A common pattern looks like this: a founder has an AI sales coaching idea.
They ask founder friends, "Would you use this if it helped reps close more deals?" Several say yes.
So they build.
Then they learn the real pain was not "sales coaching." It was narrower: new reps keep making the same objection-handling mistakes for weeks, and managers notice too late. Existing habits, tools, call reviews, CRM notes, and manager feedback already shape the workflow.
That is the difference between encouragement and market research.
Positive intent is not market proof. The market was hiding in current behavior.
TL;DR
Market research is not a report. It is the process of finding out what is already true before you build for the wrong people.
For a startup, good market research answers seven questions:
- Who can we actually reach?
- What painful problem do they already have?
- What do they do now?
- What does that workaround cost in money, time, risk, or frustration?
- Who decides whether to change?
- What alternatives and substitutes already compete for attention, budget, or habit?
- What decision should we make next: continue, narrow, change, or pause?
The mistake is asking, "Would you use this?"
The better question is: "When did this last happen, what did you do, and what did it cost?"
You cannot think market research into existence. You have to find facts outside your head.
What market research means for an early-stage startup
Market research is the process of learning who your customers are, what problems they already have, how they solve those problems today, what alternatives they compare, and whether the pain is strong enough to change behavior. For startups, the goal is not certainty. The goal is better decisions before expensive mistakes.
For a big company, market research can mean agencies, panels, analyst reports, and months of slide decks.
For a founder, it should mean something simpler and sharper: reduce the odds that you build for a market that exists only in your pitch deck.
Startup market research has two parts:
| Research type | What it helps you learn | Useful sources | What it cannot prove alone |
|---|---|---|---|
| Secondary research | What already exists in the market | Competitor sites, review pages, pricing pages, forums, search results, app stores, job posts, customer stories | Whether your reachable customers feel enough pain to act |
| Primary research | What real people have done, bought, tried, ignored, or hacked together | Customer interviews, sales calls, surveys, message tests, landing page tests | Total market size or broad statistical certainty |
Secondary research stops you from asking obvious questions. Primary research stops you from believing your own story. You need both.
How to do market research as a founder
Use this practical sequence:
- Define one reachable market.
- Map the current landscape before interviews.
- List assumptions about pain, workaround, buyer, budget, urgency, and channel.
- Run interviews about past behavior.
- Use surveys to check patterns, not to discover the truth from scratch.
- Synthesize findings into a decision.
- Repeat by segment, geography, channel, or use case.
Do not treat this as a one-time validation phase. Founders come back to market research whenever they change segment, price, positioning, product scope, or acquisition channel.
The startup market research checklist
This is the core framework. Keep it close, but keep the notes short enough that they still force a decision.
- Who can we actually reach? Look for communities, lists, search demand, your existing network, paid groups, or partnerships. You are testing whether this is a real go-to-market segment, not just a persona.
- What painful problem do they already have? Look for recent examples, repeated complaints, and visible frustration. You are testing whether the problem exists before your pitch.
- What do they do now? Look for tools, spreadsheets, agencies, internal processes, and manual workarounds. You are testing whether the problem already creates behavior.
- What does it cost? Look for money spent, hours lost, risk, errors, missed revenue, or embarrassment. You are testing whether the pain is big enough to matter.
- Who decides? Look for the last purchase approval, job title, procurement path, or founder involvement. You are testing whether you understand the buyer, user, and budget owner.
- What else competes? Look for products, services, consultants, templates, internal workarounds, and doing nothing. You are testing whether you understand direct competitors and substitutes.
- What decision should we make? Choose continue, narrow, change, or pause. If your research does not change a decision, it may not be research. It may be reassurance.
Step 1: Define one reachable market
A market is not "SMBs," "mothers," "sales teams," or "Europe."
A market is a reachable group with similar pain, buying behavior, alternatives, and channels.
