How to Do Market Research for a Startup (Step-by-Step)

Every founder believes their market is big enough. Far fewer can prove it. The gap between a confident pitch and a fundable business is almost always market research — the unglamorous work of replacing "I'm sure people want this" with numbers, quotes, and evidence. Done well, it tells you whether your idea is worth building, who exactly will pay for it, and how to reach them before you burn a year of runway. Here's how to do it properly, step by step.

What is market research?

Market research is the process of gathering and analyzing information about your customers, competitors, and industry to make better business decisions. For a startup, it answers three questions: is there real demand for what you want to build, who exactly will pay for it, and how big could this opportunity become? It combines primary research (data you collect yourself, like interviews) with secondary research (existing data, like industry reports) to replace assumptions with evidence.

Good market research is not a one-time document you write to satisfy an investor. It's an ongoing discipline that sharpens your strategy, pricing, and positioning as you learn.

Primary vs. secondary research

The two halves feed each other, and you need both.

Primary research is data you gather first-hand: customer interviews, surveys, observing how people behave, or running small experiments. It's specific to your business and impossible for a competitor to copy because you generated it. Primary research is where the real insight lives — the unexpected quote, the workaround a customer hacked together, the price someone blurted out before thinking.

Secondary research is existing data collected by others: industry reports, government statistics, competitor financials, and credible online sources. It's faster and cheaper than primary research and gives you the big picture — market size, growth rates, and benchmarks. The smart sequence is to start with secondary research to understand the landscape, then use primary research to test what a report can't tell you: whether your specific customers will actually pay.

Defining your target market

You cannot research a market you haven't defined. "Everyone" is not a target market; it's a fantasy that makes research meaningless. Start narrow: who has the problem most acutely, who has money to spend, and who is easiest to reach?

Build a sharp customer profile. For a B2B product, that means industry, company size, the buyer's role, and their budget authority. For a consumer product, it means demographics, behaviors, and the context in which the problem bites. A useful test: if you can name three real people who fit your profile and could call them today, your target market is concrete enough. Defining your customer tightly makes every later step — sizing, interviews, competitor analysis — far more accurate.

Market sizing with TAM, SAM, and SOM

Once you know who you serve, you need to know how big the opportunity is. The standard framework is TAM, SAM, and SOM — three nested circles moving from the entire universe down to what you can realistically capture.

• TAM (Total Addressable Market): the total demand if every possible customer on earth bought your product — the global revenue opportunity for your category.
• SAM (Serviceable Addressable Market): the portion of TAM you can actually serve given your business model, geography, and product. If you only sell in Europe or only to small businesses, your SAM is that slice.
• SOM (Serviceable Obtainable Market): the realistic share of SAM you can win in the next few years given competition and resources. This is the number that matters most for a startup — the revenue you can credibly chase.

Calculate from the bottom up whenever you can. Top-down ("the global market is $50B and we'll take 1%") is lazy and investors distrust it. Bottom-up ("there are 40,000 target companies, we can reach 5,000, and 10% convert at $1,200/year") is grounded and defensible. Stripe is a useful illustration: rather than claiming a slice of "all payments," it sized the specific, growing population of online businesses that needed developer-friendly payment infrastructure — a concrete, reachable market that expanded as e-commerce grew.

Customer interviews and surveys

This is the heart of primary research. Interviews tell you why; surveys tell you how many.

In interviews, the golden rule is to ask about the past, not the future. "Would you use a tool that does X?" invites polite, useless answers. Instead ask: "Tell me about the last time you faced this problem. What did you do? How much did it cost you?" Past behavior is data; future intentions are wishful thinking. Aim for 15–30 conversations with people who fit your target profile, and listen for emotion and effort — complaints, money spent on workarounds, and homemade hacks are the loudest signals of real pain.

Surveys come after interviews, not before. Once interviews reveal the patterns, surveys quantify them across a larger group: how common is the problem, what do people currently pay, which feature matters most? Keep surveys short, avoid leading questions, and reach respondents where your real customers already gather. If you're not sure which assumptions to test first, a quick AI diagnostic can map your idea and surface the riskiest questions to put to customers.

Competitor analysis

If you think you have no competitors, you haven't looked hard enough — or there's no market. Every problem worth solving already has a status quo: a rival product, a spreadsheet, or simply people doing nothing. Map them all.

For each competitor, document what they offer, how they price it, who they target, and their weaknesses. Read their reviews obsessively — one- and three-star reviews are a goldmine of unmet needs you can build around. Notion grew by studying a crowded productivity landscape and noticing users were stitching together separate tools for notes, tasks, and docs; it won by collapsing those into one flexible workspace. The lesson isn't to copy competitors — it's to find the gap they've left open and the customers they've underserved.

Also distinguish direct competitors (solving the same problem the same way) from indirect ones (a spreadsheet, a manual process, or doing nothing). Underestimating indirect competition — especially "good enough" free alternatives — is how startups misjudge how hard switching will be.

