Global market intelligence platform Industries · Countries · Cities · Companies · Data
Research Method

Market Research

Market research exists to replace assumption with evidence before money is committed — to a product launch, a market entry, a price change, or an investment thesis. It splits into two data types (primary, collected first-hand for the specific question at hand; secondary, already collected by someone else for a different purpose) and two method families (qualitative, which explains why and how, via interviews, focus groups, and ethnographic observation; quantitative, which measures how many and how much, via surveys, experiments, and statistical analysis of transaction or usage data). Good market research almost always triangulates: it starts with secondary research to frame the problem cheaply, then uses primary research to close the specific evidence gaps secondary sources cannot answer, and it pairs qualitative depth with quantitative reach rather than relying on either alone. The output feeds directly into the market-sizing, segmentation, and strategic-framework methods elsewhere on this site — those methods are only as reliable as the research data underneath them.

Definition

What this method is.

A precise definition, its boundaries, and when it applies -- before any formula or worked example.

Definition

Market research is the systematic process of gathering, recording, and analyzing data about a market's customers, competitors, and broader environment in order to reduce uncertainty in a business decision. The American Marketing Association's official definition frames it as "the function that links the consumer, customer, and public to the marketer through information—information used to identify and define opportunities and problems; generate, refine, and evaluate actions; monitor performance; and improve understanding of it as a process" (AMA, definition approved 2017). In practice, market research is the data-collection and validation layer underneath every other method on this site: before you can size a market (see Market Size, TAM/SAM/SOM), segment it, or run a PESTLE or Porter's Five Forces read on it, you need defensible primary and/or secondary data about who the customers are, what they need, who else is competing for them, and what external forces will shape demand.

Scope and exclusions

In scope: the methodology of gathering original data (primary research — surveys, interviews, focus groups, field observation, experiments) and of locating and validating existing data (secondary research — government statistics, industry association reports, company filings, syndicated data services); sampling and sample-size discipline; qualitative vs quantitative method selection; survey and interview instrument design; and the ethical/legal guardrails professional researchers operate under (informed consent, data protection, non-coercion). Out of scope, and covered by sibling method pages instead: the arithmetic of turning research data into a market-size figure (see Market Size and TAM, SAM and SOM); the strategic frameworks applied to research findings once collected (see Market Analysis, PESTLE Analysis, Porter's Five Forces, Value-Chain Analysis); segmentation logic itself (see Market Segmentation); and financial or legal due diligence, which follows different standards (audit, GAAP/IFRS, disclosure law) than market research does. Market research also does not include internal operational data analysis (e.g., analyzing your own CRM funnel) except where that data is being used as one input alongside external research.

When to use it

  • Before launching a new product or entering a new country/city market, to validate that real demand exists before building or spending against it
  • Before finalizing a market-size estimate (TAM/SAM/SOM) or growth forecast, since sizing math is only as good as the underlying demand and pricing data feeding it
  • When writing a business plan, investment memo, or board deck that will be scrutinized on its evidence base, not just its logic
  • To test pricing, packaging, or positioning with real prospective buyers before a launch, rather than after
  • To track brand perception, customer satisfaction, or competitive share over time via repeated (tracking) surveys
  • When internal data (CRM, product analytics) can explain what existing customers do but cannot explain why non-customers are not converting, or what the total addressable population looks like beyond your existing funnel
  • Before a market-entry or M&A decision, to sanity-check target-market assumptions against independent primary and secondary evidence rather than the deal team's own priors
Application

How to apply it.

A repeatable step-by-step procedure, the underlying formula where one exists, and a worked example using illustrative numbers.

