Every guide to market research methods is published by someone selling one of them, which is why the type counts differ and the costs never appear. Here is the map with the prices left in — what each method does, what it runs, and how to choose.
By the Wavefield Research team · Published Aug 16, 2026 · Updated Aug 23, 2026
The short answer
Market research methods split along two axes: primary versus secondary (collect new data, or use existing sources) and qualitative versus quantitative (understand, or measure). The working methods are surveys, interviews, focus groups, observation, and secondary research. Choose by matching the method to the question, the budget, and how much certainty the decision needs.
Whether a guide says market research methodologies, market survey methods, or counts seven types against another's eleven, everything in it sorts along the same two questions. Primary or secondary: are you collecting new data, or reading what exists? Qualitative or quantitative: are you trying to understand something in depth, or measure how much of it is out there?
The axes also give you the classic workflow: secondary research first to frame the market, qualitative next to discover what matters and how people talk about it, quantitative last to measure which discoveries are true at scale. Most real studies are a sequence across quadrants, not a choice of one box.
The method map — two axes, four families
Primary + Quantitative
Surveys, polls, concept tests, trackers
Primary + Qualitative
Interviews, focus groups, ethnography
Secondary + Quantitative
Syndicated data, sales analytics, census tables
Secondary + Qualitative
Published studies, reviews, social listening
Every method in every listicle lands in one of these four boxes. The highlighted one is where most business decisions end up, because it is the only quadrant that produces defensible numbers about your own market.
Cost figures are honest order-of-magnitude ranges as commonly reported in mid-2026 — every project varies, but the relative picture is what drives real method choices, and it is the part most guides omit.
Cost: DIY from ~$99 per study; panel sample a few dollars per complete
The workhorse of quantitative research: structured questions, real sample sizes, results you can chart and significance-test. Modern platforms automate the programming and analysis, which moved surveys from specialist work to something a small team runs in a week. The craft that remains is question wording, sampling, and honest analysis.
Cost: Commonly $150–$500 per interview all-in (recruiting, incentive, moderation)
Thirty to sixty minutes, one respondent, a skilled interviewer, and the freedom to follow surprises. Interviews find the why behind behavior and vocabulary you did not know customers used. Eight to fifteen interviews usually reach saturation for one audience — after that you hear repetition, not news.
Cost: Commonly $4K–$12K per group (facility, recruiting, incentives, moderator)
Six to ten people reacting to concepts together. The group dynamic is the point and the problem: discussion sparks reactions an interview misses, and loud voices steer quiet ones. Groups earn their cost for stimulus-response work — pack designs, ad rough cuts — less so for questions individuals can answer alone.
Cost: Labor-heavy; from store-visit cheap to multi-week studies
Watching what people actually do — in stores, in homes, in usage data — instead of what they say they do. The gap between reported and observed behavior is one of the most reliable findings in the field. Slow and unscalable, but immune to the biases that haunt every asked question.
Cost: Free to a few thousand for syndicated reports
Reading what already exists: census tables, industry reports, competitor filings, academic work. Always the first step, never the last — it frames the market and stops you paying to rediscover known facts, but it answers yesterday's questions about everyone's market, not today's about yours.
Cost: Free to ~$500+/month for tooling
Mining what customers say unprompted — reviews, forums, social posts. Honest in a way surveys are not (nobody is being polite) and biased in a way surveys are not (only the vocal appear). Good for discovering language and complaints; dangerous as a measurement of prevalence.
Cost: Mostly internal data; tooling varies
Your own transaction, usage, and traffic data. It answers what is happening with certainty and says nothing about why, or about the people who never became customers. Pairs naturally with a survey: analytics finds the anomaly, the survey explains it.
The survey column of this landscape has its own deep guides here: the six study types surveys run — trackers, concept tests, U&A, pricing — are covered with sample questions in market research surveys, type by type; how surveys reach people is covered in survey data collection methods (the fielding modes — online, SMS, phone, mail), what goes in them in survey question types, and how respondents get selected in quota sampling vs. random sampling.
