Marketing research is the systematic gathering, analysis and reporting of information to reduce risk in marketing decisions: who the customers are, what they want, what they will pay and whether the message works. The general method is covered in Research Methods; this article focuses on how it is applied to marketing.
1. The marketing research process
| Step | Key question |
|---|---|
| Define the problem and objectives | What decision depends on this? The management problem (should we launch?) becomes the research problem (what share would prefer this concept at this price?) |
| Design the research | Exploratory, descriptive or causal; sources and methods |
| Collect data | Fieldwork with quality control |
| Analyse | Clean, tabulate, test, model |
| Report and act | Clear findings and recommendations |
A marketing information system (MIS) combines internal records, marketing intelligence, research and analytics. Before commissioning a costly study, ask whether the value of the information exceeds its cost; if existing data or a quick experiment answers the question, use that.
In the real world: New Coke (1985) is the standard cautionary tale: taste tests showed people preferred the new formula, but the research missed the emotional attachment to the original, and the launch had to be reversed within months. The study answered the wrong question.
2. Qualitative research
Qualitative research explores motives, attitudes and language; it gives depth, not statistics.
| Technique | What it is | Best for |
|---|---|---|
| Focus group | 6–10 people discussing with a moderator | Reactions to concepts, language, ideas |
| In-depth interview | One-to-one, probing | Sensitive topics, experts |
| Projective techniques | Word association, sentence completion, role-play, picture response | Uncovering hidden feelings |
| Ethnography and observation | Watch people in real settings | Real behaviour; unmet needs |
| Diary studies and netnography | Self-reports; online community analysis | Usage over time |
Limits: small samples cannot be projected to a population, and results depend on the researcher’s interpretation.
In the real world: consumer-goods firms like Procter & Gamble send researchers into homes to observe how people actually clean and cook, which leads to products that fit real habits better than survey answers do.
3. Quantitative research, instruments and sampling
Surveys (in person, phone, mail, online, mobile) quantify attitudes and behaviour. Observation in stores counts traffic and dwell times. Experiments test marketing variables, such as price and promotion, with test marketing and A/B testing. Panels track the same households over time (retail audits and BARC TV ratings in India).
Questionnaire design: clear objectives, simple wording, one idea per question, logical order (funnel from broad to specific; sensitive items last), balanced answer options, and a pilot. Question types: open, closed, multiple choice, ranking, rating scales. Scales: Likert, semantic differential, constant sum, purchase intent. Net Promoter Score (NPS) asks “How likely are you to recommend us?” on 0–10 and subtracts % detractors (0–6) from % promoters (9–10).
Sampling (see Research Methods): choose a sampling frame, a probability or non-probability method, and a sample size; watch for non-response bias and bad quotas. Online panels are cheap but may skew toward the connected.
4. Key analysis techniques in marketing
| Technique | Purpose | Marketing use |
|---|---|---|
| Cross-tabulation and chi-square | Relationship between two categories | Brand preference by city |
| t-test and ANOVA | Compare group means | Satisfaction across store formats |
| Correlation and regression | Drivers and prediction | Effect of price and ads on sales |
| Factor analysis | Reduce many variables to a few factors | Underlying dimensions of service quality |
| Cluster analysis | Group similar respondents | Segmentation |
| Discriminant and logit analysis | Predict group membership | Likely buyers vs non-buyers |
| Conjoint analysis | Measure trade-offs among product attributes | Optimal feature and price mix |
| Multidimensional scaling (MDS) | Map brands on a perceptual map | Positioning |
| Time-series and market-mix modelling | Estimate effects of marketing spend | Budget allocation |
Conjoint example: respondents rank hypothetical phones that vary in price, camera and battery; the analysis infers how much each attribute and level is worth, so the firm can see that, for instance, extra battery life is worth a ₹2,000 higher price to a segment, but a bigger screen is not.
5. Applications
| Area | Typical study |
|---|---|
| Product | Concept tests, product tests, usage and attitude (U&A) studies, packaging tests |
| Pricing | Van Westendorp price sensitivity, conjoint, willingness to pay |
| Advertising | Copy testing, pre-testing, recall and recognition, brand-lift studies, tracking |
| Brand | Awareness, image, equity tracking |
| Distribution | Retail audits, mystery shopping |
| Customer | Satisfaction, loyalty, NPS, churn analysis |
| Market | Segmentation, sizing, demand forecasting |
Demand forecasting uses survey, test-market and statistical methods. Test markets limit risk before national launch but alert competitors and cost time.
In the real world: Nielsen-style audits and India’s BARC audience measurement shape which TV shows and ads survive. Online firms use A/B tests and conversion funnels to evaluate each change, a continuous form of marketing research.
6. Digital and data-driven research
Marketing research now draws on big data and behavioural data: web analytics, search queries, purchase history, app events, loyalty-card data and social listening (monitoring conversations and sentiment). Methods: sentiment analysis, text mining of reviews, cohort analysis, attribution modelling, and predictive models for churn and propensity to buy. Advantages: speed, scale and real behaviour. Cautions: privacy and consent, bias in who generates data, correlation mistaken for cause, and the need to validate against sound research.
7. Report, ethics and quality
A research report opens with an executive summary (findings and recommendations in one page), then objectives, method, results with clear tables and charts, limitations and appendices. Tell the story: what we learned, why it matters, what to do.
Ethics: honest reporting (no selective results), informed consent, anonymity and confidentiality, no deception that harms participants, protection of children and vulnerable groups, and no disguising selling as research (“sugging”) or fundraising as research (“frugging”). Quality depends on the right sample, tested questions, supervised fieldwork, data checks and clear interpretation. Typical errors are leading questions, social desirability bias (people say what looks good), and over-reading small differences.
One-minute recap
| Topic | Key idea | Where you meet it |
|---|---|---|
| Process | Define the decision problem first | Launch and pricing decisions |
| Qualitative | Focus groups, interviews, ethnography | Concept and language tests |
| Quantitative | Surveys, panels, experiments, scales, NPS | Tracking and measurement |
| Analysis | Cross-tabs, regression, cluster, conjoint, MDS | Segmentation, pricing, positioning |
| Applications | Product, price, ad, brand, distribution | Market studies |
| Digital research | Social listening, analytics, A/B testing | Online businesses |
| Ethics and quality | Consent, honesty, bias control | Every study |
Next in MBA Revisited: Product and Brand Management.
Written by Claude, Anthropic’s AI assistant, as part of MBA Revisited, a quick refresher of the PGDM subjects one at a time.