Step by step

Define the research problem and objectives
DetailWrite down the specific decision this research needs to inform and the exact questions that must be answered to make it (e.g., "will mid-market logistics firms in Brazil pay for a route-optimization add-on, and at what price"). This is step one of Philip Kotler's widely taught marketing-research process (Kotler & Keller, Marketing Management) and the step most projects under-invest in — a vague objective produces data nobody can act on.
Develop the research plan
DetailDecide primary vs secondary, qualitative vs quantitative, and the specific instruments (survey, discussion guide, experiment design), sample frame, and sample size needed to answer the objective at an acceptable confidence level and cost. This is also where legal/ethical review happens: the ICC/ESOMAR International Code on Market, Opinion and Social Research and Data Analytics requires research to be legal, honest, transparent, and conducted with due care toward the people being studied.
Identify and evaluate secondary sources first
DetailBefore commissioning anything original, exhaust free/cheap existing data: national statistics agencies, central banks, trade and industry associations, regulator filings, company 10-Ks/annual reports, and prior syndicated studies. Secondary research is faster and cheaper than primary research and often narrows or eliminates the questions primary research would otherwise need to answer.
Design the primary research instrument and sampling plan
DetailDraft the survey or interview guide, pilot it on a small group to catch ambiguous or leading questions, define the target population and sample frame, and calculate the required sample size (see Formula below) at the confidence level and margin of error the decision requires.
Collect the data (fieldwork)
DetailField the survey, run the interviews/focus groups, or execute the experiment, monitoring response rates and data quality as they come in rather than only at the end. Both quantitative fieldwork (survey platforms, panel providers) and qualitative fieldwork (recruiting and moderating interviews) need a documented, repeatable process so the results can be audited later.
Analyze the data
DetailQuantitative: run descriptive and inferential statistics, check for non-response and sampling bias, and test whether findings are statistically significant, not just directionally interesting. Qualitative: code and theme the interview/focus-group transcripts and check whether the sample has reached thematic saturation — the point where additional interviews stop surfacing new themes (Guest, Bunce & Johnson, 2006, found saturation was frequently reached within the first 12 interviews in their study, though the right number varies with population homogeneity and topic complexity).
Report findings and translate them into the decision
DetailPresent findings against the original objective from step 1, explicitly flag confidence levels/margins of error and any known sampling limitations, and hand the validated data to whichever downstream method (Market Size, TAM/SAM/SOM, Market Segmentation, PESTLE, Porter's Five Forces) will turn it into a sizing figure or strategic read.
Follow up
DetailTrack whether the decision made on the research played out as the data predicted, and feed that outcome back into how the next research plan is scoped — Kotler's process explicitly includes this follow-up step, and it is the one most often skipped in practice.

Formula

Name

Cochran's sample-size formula (for quantitative survey research)

Expression

n₀ = Z² × p(1 − p) / e² — with finite-population correction: n = n₀ / [1 + (n₀ − 1)/N]

Variables
N₀

required sample size for an infinite/unknown population

Z

z-score for the desired confidence level (1.96 for 95% confidence, the standard used in most market and social research)

P

expected proportion of the population with the attribute being measured; use 0.5 when unknown, since p(1−p) is maximized at p=0.5, which yields the most conservative (largest) sample size

E

acceptable margin of error, expressed as a decimal (e.g., 0.05 for ±5 percentage points)

N

known size of the finite target population, used to correct n₀ downward when the population itself is small relative to n₀

Originating source

William G. Cochran, Sampling Techniques, 3rd edition (Wiley, 1977)

Note

This formula sizes a quantitative survey sample for a given confidence level and margin of error; it is not a market-sizing formula — for turning research data into a dollar market-size estimate, see the Market Size and TAM, SAM and SOM method pages.

Worked example ILLUSTRATIVE

Scenario

A B2B SaaS company sells operations software to mid-market logistics firms and has identified a total addressable population of 8,000 companies in its target country. Before greenlighting a new route-optimization module, it wants to survey this population to validate willingness to pay, at 95% confidence and a ±5 percentage-point margin of error.

Calculation steps
  • Step 1 — base sample size (infinite population assumption): n₀ = Z² × p(1−p) / e² = 1.96² × 0.5 × 0.5 / 0.05² = 3.8416 × 0.25 / 0.0025 = 384.16, rounded up to 385.
  • Step 2 — finite population correction, since the target population (N = 8,000) is known and finite: n = n₀ / [1 + (n₀−1)/N] = 385 / [1 + 384/8000] = 385 / 1.048 ≈ 367.
  • Step 3 — inflate for expected non-response: if the company's prior surveys of this segment saw a 25% completion rate, it needs to reach roughly 367 / 0.25 ≈ 1,468 companies to end up with 367 completed responses.
  • Step 4 — complement with qualitative depth: alongside the quantitative survey, the company runs 12–15 in-depth interviews with prospective buyers to understand the 'why' behind stated willingness to pay; per Guest, Bunce & Johnson (2006), thematic saturation in reasonably homogeneous samples is often reached within a similar range of interviews, though the team keeps interviewing past the point new themes stop appearing rather than stopping at a pre-fixed number.
Result

The company needs approximately 367 completed survey responses (from an outreach pool of roughly 1,468, assuming a 25% completion rate) to be 95% confident that its measured willingness-to-pay figure is accurate to within ±5 percentage points, plus 12–15 qualitative interviews to explain the reasoning behind that number.