| Method | Cost | Speed | What you get | Run it yourself? |
|---|---|---|---|---|
| Survey | $ (DIY) – $$ | Days | Numbers you can test | Yes — platforms automate it |
| Interviews | $$ | 1–3 weeks | Depth, vocabulary, the why | Partly — moderation is a skill |
| Focus groups | $$$ | 2–4 weeks | Group reactions to stimuli | Rarely — facility + moderator |
| Observation | $$–$$$ | Weeks | Actual behavior | Sometimes |
| Secondary research | Free – $$ | Hours–days | Market context | Yes |
| Social listening | Free – $ | Ongoing | Unprompted language | Yes |
| Analytics | Internal | Ongoing | What is happening | Yes |
Most method guides are published by companies selling one of the methods, so prices stay vague. The honest version: secondary research is nearly free, surveys went from five figures to double digits when programming and analysis were automated, interviews cost hundreds per conversation, and focus groups thousands per room. Method choice is budget allocation — pretending otherwise serves the vendor, not you.
A survey of 40 volunteers and a survey of 400 quota-managed respondents are different instruments wearing the same name. Whatever method you choose, the questions that decide whether the output is evidence are the same: who was asked, how many, and what was done to correct the skews — sample size, quotas, weighting.
The classic sequence runs qualitative first to find the categories, language, and hypotheses, then quantitative to measure which of them are true at scale. Reversing it — running a survey before you know what to ask — produces confident numbers about the wrong questions. The sequence matters more than the method list.
An afternoon of desk research before any primary spend is the highest-ROI step in the field. It kills questions that are already answered, sharpens the ones that are not, and gives you the population numbers you will need later for quotas and weighting targets anyway.
The rigor layer has its own guides: how many respondents a study needs (sample size calculator), how skewed samples get corrected (survey weighting, explained), and the full workflow from raw responses to defensible findings (how to analyze survey results).
Start from what the decision needs. How many, how much, what share? That is a survey — the only affordable method that produces testable numbers about your specific market. Why, and in what words? Interviews, or an open-ended layer inside a survey. How do people react to this thing? Concept tests with rotations in a survey, or a focus group when the stimulus needs a room. What is already known? Desk research, before any money moves.
Then let the budget arbitrate. A team with five thousand dollars can afford one focus group, or a dozen interviews, or several properly weighted surveys with room left over — and for most commercial questions, the survey path produces the evidence that survives a boardroom. That arithmetic, more than any taxonomy, is the real method decision, and it is the arithmetic the vendor-written guides skip.
When the answer is a survey, that is what Wavefield does end to end: an AI agent programs the questionnaire from a brief, fielding runs through your lists or links with quotas enforced, and the analysis comes back weighted with significance testing on honest bases — from $99 per project. Feedback programs live in customer feedback surveys; polling in political polling.
It spans four orders of magnitude, which is why market research pricing feels opaque. Secondary research runs free to a few thousand dollars for syndicated reports. A self-served survey on a modern platform costs $99 to $299 per study, plus panel sample at a few dollars per complete if you need respondents sourced. In-depth interviews commonly run $150 to $500 each all-in. Focus groups run $4,000 to $12,000 per group, and a full-service agency study typically starts around $15,000 to $30,000. The pattern: you are paying for skilled labor, so the more of the programming and analysis a platform automates, the further your budget goes.
Five working methods cover practice: surveys (quantitative measurement), in-depth interviews and focus groups (qualitative understanding), observation (actual behavior), and secondary research (existing sources) — with social listening and behavioral analytics as newer supplements. They all sort along two axes: primary versus secondary, and qualitative versus quantitative. Most studies combine two or three methods in sequence rather than choosing one.
Because the counts are taxonomy choices, not facts. One guide counts "branding research" and "product research" as separate types; another folds them into surveys and interviews by use case. The famous numbered lists are headlines, and each vendor's number conveniently matches its table of contents. What is stable underneath: two axes (primary/secondary, qualitative/quantitative) and a half-dozen working methods. Learn the axes and every listicle becomes readable.
The "5 P's" (product, price, place, promotion, people) are really the marketing-mix framework, borrowed as a checklist of topics a research program should cover — they are not research methods. They are useful for scoping what to study; the methods on this page are how you study it.
Primary research collects new data for your specific question — your survey, your interviews, your observation. Secondary research uses data that already exists — census tables, industry reports, published studies. Secondary is faster and cheaper and should come first; primary is the only way to answer questions nobody has asked before about your specific market, customers, or concepts.
Upload a brief and the agent programs the study — questions, quotas, rotations — with weighted, significance-tested results built in. Free to build, from $99 to field.
Start a project