You need a haircut. I'm serious — the "did you fall asleep on a calculator?" look. Before a serious investor will talk to you about your market size, they want to see the whole picture. That's the startup trade-off in one sentence: you spend 80% of your time on 20% that matters, including the awkward parts.
I know the counter-argument. "We can't afford research, we're pre-revenue." So let me describe what actually happens when you skip it. A team I worked with — solid product, real design taste — burned eleven months and roughly €40,000 building a scheduling tool for independent dentists. They'd done zero discovery. When they finally called twenty dentists, they learned most practices already used their bookkeeping software for appointments, and the first question wasn't about UI — it was "who's liable for a double-booking?" That single question would have reshaped the product in week two. They found it in year one.
How to conduct market research for startups without a research team
Here's the honest version: most market research advice is written for people with a budget and a research department. You have neither. So the useful question isn't "how do I do it properly?" — it's "what is the cheapest test that could kill my idea this week?"
Key takeaways
- Prioritize a baseline of 10–15 conversations over any report you could buy.
- Secondary research (public data, competitor pricing, review sites) is free and should take you 2–3 days maximum.
- Define your go/no-go criteria before you collect data, not after.
- AI tools are excellent at synthesis and terrible at original insight — do not outsource the interviews.
- Your sample does not need to be statistically clean. It needs to be specific enough to be falsifiable.
Primary vs secondary research: the practical split
Secondary research is everything already published: industry figures, competitor pricing pages, company filings, review sites, Reddit threads, job postings (a competitor hiring three sales reps tells you something). Primary research is you talking to people. For a startup under five people, the ratio should be roughly two days of secondary for every two weeks of primary conversation.
Why so lopsided? Because secondary research tells you whether a market exists. It almost never tells you whether people will pay. Only a human on the other end of a call does that.
A realistic two-week plan
- Days 1–2: Write your riskiest assumption as a sentence. Example: "Independent pharmacists will pay €150/month to automate reordering."
- Days 3–4: Pull secondary data — how many pharmacies in your country, what their margins look like, what the incumbent software charges.
- Days 5–9: Ten to fifteen conversations. Not surveys. Calls.
- Days 10–11: Ask for a commitment — a pilot, a deposit, a signed letter of intent.
- Day 12: Kill or continue. Write one page on what changed in your thinking.
Budget: under €300 if you're paying for interview incentives or a small ad test. That's it.
The interview questions that actually produce signal
Ask about the past, not the future. "Would you use this?" is a polite lie generator. I've had people tell me enthusiastically they'd pay for a product and then ghost me for six weeks. Every founder I know has that story.
- Walk me through the last time this problem cost you money.
- What did you do instead, and what did that cost?
- Who else is involved in this decision?
- What's the worst part of your current workaround?
- (Last five minutes) Could I show you a rough version and watch you try it?
Notice the syntax. Concrete, past-tense, no hypotheticals. The gap between what people say and what they do is where most startup market research dies.
What are the 5 P's of market research?
The five P's are product, price, place, promotion, and people — the classic marketing mix, extended with the human element that matters most for a young company. Not every framework uses "people," and honestly the version I've found most useful swaps it for problem: is this a real, painful, recurring problem?
For a startup, that reframing is the whole point. Product-market fit is not about the mix being elegant. It's about whether the problem is sharp enough that people open their wallet without being convinced.
Can ChatGPT do market research?
Partly. It can summarize the competitive landscape of a category, draft an interview script, organize forty messy interview notes into themes, and pressure-test your pricing logic. That's genuinely useful — I use it for synthesis constantly.
What it cannot do: tell you what a specific pharmacist in a specific city actually does on a Tuesday morning. Every time I asked a model to estimate a market size for me, I got a number, formatted beautifully, and untraceable. Treat every AI-generated figure as a hypothesis, not a fact. Verify against a primary source or drop it.
What is the 80/20 rule for startups?
Roughly 80% of your results come from 20% of your inputs — and in early-stage market research, that 20% is direct conversation with people who have the problem. Everything else (market sizing models, persona decks, competitor matrices) is support material.
In practice I decided, after a few painful cycles, that I would not touch a spreadsheet model until I'd spoken to at least ten real prospects. Every single time I broke that rule, I ended up rebooting the model anyway because the assumptions were wrong.
The numbers nobody quotes correctly
Is it true that 90% of startups fail?
No — that figure is a stubborn simplification that gets repeated until it sounds like law. Long-term survival studies of new employer businesses show failure rates climbing steadily over the first decade, and the honest number depends entirely on how you define "startup," "fail," and the measurement window. Some cohorts land near 50% within five years; others look better or worse. Anyone quoting a single clean percentage without those definitions is guessing.
The useful takeaway is directionally true and easy to act on: a large share of young companies don't survive their first decade, and the common failure modes — no market need, running out of cash, wrong team, outcompeted — are exactly the ones discovery interviews catch. Don't memorize the number. Learn the failure modes.
What research method fits which question
| Question you're asking | Right method | Time cost | Reliability |
|---|---|---|---|
| Does this market exist at all? | Secondary data (filings, industry bodies, pricing pages) | 2 days | High for size, low for demand |
| Does this person feel the pain? | 5–10 problem interviews | 1 week | High if you ask about past behavior |
| Will they pay this price? | Pre-sale, deposit, or letter of intent | 2 weeks | Highest available |
| Is my messaging clear? | Cheap ad test to a landing page | 3–5 days, small budget | Moderate; noisy at low volume |
| Who is actually winning? | Competitor teardown: pricing, onboarding, review complaints | 1 day | High on facts, low on motives |
That last row is where I've found the cheapest competitive intelligence. Read the one-star and three-star reviews of your competitors. The complaints are your feature roadmap, written for free by people who already paid.
Where founders get research wrong
The most common mistake is not laziness. It's asking leading questions and calling it validation. "Wouldn't it be great if…" is a sentence that has never produced a usable data point.
- Interviewing friends. They're polite. Find strangers through communities, forums, or cold outreach.
- Confusing interest with demand. "That sounds cool" is worth nothing. A €50 deposit is worth a lot.
- Defining success criteria after the fact. Decide in advance what result would make you abandon the idea. If nothing would, you're not researching.
- Stopping at ten conversations because you got the answer you wanted. Keep going until responses start repeating or contradicting.
I once ran a "validation" survey where 82% of respondents rated the concept 4 or 5 out of 5. I felt great for about four days. Then I realized the survey had been sent to my own mailing list — people who liked me, not necessarily the product. Trash data, wrapped in a nice chart.
What you should actually remember
The best market research a startup can do is embarrassingly cheap and slightly uncomfortable: get on a call with someone who has the problem, ask them what they did last time, and shut up while they answer. Everything else — the frameworks, the five P's, the tools, the polished market-size slide — is scaffolding around that one conversation.
If you take nothing else from this piece, take this: write down, today, the assumption you're most afraid to test. Then book three calls this week to test it. The fear doesn't shrink. Your picture of reality does.