That word "reachable" matters. The best theoretical customer is useless if you cannot get in front of them.
| Weak definition | Why it fails |
|---|---|
| "Small businesses" | A 12-person law firm and a 20-person Shopify brand buy differently. |
| "Mothers" | First-time mothers in paid postpartum communities behave differently from mothers of teenagers in a local school group. |
| "Europe" | The UK, Germany, France, Spain, and the US can differ in language, regulation, tools, buyer expectations, and channels. |
| "Sales teams" | A five-person founder-led sales motion is not the same as a 300-person SDR org. |
Better definitions:
| Better definition | Why it works |
|---|---|
| "Seed-stage B2B SaaS teams hiring their first two account executives" | Shared timing, pain, buyer, and reachable channels. |
| "Independent Shopify brands with monthly overstock issues and no dedicated inventory planner" | Clear trigger, workaround, and likely economic pain. |
| "First-time mothers in paid local postpartum groups who already buy planning or support services" | Reachable and behavior-based. |
| "Sales managers onboarding new reps in founder-led B2B sales teams" | Specific pain and identifiable owner. |
The question is not "What hurts the category?" The question is "What hurts the people we can actually reach?"
Step 2: Map the current landscape before talking to customers
Founders often rush into interviews because they want fresh signal. That is understandable, but it leads to lazy questions.
Before you interview, spend a few hours mapping what is already true:
- Direct competitors: products that claim to solve the same problem.
- Substitutes: agencies, freelancers, spreadsheets, templates, consultants, internal staff, communities, or doing nothing.
- Pricing: public plans, sales-led pricing clues, usage limits, and annual contracts.
- Customer language: review complaints, forum posts, comparison pages, Reddit threads, and LinkedIn comments.
- Buying triggers: compliance deadlines, hiring, churn, missed revenue, regulation, team growth, or a failed manual process.
- Channels: where customers spend time, search, complain, buy, ask peers, or hire help.
- Proof of urgency: job posts, budget ownership, public complaints, expensive workarounds, and repeated manual tasks.
A founder building compliance automation for small fintechs should not only list compliance software. They should also look at lawyers, spreadsheets, internal checklists, ops people, audit consultants, and "we will deal with it later." Named tools can help you find the category, but they should not become the whole competitor map.
No competitors usually means no direct competitors. The real competition may be a spreadsheet, an agency, an intern, a template, an internal process, apathy, or doing nothing.
A useful shortcut: a social app may compete with other social apps, but it can also compete with streaming, games, or anything else that consumes attention. Your startup has a similar problem. The competitor set is wider than the product category.
Step 3: Write down your riskiest assumptions
Do this before interviews so you do not bend every answer into support for your idea.
Use this one-page startup market research plan:
- Segment problem: What evidence would show this segment has the problem often enough to care? Check interviews, reviews, and forums. Decision: continue, narrow, change, or pause.
- Current workaround: What evidence would show the workaround is painful? Check time spent, money spent, errors, risk, embarrassment, or delay.
- Budget owner: What evidence would show the buyer owns budget? Ask who approved the last workaround or tool, and compare that with job titles and pricing clues.
- Urgency: What evidence would show the pain matters now? Look for deadlines, trigger events, missed revenue, compliance risk, churn, or a new hire.
- Reachability: What evidence would show you can reach this segment? Check search demand, LinkedIn, communities, partnerships, outbound lists, and your existing network.
- Alternative weakness: What evidence would show existing alternatives leave room? Look for complaints, switching stories, manual patches, and unmet needs.
- Customer language: What evidence would show you can sell with their words? Capture repeated phrases, emotional words, and concrete moments from interviews, reviews, and sales calls.
The output is not a beautiful research doc. Notes are only useful if they change the segment, product, message, price, channel, or next call.
Step 4: Run interviews about past behavior
The main rule: ask what happened last time, not what they would do next time.
Weak interview questions invite politeness. Strong questions make the person describe reality.
- Instead of "Would you use a tool that fixes this?" ask "When did this last happen?"
- Instead of "Would your team pay for this?" ask "What did you pay for the last workaround?"
- Instead of "Is this a big problem?" ask "What happened the last time this caused trouble?"
- Instead of "Do you dislike your current tool?" ask "What do you still use it for, even though it annoys you?"
- Instead of "Would automation help?" ask "What parts are still manual today?"
- Instead of "Would you switch?" ask "When was the last time you switched tools like this?"
- Instead of "Who is the buyer?" ask "Who approved the last purchase or workaround?"
- Instead of "How painful is this?" ask "What did it cost in time, money, risk, or reputation?"
A raw note like "they hate spreadsheets" is incomplete. Spreadsheets may still be good enough.
