School Me MarketingSMM
FoundationsBeginner10 min deep read · 30-sec skim available

Market Research Basics

The discipline of replacing 'we think' with 'we checked': asking, watching, and testing your way to decisions before the budget gets bet on a guess.

⚡ Understand it in 30 seconds

  • Research exists to improve a decision, not to satisfy curiosity. The first question is never 'what do we want to know?'. It's 'what are we about to decide, and what does it cost us to decide it wrong?'
  • There are only three ways to learn about a market: ask people (surveys, interviews, focus groups), watch people (observation, ethnography, analytics), and test on people (experiments, pilots, A/B tests).
  • The central trap is the say-do gap: people are honest but unreliable witnesses to their own future behaviour. Watch for behaviour, ask for meaning, test for causation.
  • HUL's famous rural immersions, where managers live in villages to see how products actually get used, read as the core lesson institutionalised: go and look beats sit and assume.

Go deeper

The core idea

Anchor everything to the decision. 'Should we launch the ₹10 pack?' is researchable. 'Let's understand the youth' is a budget with no brakes. Working backwards from the decision tells you how much rigour to buy: a tagline test doesn't deserve six weeks, and a ₹50-crore launch doesn't deserve a Google Form. The cost of the research should scale with the cost of being wrong.

The three learning modes have different superpowers. Asking is the only way into meaning: why something matters, what it signifies, what buyers fear. But it's polluted by memory, politeness, and self-image. Watching (shop-alongs, home visits, analytics, sales data) captures true behaviour but not its reasons. Testing, the youngest mode, is the only one that proves causation: change one thing, hold the rest, watch what happens. Serious research questions usually need two modes triangulated: ask to generate hypotheses, watch or test to validate them.

Sequence qualitative before quantitative. Interviews and observation on a small scale surface what matters and the words customers use for it. Surveys then measure how much and how many, in those same words. Invert the order, and you launch a big survey before anyone has sat with ten real customers. That produces statistically precise answers to the wrong questions, which is worse than no research because it arrives wearing confidence.

Respect the say-do gap as a law, not a nuisance. Buyers overstate future virtue (healthy eating, higher spending, trying new things) and understate habit's grip. They can't reliably report their own price sensitivity or how an ad worked on them. So calibrate by question type. Behaviour and money questions deserve behavioural evidence, what people did or what they do in a test, while feelings, associations, and language genuinely need asking. New Coke is the eternal reminder: the taste tests answered the sip question correctly and the meaning question not at all.

The business case

Why marketers care

It's the cheapest insurance marketing can buy. Campaigns, launches, and repositionings are expensive hypotheses, and research is the low-cost preview of how they fail. A week of customer interviews routinely kills an assumption that would have cost a quarter's budget to disprove in the market. The ROI of research is measured in disasters not shipped.

It's also the marketer's authority in the room. Product argues from roadmaps, finance from spreadsheets. The marketer who argues from 'here's what twenty buyers said and here's what ten thousand did' stops being the department of opinions. Customer evidence is the one currency every function accepts.

And in a market like India, intuition doesn't scale across the distances. Insight formed in a metro office reliably misfires across languages, income tiers, and town classes. The histories of sachets, rural distribution, and vernacular content are all stories of research (formal or founder-led) beating boardroom assumption. The market keeps punishing companies that extrapolate themselves onto everyone else.

See it

The visual model

The research loop

Research isn't a report. It's a loop that starts and ends at a decision.

Read this diagram as text

A circular loop of six stages: define the decision this research serves; design the study by choosing ask, watch, or test methods to fit the question; collect the evidence; analyse it honestly, hunting for disconfirmation rather than support; decide and act on what was learned; then measure the outcome and refine the next question, which feeds back into defining the next decision.

The receipts

Where it comes from

Formal market research is usually dated to Charles Coolidge Parlin (Curtis Publishing, 1911), often called the first professional market researcher. The qualitative tradition grew through the motivational research era (Ernest Dichter, 1950s). The experimental tradition arrived with direct marketing and matured into today's A/B testing culture.

Parlin's innovation was treating buyers as a subject of systematic study, famously counting garbage to see what people actually consumed rather than what they claimed. The founding gesture of the field was, fittingly, a preference for behavioural evidence over testimony.

The field then split into complementary traditions. The quantitative survey tradition (sampling theory, panels, tracking studies) built measurement at scale. The qualitative tradition (depth interviews, ethnography, focus groups) built understanding of meaning. The methodological wars between them got settled in practice as 'use both, in the right order'.

The digital era added always-on behavioural data and cheap experimentation. That changed the economics but not the underlying logic: analytics tell you what is happening in magnificent detail and stay silent on why. The say-do gap survives every new tool. Modern practice triangulates: behavioural data for what, qualitative for why, experiments for what-if.

Brands you know

Seen in India

Educational readings of familiar brands — how the concept helps you see what they do, not claims about their current campaigns.

Hindustan Unilever

India's FMCG bellwether, famous for grounding managers in rural reality.

Its long-running immersion programmes, where young managers are posted to villages, live with families, and watch how a sachet is cut, stored, and stretched, are commonly cited as ethnography institutionalised: the company buying its future leaders a permanent correction against boardroom extrapolation. The method reads as a bet that watching one home closely beats surveying a thousand at arm's length, at least for finding the questions worth surveying.

What to steal: Build 'go and look' into the operating rhythm, not just into projects. The cheapest way to stay right about a market is to keep physically meeting it.

Nirma

The founder-observed gap that became a detergent empire.

