How to Do Market Research: A Practical Guide
This guide is for brand managers, product teams, and founders who need reliable consumer data and don't have a research department. It covers the types of market research, when to use each method, how to design a study that produces usable results, and how to collect data without a six-figure agency budget.
What market research is — and what it isn't
Market research is the practice of collecting and analyzing information about your target consumers, your market, and your competitors to make better business decisions. It's the difference between building and pricing a product based on what you think the market wants versus what the market tells you it wants.
That gap — between assumption and evidence — is where most product and marketing decisions go wrong. Market research doesn't eliminate uncertainty, but it narrows it substantially before you commit resources.
What market research is not:
It's not just competitive analysis. Tracking what competitors are doing is useful, but it tells you what they think the market wants — not what the market actually wants. Real market research goes directly to consumers.
It's not focus groups as a primary method. Focus groups are qualitative and directional — they surface hypotheses, not conclusions. Quantitative survey research is what produces statistically reliable data. The two complement each other; qualitative research is best used to inform what to ask in quantitative research, not to replace it.
It's not one-size-fits-all. Different business questions require different research methods. The right research design depends on what you're trying to learn.
Secondary research
(also called desk research) uses existing data — industry reports, government statistics, competitor analysis, published studies, and market sizing reports. It's fast and often free or low-cost. It's useful for understanding market size, industry trends, and competitive context. Its limitation: it tells you about the market in general, not about your specific product, your specific brand, or your specific target segment.
Primary research
is research you conduct yourself, directly with your target consumers. Surveys, interviews, observation, usability tests, concept tests — any method that generates new data specific to your question. Primary research is slower and costs more than secondary research, but it answers questions secondary research can't: Does our target market want this product? How do they perceive our brand relative to competitors? What price will they accept?
For most brand and product decisions, both are needed. Secondary research gives you the market context; primary research gives you the consumer data specific to your situation.
Quantitative vs. qualitative market research
Quantitative research produces numerical data: percentages, averages, scores, distributions. It's collected via structured surveys with predefined answer choices. The goal is to measure something — how many people, what proportion, how strong — with statistical confidence.
Use quantitative research when you need to know:
How many consumers would buy this product at this price?
What percentage of our target market is aware of our brand?
Which of these three product concepts scores highest on purchase intent?
How does our CSAT score compare to last quarter?
Qualitative research produces non-numerical data: words, themes, narratives. It's collected via interviews, focus groups, open-text survey responses, and observation. The goal is to understand why — what motivates consumers, how they think about a category, what language they use to describe a problem.
Use qualitative research when you need to know:
Why are customers churning from our product?
What words do our target consumers use to describe this problem?
What's missing from our product that would make someone switch from a competitor?
Why did our concept test score poorly on uniqueness?
In practice, they complement each other. Qualitative research generates hypotheses; quantitative research tests them at scale. A common research process: run 5–8 customer interviews to understand the problem space, then design a quantitative survey based on what you learned, then field it to 300+ respondents to measure which findings are representative.
The main market research methods and when to use each
Survey Research
The most common primary research method. Structured questionnaires administered to a defined sample of consumers. Can be fielded online to your own audience or to a panel of pre-recruited respondents.
Best for: quantitative measurement of any kind — brand awareness, purchase intent, price sensitivity, customer satisfaction, concept appeal, feature preferences.
Not ideal for: deep exploration of motivations (use interviews) or observational research (use usability testing).
When to use it: any time you need statistically reliable data from a defined sample — which is most of the time in brand and product research.
Customer Interviews
One-on-one conversations with customers or target consumers. Unstructured or semi-structured, usually 30–60 minutes. The interviewer probes open-ended responses to understand the "why" behind behaviors and attitudes.
Best for: understanding motivations, uncovering unmet needs, generating hypotheses for quantitative research, understanding churn reasons in depth.
Not ideal for: measuring how common a finding is — you can't statistically generalize from 8 interviews.
When to use it: early-stage research before a quantitative study, or after a quantitative result you need to explain.