Analyzing demand and trends

A great product in a shrinking market is a hard life. Study where the market is heading. Is demand growing, flat, or declining? What technological or behavioral shifts are reshaping the category? Search-trend data, industry growth reports, and public funding activity all reveal momentum.

You want a market with a tailwind — growing demand, an emerging behavior, or a recent shift that makes your timing right. Airbnb rode the rise of the sharing economy and travelers' growing comfort with peer-to-peer trust; the trend made a once-strange idea feel inevitable. Dropbox caught the early wave of people working across multiple devices and needing their files everywhere. Timing isn't luck — it's something you can research. Answer honestly: why is now the right moment, and what trend is on your side?

A step-by-step market research process

1. Define your target customer. Write a specific profile — industry, role, demographics, behaviors. If you can't name three real people who fit, refine it.
2. Do secondary research first. Pull industry reports, market-size data, and trend signals to understand the landscape before you talk to anyone.
3. Size the market with TAM, SAM, SOM. Calculate bottom-up. Your SOM — the realistic, obtainable slice — is the number that should drive your plan.
4. Run customer interviews. Hold 15–30 conversations with real target customers about their past behavior and genuine pain. Learn, don't pitch.
5. Quantify with a survey. Measure how common the problem is, what people pay, and which needs rank highest.
6. Analyze competitors. Map direct and indirect rivals, their pricing and positioning, and the gaps in their reviews.
7. Study demand and trends. Confirm the market is growing and that timing is on your side.
8. Synthesize and decide. Turn findings into a clear view of the opportunity — then proceed, adjust, or rethink.

Common mistakes to avoid

• Defining "everyone" as your market. A vague target makes every number meaningless. Start with a sharp, specific customer profile.
• Top-down market sizing. "It's a $50B market and we'll grab 1%" signals laziness. Build your numbers bottom-up from real, reachable customers.
• Only doing secondary research. Reports tell you the landscape, not whether your customers will pay. They can't replace conversations.
• Asking leading questions. "Wouldn't this be amazing?" guarantees flattering answers. Ask about what people actually did in the past.
• Ignoring indirect competitors. Spreadsheets, manual processes, and "good enough" free tools are real competition — underestimating switching costs is fatal.
• Treating research as one-and-done. Markets move. Revisit your research as you learn, not once to satisfy an investor.

FAQ

How much does market research cost for a startup? It can cost almost nothing but time. Secondary research draws on free public reports, government statistics, and competitor websites. Primary research — interviews and surveys — costs mainly the hours you spend talking to people and, optionally, a small budget for survey tools or modest ads to reach respondents. Expensive research firms exist, but early-stage founders rarely need them. The most valuable insights usually come from conversations you can run yourself for free. The real investment is effort and honesty, not money.

What's the difference between TAM, SAM, and SOM? They're three nested measures of market size. TAM (Total Addressable Market) is the entire global demand for your category if everyone bought. SAM (Serviceable Addressable Market) is the portion you can realistically serve given your geography, model, and product. SOM (Serviceable Obtainable Market) is the slice of SAM you can actually capture in the next few years given competition and resources. For a startup, SOM matters most because it reflects achievable revenue, and investors trust bottom-up calculations far more than top-down guesses.

How many people should I interview for market research? A practical baseline is 15–30 in-depth conversations with people who genuinely fit your target customer profile. By that point clear patterns emerge — either the same painful problem keeps surfacing, or you learn that few people truly care. Quality beats quantity: ten interviews with exactly the right customers beat fifty with random people. Focus on what interviewees actually did in the past rather than what they imagine they'd do, and follow interviews with a broader survey when you want to quantify the patterns you've found.

Is market research the same as validation? They overlap but aren't identical. Market research is the broad process of understanding your customers, competitors, market size, and trends — the full landscape around your idea. Validation is the narrower act of gathering behavioral proof (sign-ups, pre-orders, paying customers) that a specific solution is wanted. Research tells you whether the opportunity is real and how to position for it; validation tests whether your product will actually sell. You research the market first, then run validation experiments inside it.

In summary

Market research is how you replace confidence with evidence. Define a sharp target customer, use secondary research to understand the landscape, then size the opportunity bottom-up with TAM, SAM, and SOM — paying closest attention to the SOM you can realistically obtain. Interview 15–30 real customers about their past behavior, quantify what you learn with a survey, and map both direct and indirect competitors to find the gap they've left open. Confirm the market is growing and that timing is on your side, then synthesize everything into a clear view of the opportunity. The companies that scaled — Stripe, Notion, Airbnb, Dropbox — all understood their customers and timing deeply before they bet everything on the build.

Ready to turn research into a real business? Run a free AI diagnostic to map your market and find the riskiest questions to test, explore our plans for end-to-end startup support, or get matched with the co-founder you'll need to chase the opportunity you've uncovered.