The origin story is regularly told as informal research done perfectly: direct observation that lower-income households wanted clean clothes but found premium detergent prices impossible. That need-and-price gap no survey of existing detergent buyers would have surfaced, because the segment wasn't buying detergent at all. The insight came from watching the market's edge, not its centre.

What to steal: Research your non-customers. Surveys of existing buyers describe the market you have; observation of the excluded describes the market you could create.

Swiggy & Zomato

Food delivery platforms running on continuous behavioural telemetry.

Every session is a study: search terms reveal demand language, drop-off points expose friction, reorder gaps flag churn risk, and citywise patterns steer selection and pricing. The platforms read as 'watch' mode industrialised, with the standing caveat that telemetry explains everything except motive. That's why such companies still run interviews and surveys for the why behind the what.

What to steal: Analytics is observation at scale, not a replacement for asking. Pair the dashboard's what with a human's why before acting on either alone.

Paper Boat

Packaged traditional drinks built on memory and meaning.

The brand's premise, aam panna and jaljeera and childhood nostalgia in a pouch, reads as insight from the 'ask' tradition: the value being packaged is emotional association, which no analytics stream or sales report could have surfaced. Understanding what a drink means (summers, grandmothers, train journeys) is interview-and-culture work, the mode's home turf.

What to steal: When the product is meaning, qualitative isn't the soft option. It's the only instrument that measures the asset you're building.

Beyond India

The global lens

New Coke

The 1985 reformulation backed by 190,000 taste tests — and reversed in 79 days.

The canonical research-misread. The sip tests were well executed and correctly showed a preference for the sweeter formula. What the design never measured was what Coke meant (identity, ritual, ownership), because blind sips can't ask that question. Consumers who 'preferred' the new taste revolted at losing the old brand. The method answered precisely, and answered the wrong question.

What to steal: Rigour can't rescue a mis-specified question. Before trusting any result, ask what the method was structurally unable to see.

Netflix

A product culture that tests what it could have merely asked.

Widely described as an experiment-first shop: artwork, autoplay, ordering, and recommendations decided by controlled tests on real behaviour rather than stated preference, on the working theory that members' clicks and watch-hours testify more honestly than their survey answers. Reads as the say-do gap adopted as an operating principle.

What to steal: Where testing is cheap, prefer revealed preference to stated preference. Ask users for meaning and language, and let experiments decide the behaviourally answerable questions.

From theory to Monday morning

How to use it

  1. Write the decision memo before the research brief

    One page: the decision to be made, the options on the table, what evidence would change the choice, and the cost of choosing wrong. If no plausible result would alter the decision, cancel the research. You're about to buy justification, not information.

  2. Pick the mode by question type

    Route each question deliberately. Meaning, motivation, and language go to ask (interviews first, groups with care). Actual behaviour, frequency, and context go to watch (analytics, observation, sales data). Causation, pricing, and creative choices go to test (experiments, pilots, A/Bs). Flag any behaviour question currently being answered by a survey. That's the say-do gap eating your budget.

  3. Run five interviews before any survey

    A little qualitative first: five to ten conversations to learn what matters and the exact words customers use. Build the questionnaire from their vocabulary, not your jargon. A survey asking about 'value propositions' measures confusion. Then quantify with a sample that mirrors the market, not just your followers.

  4. Hunt disconfirmation in analysis

    Before reading the results, write down what you expect and what would prove you wrong, then look for the second thing first. Report the base ('43 of 61 target-segment respondents'), keep verbatims attached to the numbers, and quarantine findings from anyone's pet conclusion, including yours.

  5. Close the loop after the decision

    When the decision ships, log the prediction the research made and check it against what happened. A simple running scorecard of 'research said / market did' compounds into the rarest asset: knowing which of your methods, vendors, and instincts to actually trust.

Watch out

Common mistakes

✕ Commissioning research to justify a decision already made.

Fix: Justification research is expensive theatre with a side effect: it teaches the organisation to ignore all research. If the decision is locked, save the money. If it's genuinely open, protect the study's power to say no.

✕ Asking people to predict their own behaviour.

Fix: 'Would you buy this at ₹99?' collects politeness, not demand. Substitute behavioural evidence: past purchases in the category, smoke tests, pre-orders, pilot markets. Reserve asking for what asking measures well: meaning, associations, objections, language.

✕ Sampling the reachable instead of the representative.

Fix: Your followers, app users, and English-comfortable metro respondents are systematically unlike the market you're trying to grow into. Define who the decision is about first, then buy or build access to those people. An inconvenient sample beats a convenient bias.

✕ Confusing statistical significance with importance.

Fix: A 0.3% lift can be significant and worthless. A pattern in eight interviews can be unquantified and decisive. Judge findings by decision impact, whether this would change what you do, and treat significance as a supporting check, not the verdict.

Don't just read it

Practice task — 10 minutes

Take one marketing opinion your team treats as fact: 'our buyers are price-sensitive', 'young people love the new pack'. Design the cheapest possible check for it this week: five customer calls, one afternoon watching a store, or a small A/B. Run it. Whatever you find, you'll have practised the core move, converting a belief into a checkable claim before it converts itself into a budget.

If you remember five things

  • Research serves a decision. Scale its rigour to the cost of deciding wrong, and cancel it if no result would change the choice.
  • Three modes: ask for meaning, watch for behaviour, test for causation. Most real questions need two of them, triangulated.
  • Qualitative before quantitative: find what matters and the customer's words first, then measure how much and how many.
  • The say-do gap is a law: never let a stated intention answer a behavioural question when behavioural evidence is buyable.
  • Close the loop. Score research predictions against market outcomes, and your judgement compounds along with your data.