Common mistakes

Where analysts go wrong.

The most frequent errors made when applying this method, so you can check your own work against them.

Common errors

Sampling only existing customers when the research question is about the broader addressable market — this systematically excludes the non-customers whose behavior the research is usually trying to explain (survivorship bias).
Writing leading or double-barreled survey questions that push respondents toward the answer the research sponsor wants to hear, rather than measuring their actual view.
Treating a convenience sample (e.g., your own email list or social followers) as if it were a representative sample of the target market, then quoting the result with false statistical precision.
Confusing stated intent with revealed behavior — the 'say-do gap': survey respondents routinely overstate willingness to pay or intent to purchase relative to what they do when actually asked to spend money, so primary research on price/demand should be triangulated against behavioral or experimental evidence where possible.
Skipping secondary research and re-collecting, at primary-research cost, data that a government statistics agency, industry association, or public company filing already publishes for free.
Running only one method (e.g., only a survey, or only interviews) rather than triangulating primary and secondary, or qualitative and quantitative, evidence against each other.
Under-sizing the sample for the confidence level being claimed, or over-claiming statistical significance from a qualitative sample that was never designed to be statistically representative.
Conflating market research with market sizing — assuming a demand survey by itself produces a TAM/SAM/SOM figure, when sizing requires a separate, documented methodology (see Market Size and TAM, SAM and SOM).
Ignoring non-response bias: the segment of the population that declines to respond to a survey or interview request is rarely identical to the segment that responds, and low response rates should be flagged as a limitation, not silently ignored.
Skipping the ethical/legal basics — informed consent, data protection, and transparency about who is sponsoring the research — that the ICC/ESOMAR Code sets as baseline professional obligations, not optional extras.
Related

Related methods and tools.

Other frameworks that pair with this one, and the calculators/tools that implement it.

Related tools

Not yet available.

Further reading

  • Kotler, P. & Keller, K.L., Marketing Management (Pearson) — Chapter on the marketing research process (the 6-step problem-definition → plan → collect → analyze → report → follow-up sequence referenced above)
  • ICC/ESOMAR International Code on Market, Opinion and Social Research and Data Analytics (2025 edition) — the global self-regulatory standard for research ethics and methodology
  • Cochran, W.G., Sampling Techniques, 3rd ed. (Wiley, 1977) — originating source for the sample-size formula above
  • Guest, G., Bunce, A. & Johnson, L. (2006), "How Many Interviews Are Enough? An Experiment with Data Saturation and Variability," Field Methods 18(1)
Trust & methodology

Sources and review.

Every important figure on this page is traceable to a dated source. This page was last human-reviewed on an unrecorded date.

American Marketing Association, Definitions of Marketing Research (approved 2017) American Marketing Association · Published 2017-01-01 · Accessed 2026-07-15 View source →
ICC/ESOMAR International Code on Market, Opinion and Social Research and Data Analytics, 5th edition ESOMAR / International Chamber of Commerce · Published 2025-01-01 · Accessed 2026-07-15 View source →
Guest, G., Bunce, A. & Johnson, L. (2006), "How Many Interviews Are Enough? An Experiment with Data Saturation and Variability," Field Methods 18(1), pp. 59–82 SAGE Publications / Field Methods · Published 2006-02-01 · Accessed 2026-07-15 View source →
Cochran, W.G., Sampling Techniques, 3rd edition John Wiley & Sons · Published 1977-01-01 · Accessed 2026-07-15
Kotler, P. & Keller, K.L., Marketing Management (marketing research process chapter) Pearson · Published 2016-01-01 · Accessed 2026-07-15
Search captured locally. No data was submitted.