before investing time in direct customer contact.

Industry reports and market data. Research firms like Gartner, IBISWorld, Euromonitor, and Statista publish detailed analyses across nearly every industry. These reports are expensive to purchase outright but are often accessible in summary form for free, through university library memberships, or through business support organizations. They give you baseline market size figures, compound annual growth rates, and competitive landscape overviews that provide essential context.

Competitor analysis done seriously. Your competitors are the most underused research resource most founders have access to. Study every competitor you can identify: their websites, pricing pages, case studies, product documentation, customer reviews (particularly the critical ones on G2, Capterra, Trustpilot, or app stores — these are unfiltered customer voice). Sign up for their email newsletters. Try their products. Talk to their customers where possible. Your competitors have already done significant market validation; understanding what they've built and what customers say about it tells you an enormous amount about what works, what doesn't, and where the gaps are.

Government and statistical data. National statistics bureaus publish census data, business formation statistics, industry employment figures, import/export data, and consumer spending breakdowns that are free and often dramatically underused by early-stage founders. In Kenya, the Kenya National Bureau of Statistics publishes comprehensive economic and demographic data. In the US, the Census Bureau and Bureau of Labor Statistics cover virtually every sector. These sources give you population data, income distributions, and industry size figures that underpin credible market sizing.

Online communities where your customers gather. Spend structured time in the forums, Facebook groups, Reddit communities, LinkedIn groups, and WhatsApp communities where your target market congregates. Read what they complain about. Note what questions they ask repeatedly. Observe which solutions they recommend and which they criticize. This is unsolicited, unfiltered customer voice — often more honest and specific than any survey or interview you'll run.

Primary Research: Talking to Real Customers

Secondary research orients you. Primary research — direct conversations with people who represent your target market — is where genuine insight comes from. There is no substitute.

Customer Discovery Interviews

The goal of a customer discovery interview is explicitly not to pitch your idea. Your job is to understand the customer's world: how they currently experience the problem you're trying to solve, what they've tried, what's worked and failed, how much friction the problem creates in their life or business, and what a solution would be worth to them.

The methodology popularized by Steve Blank in The Four Steps to the Epiphany and refined by Rob Fitzpatrick in The Mom Test is built around a fundamental insight: people are unreliable predictors of their own future behavior but quite reliable reporters of their past behavior. Ask about the past, not the hypothetical future.

Bad question: "Would you use an app that tracked your freelance invoices and chased late payments automatically?"

Better questions: "Walk me through how you handled your invoicing last month. What tools did you use? When did you last have a client pay late — what happened, and what did you do about it? How much time did that situation take you? What did you try that didn't work?"

The second approach extracts real behavior, real friction, real workarounds, and real emotional context. You'll learn things about the problem that no survey or Google search would reveal. You'll hear the exact language customers use to describe their frustration — language that should go directly into your marketing copy.

Conduct at minimum 20 interviews before drawing conclusions. One person's experience is an anecdote. Five people describing the same problem in the same language is a pattern. Twenty people consistently surfacing the same frustrations across different backgrounds and contexts is a finding you can build on.

Finding interview subjects is the part most founders find hardest. Approaches that consistently work: LinkedIn outreach filtered by job title and industry, posts in relevant online communities offering a small gift card for 30 minutes of their time, friends and extended-network referrals, and direct outreach at industry events. Always be transparent in your approach: you are researching a problem, not selling something. This framing dramatically increases acceptance rates.

Surveys

Surveys are not a substitute for interviews — they are a complement. Surveys are most useful for quantifying patterns you've already identified through interviews: testing how widely a finding holds, ranking priorities, or establishing benchmarks across a larger population.

Design surveys with discipline. Keep them under ten minutes. Ask specific behavioral questions. Avoid leading questions that signal the answer you're hoping for. Include at least one open-text field where respondents can say something you didn't anticipate. Distribute them to audiences that actually match your target customer — a survey answered by random people is noise dressed up as data.

Google Forms works for simple surveys at no cost. For targeted distribution to specific demographic profiles, platforms like Pollfish or SurveyMonkey Audience let you pay to survey filtered respondents — useful when your network doesn't naturally include your target customer type.

Validating Willingness to Pay

One of the most neglected dimensions of pre-launch market research is pricing validation. Founders routinely confirm that people have the problem they're trying to solve without confirming that those people will pay — at the price point the business model requires — to have it solved.

The most robust test is pre-selling. Before you build the product in full, ask people to pay for it. This feels uncomfortable, but it is the most honest signal available. Kickstarter campaigns do this explicitly and publicly. Many B2B founders close pre-sales — even at a significant discount — before writing a line of code. If ten businesses will sign a letter of intent (or better, pay a deposit) for a product that doesn't yet exist, you have genuine market validation that no amount of positive survey responses can match.

If pre-selling isn't feasible for your context, use the Van Westendorp pricing model in your interviews: ask each subject four questions. At what price would this feel expensive? At what price would it feel too expensive to consider? At what price would you begin to question its quality? At what price would it feel like a bargain? The overlapping answers from 20-30 respondents give you a price range that the market will accept without heavy friction.

Building and Testing a Minimum Viable Experiment

Beyond interviews and surveys, the most compelling market research takes the form of a working experiment that produces real behavioral data.

A freelance consultant testing a new service offering can provide it to two or three clients for a reduced fee before building infrastructure around it. A product startup can create a landing page that describes the product in specific terms, drive paid traffic to it, and measure email signup rates — treating signups as a proxy for genuine interest. A retail concept can be tested at a weekend market before committing to a lease and fit-out. A SaaS product can be validated with a working prototype shown to paying customers before the full product is built.

These experiments are not launches. They are controlled, time-bounded tests designed to generate specific data against specific hypotheses. They cost a fraction of building the real thing and produce data that is far more actionable than any survey.

Synthesizing What You've Learned

After 20-30 interviews, one or two surveys, and perhaps a live experiment, you have a large volume of information. The work of analysis is separating signal from noise.

Affinity mapping is the most practical technique: write each distinct finding or theme on a note (digital tools like Miro work well for remote teams), then group related notes together. The clusters with the most notes pointing to the same insight are your strongest signals. Pay particular attention to findings that emerged unprompted — problems or needs that interviewees raised without being specifically asked, because those reflect the issues that are top of mind rather than issues that feel relevant when prompted.

Look for: problems that multiple customers described in strikingly similar language, the current solutions they use and their specific complaints about each, the moments in their day or workflow where the problem causes the most friction, and the specific dollar amounts or time costs they associate with the problem. These are the raw materials of a compelling value proposition.

Knowing When You've Done Enough

There is no point at which market research is complete. You could conduct interviews indefinitely. The discipline is knowing when you have enough confidence to justify the investment the next stage requires — not certainty, which is never available, but a grounded foundation.

You have done enough preliminary research when you can name your target customer with real specificity, you have spoken with enough of them to hear genuine patterns, you understand what they currently use and why it falls short, you have a credible hypothesis about pricing that is grounded in what actual humans told you they'd pay, and you have at least a handful of people who expressed something closer to enthusiasm than polite interest.

That is not a guarantee. Nothing in early-stage business is. But it is a foundation — and building on a foundation is a profoundly different activity from building on hope.