A better signal sounds like this: "They spend six hours every Friday cleaning the sheet, the CFO checks it manually, and invoicing was delayed twice because of errors." Treat that as an illustrative example, not a benchmark.
For a fuller interview bank, use these customer research questions to stay focused on real behavior instead of hypothetical interest.
Step 5: Look for economic pain and emotional pain
Founders often look only for money pain: wasted budget, lost revenue, high labor cost.
That matters, but it is not the whole picture.
People also act because a problem makes them feel exposed, embarrassed, slow, guilty, irresponsible, or out of control.
An ecommerce founder with leftover inventory does not only have cash tied up in stock. They may also feel like they made a bad call, over-ordered, disappointed the team, and now have stale inventory staring at them every week.
That emotional weight can create urgency. It can also shape messaging.
- Money: "We paid a contractor $4,000 to clean this up."
- Time: "My ops lead loses every Friday afternoon to this."
- Risk: "If we miss this deadline, the audit gets ugly."
- Status: "I look unprepared in the leadership meeting."
- Frustration: "Everyone knows this is broken, but nobody owns it."
- Fear: "I am worried we will make the same mistake again."
- Guilt: "I approved the process, so now the cleanup is on me."
This is where voice-of-customer work becomes useful. If several customers describe the same pain in plain language, do not translate it into abstract product language too quickly. Use their words. These voice-of-the-customer interview questions can help you capture phrasing that later becomes positioning, landing page copy, and sales messaging.
Step 6: Use surveys carefully
Surveys are usually better for pattern-checking than early discovery.
A survey can help answer:
| Good survey use | Example |
|---|---|
| Frequency | "How often did this happen in the last 30 days?" |
| Segmentation | "Which role owns this today?" |
| Current behavior | "Which workaround have you used in the past six months?" |
| Prior spend | "What did your team spend on this last year?" |
| Ranking | "Which of these problems caused the most operational pain last month?" |
A survey is much weaker when it asks:
| Weak survey use | Why it misleads |
|---|---|
| "Would you buy this?" | People are generous with imaginary money. |
| "Do you like this idea?" | Liking is not buying. |
| "How much would you pay?" | Answers are often detached from real budget behavior. |
| "Is this important?" | Almost everything sounds important in isolation. |
If you want survey prompts that avoid the worst traps, use these market validation survey questions. Treat survey interest as a clue, not purchase proof.
Step 7: Turn research into a decision
Research should end in a decision. Otherwise, it becomes founder theater.
- Pain is real, frequent, expensive, and reachable: continue.
- Pain is real but only in a narrower group: narrow.
- Pain is real but the buyer, urgency, or channel is different: change positioning, buyer, channel, or wedge.
- People complain but do not act, spend, or change behavior: pause or find a sharper trigger.
- Competitors are strong, but customers still patch around them: find the unmet job.
- Channel test failed early, but customer pain is strong: learn the channel or test another route before killing the idea.
- Nobody has recent examples, workarounds, or spend: stop treating this as validated.
A failed early channel test does not automatically mean there is no market. It may mean the audience, offer, creative, trust level, timing, or funnel was wrong. Channels take time to learn.
But do not use that as an excuse to ignore weak evidence. If nobody has done anything about the problem before, your burden of proof goes up.
A practical market landscape matrix
Use this when you are comparing segments or geographies. Keep each row short enough that you can actually make a decision from it.
- UK seed-stage SaaS sales teams: alternatives might include call reviews, founder coaching, CRM notes, or sales enablement software. If managers mention repeated objection mistakes, the next decision may be continue.
- Germany mid-market SaaS sales teams: alternatives might include sales enablement tools, internal training, and localized vendors. Do not merge this with UK research without separate evidence.
- US ecommerce brands with overstock: alternatives might include spreadsheets, agencies, discounting, and warehouse staff. If emotional and financial pain both show up, the next decision may be continue.
- Consumer habit app for new mothers: alternatives might include calendars, notes, partner reminders, and chat groups. If the group is still too broad, narrow before building.
This is where founders fool themselves. "We validated Europe" can mean a few calls in one country, a few replies from another, and a competitor scan somewhere else.
That is not one market. It is a set of half-signals from different markets. Validate one market at a time.