Competitive Analysis
Systematic evaluation of competitors' products, pricing, positioning, and market presence. Sources include competitor websites, pricing pages, product reviews, sales call recordings, and analyst reports.
Best for: understanding the competitive landscape, identifying positioning gaps, benchmarking your product against alternatives.
Not ideal for: understanding what consumers want (it tells you what competitors offer, not what the market prefers).
When to use it: before positioning decisions, pricing decisions, and product roadmap prioritization.
Brand Tracking
Recurring survey research that measures brand health metrics (awareness, favorability, purchase intent) at regular intervals — monthly or quarterly.
Best for: measuring whether brand investments are working over time, detecting brand health erosion before it becomes a business problem.
Not ideal for: one-time decisions (use a one-time brand validation study instead).
When to use it: when you have a baseline brand health measurement and want to track movement.
Concept Testing
Quantitative survey research that measures consumer reaction to a product idea — appeal, uniqueness, and purchase intent — before the product is built.
Best for: go/no-go decisions on product development investment, comparing multiple product concepts, validating a pivot before executing it.
Not ideal for: usability feedback (the product doesn't exist yet) or pricing decisions (use Van Westendorp for that).
When to use it: before committing development resources to a product concept.
Price Sensitivity Research
Quantitative survey research that measures the range of acceptable prices for a product. The most common methodology is the Van Westendorp Price Sensitivity Meter.
Best for: setting the right price for a new product, testing how much headroom you have for a price increase, entering a new market with an existing product.
Not ideal for: feature-level pricing trade-offs (use conjoint analysis for that).
When to use it: any time you're setting or changing a price and want data before committing.
→ Pricing Study template — Van Westendorp → → Free Van Westendorp calculator →
Conjoint Analysis
Quantitative survey research that measures how consumers value different product features by forcing trade-off choices between product configurations. Reveals the relative importance of each feature and the willingness to pay for each.
Best for: feature prioritization decisions, packaging and tier design, understanding price-feature trade-offs.
Not ideal for: overall pricing decisions where the product is already defined (use Van Westendorp for that).
When to use it: when you're deciding which features to include in a product or how to structure a product lineup.
How to design a market research study — step by step
Step 1 — Define the decision.
Before choosing a research method, be specific about the decision the research will inform. "We want to understand our customers" is not a research objective. "We want to know whether to set our new product's price at $29 or $49" is a research objective. The more specific the decision, the easier it is to design the right research.
Step 2 — Choose the right method.
Match the method to the question. Price decision → price sensitivity study. Product concept → concept test. Brand health → brand awareness or brand tracking survey. Feature prioritization → conjoint analysis. If you're not sure what to measure first, start with brand validation or idea validation — they establish a baseline for everything else.
Step 3 — Define your target audience.
Your research is only valid for the audience you study. Define precisely who should be in your sample: what category they buy, what demographics they match, what geography they're in. Write a screener question that filters for this audience before they reach your research questions.
Step 4 — Design the questionnaire.
Question order matters. Put screener questions first. For awareness studies, put unaided recall before aided. For concept tests, show the concept before rating questions. Keep surveys under 15 minutes — response quality degrades significantly beyond that. Include one attention check to filter inattentive responses.
Step 5 — Decide on your sample size.
Use the sample size calculator → to find the right number for your target confidence level and margin of error. For most market research studies, 200–500 respondents is the working range. More is not always better — over-sampling a study with weak concept clarity doesn't produce better data.
Step 6 — Field the study.
Either to your own audience (share link via email or your CRM) or to a panel. A research panel is a pre-recruited group of consumers available for survey research. SegmentOS includes a 30M+ respondent panel with demographic targeting across 127 countries. Panel responses from $0.73/response. Cost confirmed before you launch.
Step 7 — Analyze and act.
Quantitative survey data is only useful if it changes or confirms a decision. After fielding, return to the decision you defined in Step 1 and answer it directly with the data. Then share the finding — not the full dataset — with the people who need to act on it. A two-page summary with the key finding, the supporting data, and the recommended decision is more valuable than a 40-slide deck.
How much does market research cost?
