Market research techniques: how to choose the right one for each question

Market research techniques are the methods for reducing uncertainty before making a decision: launching a product, adjusting pricing, choosing a channel, writing copy, prioritizing segments. They aren't ends in themselves. They are tools for answering a specific question, and that's the first thing people forget.

Most research projects fail before they begin—not because the technique is poorly executed, but because no one defined what decision the result will unlock. Without that question, any technique produces a report that gets printed and filed away. That's why this article isn't an abstract catalog: it's a guide to choosing.

The first step isn't the technique, it's the question

Before deciding whether to run interviews or surveys, it's worth answering three internal questions:

1. What decision will you make with the results? If you don't know, don't research; clarify the problem first.
2. What kind of answer do you need—language, magnitude, or cause? Each one points to a different type of technique.
3. What precision do you need and how much can you spend? A well-run interview costs more in time than an automated survey, but it answers something different.

Skipping this is what produces the classic "we have tons of data but we don't know what to do with it."

The two families: qualitative and quantitative

Qualitative research

Answers the why and the how. Its unit is the case, not the figure. It's the best technique when you don't yet know exactly what to look for and you need to understand language, motivations, contradictions.

• In-depth interviews. An open conversation with non-leading questions. 5-10 done well usually reveal 80% of the patterns. Beyond that you hit diminishing returns.
• Direct observation. Watching how the customer uses a product or solves the problem today without your solution. It tends to reveal workarounds that no survey captures.
• Transcript analysis. Sales calls, demos, support. Material you already have and almost no one mines.
• Active listening on social media. Spontaneous conversations in forums, communities, reviews. The most honest opinions rarely reach your direct channel.

Quantitative research

Answers how much and how many. Its unit is the number, and it needs sufficient samples to be reliable.

• Surveys. Useful when you already know which hypothesis to validate. Useless for discovering what you don't know—the question limits the answer.
• Web and product analytics. Real behavioral data, not stated opinion. The difference between what people say they do and what they actually do is usually enormous.
• CRM analysis. Close rates by segment, average cycle, reasons for no-close. The truth about your funnel is here, not in intuitions.
• Keyword research. Search volume reveals real demand. It's one of the few places where the customer has already voted with their time.
• A/B testing. Comparing two versions of an offer, copy, or pricing. The only reliable way to know what works is to put it to the test.

The practical rule: qualitative to discover, quantitative to validate. If you swap them, you discover the little you already knew and validate what you didn't understand.

Internal sources: the gold that goes unmined

Before doing new research, almost every company has untapped deposits:

• Support tickets — real frustration, unfiltered.
• Lost sales calls — the reasons for not buying that no form captures.
• Internal searches on your website — what people looked for and didn't find.
• Search Console — the literal language people use to reach you.
• Untagged social mentions — criticism that doesn't tag you but talks about you.
• Competitor reviews — what people buy from them, what's excessive, what's missing.

Looking at what you already have for a week before launching new surveys usually saves half the project.

Common mistakes that kill research

• Asking about the future. "Would you pay for X?" is the most useless question. People overestimate what they'd do. Ask about the past: "what did you do the last time you had this problem?"
• Leading questions. "Would you like our product to be faster?"—everyone will say yes. That's not data, it's a bias.
• Not asking "and then what." The first "just because" isn't the real cause. Follow-up questions are where the insight appears.
• Confusing opinion with data. A conversation with three friendly customers is a hypothesis, not a conclusion.
• Small sample sizes in surveys. 30 responses to a survey validate nothing statistically, though they may be useful qualitatively—but you have to say so explicitly.
• Not leaving room for "none of the above." Forcing the user to choose between options that don't represent them produces garbage data.
• Not documenting prior hypotheses. Without that, the team reads the results confirming what they already believed.

How to choose a technique based on the question

A quick guide:

• How does the customer describe the problem? → Interviews + analysis of sales transcripts.
• What percentage of the market has this problem? → Survey of a representative sample.
• Why does the landing page convert at 1%? → Web analytics (where they drop off) + A/B testing (what works better).
• What solutions do they consider before ours? → Interviews + Google search + competitor analysis.
• What matters more when deciding, price or speed? → Conjoint survey or A/B testing with segments.
• Why do customers cancel? → Interviews with churned customers + CRM analysis of patterns.
• What content is missing to close? → Direct question to sales + analysis of the last 20 lost opportunities.

One technique answers one question. If your project contains five distinct questions, that's five mini-investigations, not one.

Research and creative operations

Research only adds value if the findings travel to the team that produces content, message, or product. A perfect report that lives in a Drive folder changes nothing. An insight that reaches the brief, the editorial calendar, and the next campaign does.

That's why research is part of the creative operations cluster: it feeds the empathy map, informs brand management, and connects with creative KPIs (what gets measured and why). Without that circuit, research is theater.

At Polimake that logic is spread across three surfaces of the same product: Studio to connect findings with campaign and roadmap decisions, Studio so the customer's real language makes it into the pieces, and Media as a repository where research, transcripts, and guides are accessible to whoever needs them.

When to research and when not to

Research has a cost. It doesn't always pay off. Three situations where it does:

1. You're about to make an expensive decision and you have no data on one of the key risks (price, market, segment).
2. Your hypothesis is based on shared intuition and no one has tested it against recent reality.
3. Something changed (market, behavior, competition) and your mental model is already outdated.

And two situations where it doesn't:

1. The decision is reversible and cheap. Launch it, measure, adjust. Researching something that costs the same as testing it is procrastination.
2. You already have the data but you don't look at it. What's missing isn't new data; it's the discipline to look at what's there.

Related concepts

• Market research
• Empathy map
• Active listening on social media
• Market segmentation
• Market segmentation criteria
• Targeting

---

This piece is part of the Polimake glossary and the cluster on creative operations. If you lead strategy or product and need to decide with less intuition and more data, also read empathy map and creative KPIs.