Competitor research should include substitutes
A competitor map should have four columns.
| Type | Examples | What to learn |
|---|---|---|
| Direct competitors | Similar SaaS tools, apps, platforms | Positioning, pricing, feature promises, target customer |
| Substitutes | Agencies, consultants, freelancers, communities, training, entertainment, other priorities | What budget or attention you must displace |
| Manual workarounds | Spreadsheets, docs, chat threads, interns, internal staff | What pain people tolerate and why |
| Do nothing | Delay, ignore, accept errors, keep current process | Whether the problem is urgent enough |
The "do nothing" option is especially important. Assume do-nothing is a serious competitor until customer behavior proves otherwise.
Your job is to learn whether the mess is annoying or intolerable.
What strong validation looks like
Strong market research usually contains evidence like this:
| Weak signal | Stronger signal |
|---|---|
| "People said they would try it." | "Several people described the same recent problem, and some already pay for a workaround." |
| "They liked the demo." | "They asked for a pilot with their real data and named the person who approves it." |
| "They hate their current process." | "They spend hours on it each week and missed a deadline last month." |
| "No one else is doing this." | "Direct competitors are weak, but substitutes are common and painful." |
| "The first ad test did not work." | "One channel test failed before we knew the segment, promise, creative, or landing page." |
| "Moms need better planning tools." | "Working mothers in a reachable paid group already buy planning help, but still miss school deadlines." |
A good discovery conversation often ends with a concrete next step when the person has real urgency: an introduction, a second call with the buyer, a workflow walkthrough, a pilot discussion, a request for pricing, or permission to follow up after a trigger date.
Vague "interesting, let's keep in touch" is not useless, but it is not strong evidence.
Common ways founders fool themselves
Founders usually do not skip research because they are lazy. They skip the painful parts because fake confidence feels better.
- Asking friends: you are measuring politeness and relationship warmth.
- Asking "Would you use it?": you are collecting imaginary future behavior.
- Treating a broad category as a market: you are hiding differences in pain, buying, and reachability.
- Looking only at direct competitors: you are missing substitutes and do-nothing behavior.
- Overbuilding the research doc: you are polishing notes instead of making decisions.
- Choosing tools for days: you are avoiding the harder work of finding facts.
- Treating one channel failure as market failure: you may only have learned that your first channel attempt was weak.
- Treating no competitors as good news: you may have found a problem people do not care enough to solve.
The antidote is simple but uncomfortable: look for behavior.
What did they do last time? What did they pay? Who approved it? What broke? What happens if they ignore it? Where do they look for help? Why now?
Useful external references
For startup context, Y Combinator's startup advice is a useful companion for founders thinking about whether they are building something people want.
For a broader business research view, the U.S. Small Business Administration guide to market research and competitive analysis covers market research and competitive analysis categories useful for business planning. Startup founders should keep it lighter than a formal planning exercise, but the categories are still useful.
Market research is a repeatable system
The first version of market research helps you decide whether an idea deserves more work.
Later versions help you decide:
- Which segment to pursue first
- Which geography needs separate validation
- Which competitor to position against
- Which pain to lead with
- Which channel to test next
- Which buyer to sell to
- Which feature to cut
- Which pricing story makes sense
- Which customers are not worth chasing
That is why market research should not feel like homework. It should feel like steering.
When the product changes, research again. When the segment changes, research again. When the channel changes, research again. When the price changes, research again.
The goal is not to become a research expert. The goal is to stop guessing about your market.
FAQ
- What is market research for startups?
Market research for startups is the process of finding evidence about a specific reachable market: customer pain, current alternatives, buying behavior, urgency, budget, channels, and competition.
- How do startups do market research?
Start with one reachable segment. Map current alternatives and substitutes. Write down assumptions. Interview people about past behavior. Use surveys to check patterns. Then decide whether to continue, narrow, change, or pause.
- What is the market research process?
The market research process is: define the market, study the landscape, list assumptions, interview customers, check patterns with surveys, synthesize evidence, and make a decision.
- Can surveys validate a startup idea?
Surveys can support validation, but they should not carry it alone. They are better for checking frequency, segmentation, current behavior, and ranking than for proving people will buy.
- How much market research is enough before building?
Enough to name a reachable segment, describe the current workaround, identify the buyer, explain the urgency, understand alternatives, and make a clear next decision.