The range is enormous — from near-zero to $100,000+. What drives the cost:
Agency research: A full-service research agency charges $5,000–$25,000 for a concept test, $10,000–$40,000 for a brand tracking study, and $15,000–$50,000 for a conjoint analysis. The price covers questionnaire design, fieldwork management, data analysis, and a presentation deck. The methodology is the same as what you can run yourself — you're paying for someone else's time, not a fundamentally better method.
Self-serve panel research: With a platform like SegmentOS, you pay for the platform subscription ($79/month) plus panel responses (from $0.73/response B2C). A 300-respondent concept test in a B2C category runs approximately $220–$300 in panel costs. A 500-respondent brand tracking study runs $365–$500. The methodology is research-grade — the same Van Westendorp, conjoint, and brand tracking methods that agencies use.
Your own audience: If you're surveying your own customers or employees, panel costs are zero. You share a link, they respond, you analyze the results. The Premium plan ($29/month) covers unlimited responses from your own audience.
Surveying the wrong audience
Asking people who'd never buy your product what they think of your product produces meaningless data. A category screener question at the start of every study is non-negotiable.
Leading questions
"How much do you love our product?" tells respondents what answer you want. "How would you rate your overall experience with our product?" doesn't. Every question should be neutral in framing.
Asking about the past to predict the future
"Have you bought [category] in the last 6 months?" is more reliable than "Would you buy [product]?" because past behavior is a better predictor of future behavior than stated intent. Ground screening and segmentation questions in observable past behavior, not hypothetical future intent.
Reading too much into small samples
A survey of 47 people is not statistically representative of anything. Below 150 respondents, don't quote percentages as if they're reliable — the margin of error is too large.
Asking too many questions
Survey response quality drops off after 10–12 minutes. Keep surveys focused on the decision you're trying to make. Everything else is noise.
Treating stated purchase intent as actual demand
Respondents consistently overstate their likelihood to buy new products. Top-2-box purchase intent scores need to be interpreted relative to benchmarks, not read as literal demand forecasts.
The study your agency quotes at $5,000. Self-serve.
Research agencies charge $5,000–$15,000 per study for Van Westendorp pricing analysis, concept tests, and brand tracking. SegmentOS gives you the same instruments, self-serve: free to start, plans from $29/month, panel responses from $0.73 each — with the full cost shown before you launch.
Frequently asked questions
How long does market research take?
A self-serve survey study with a panel typically takes 2–5 days from design to results: 1–2 hours to build and launch the study, 24–48 hours for the panel to deliver responses, and a few hours to analyze. A full agency project — briefing, questionnaire design, fieldwork, analysis, presentation — takes 4–8 weeks. The methodology produces similar results; the timeline difference is primarily administration.
Do I need a research background to run a market research study?
No. Research templates with pre-built question logic, screeners, and attention checks remove the methodological decisions from the process. You need to know what decision you're trying to make, who your target audience is, and how many respondents you need. The template handles the rest.
What's the minimum budget for market research?
The Free plan includes 4 research-grade templates and 500 responses/month at no cost — enough for small-scale research on your own audience. Panel access (to buy respondents) starts at $79/month (Pro plan) plus per-response costs from $0.73. A 200-respondent B2C study with panel costs approximately $150–$200 in panel fees.
Is online survey research reliable?
Yes, with the right quality controls. The risks in online survey research — inattentive respondents, duplicate responses, professional survey takers — are well understood and manageable. Quality controls that address these risks: screener questions (filter for the right audience), attention checks (catch inattentive respondents), device fingerprinting (block duplicates), and speeding detection (remove respondents who completed too quickly). These run automatically on every SegmentOS study.
What's the difference between market research and user research?
The terms overlap but have different emphases. Market research tends to be about the market — consumers in aggregate, brand positioning, purchase intent, competitive landscape. User research (also called UX research) tends to be about specific product interactions — usability, task completion, interface comprehension. Survey-based concept testing and brand awareness studies are market research; usability testing and prototype walk-throughs are user research. The methods and tools differ; both are valuable depending on the question.