What marketing research is for
Most marketing decisions are made on a mix of experience, habit and the loudest opinion in the room. Research exists to put evidence into that mix at the moments when a wrong call is expensive: entering a market, choosing which segment to serve, setting a price, rewriting a proposition, deciding whether a campaign idea will land. It does not replace judgement. It gives judgement something to work with, and it leaves a record of why a decision was made that the team can return to when circumstances change.
The professional bodies frame the discipline in the same way. ESOMAR and the Market Research Society both describe research as the systematic gathering and interpretation of information about people and markets, carried out under codes that protect participants and the honesty of the findings. The word systematic is the one to hold on to. A conversation with a customer is useful; a conversation whose purpose was decided beforehand, whose questions were written to avoid leading, and whose notes were compared with other conversations is research.
Research also has a job the team rarely names: it tells the business which of its beliefs are wrong. The most valuable study is often the one that stops a launch, cuts a segment or kills a message the team was fond of. A research programme that only ever confirms what the sponsor hoped for is not working.
The research process
Every study, from a two-day desk exercise to a multi-country survey, passes through the same steps. The table gives the question each step answers and the guide in this hub that covers it.
| Step | The question it answers | Guide |
|---|---|---|
| Decision | What are we about to decide, and what would change our mind? | Research design and the brief |
| Research questions | What exactly do we need to know, in a form a method can answer? | Research design and the brief |
| Secondary scan | Has someone already answered this, and can we trust them? | Market and competitor research, Running a desk research pass |
| Method choice | Do we need depth and reasons, or breadth and counts, or both in sequence? | Qualitative vs quantitative research |
| Sampling and recruiting | Who do we ask, how many, and how do we avoid asking only the easy people? | Customer research, Recruiting research participants |
| Fieldwork | How do we run interviews, surveys or observation without biasing the answers? | Customer interviews, Survey design |
| Analysis and synthesis | What do the findings say, how confident are we, and what do they not say? | Customer research, Coding open-text answers |
| Decision and record | What do we now do, and where is the evidence kept for the next person? | Research design and the brief, Reporting research findings, Building a research repository |
The steps are sequential but not equal. Teams that get the first two right rarely go badly wrong afterwards; teams that skip to fieldwork usually produce a report nobody can use, because the question it answers was never agreed.
Methods: designing and choosing
The methods section covers the choices that come before any fieldwork. The first is whether to look for existing evidence (secondary research) before collecting new evidence (primary research), and the answer is nearly always yes: published reports, regulator data, trade bodies, the business's own analytics and sales records, and the research the team ran last year all cost less than a new study and often narrow the question. The second is the method family. Qualitative methods (interviews, groups, observation) explain why people do what they do and surface things the team did not know to ask about; quantitative methods (surveys, experiments, behavioural data) measure how many and how much. The comparison of qualitative and quantitative research covers when each wins and why the usual sequence is explore first, then measure.
The third choice is the design itself: sampling, timing, budget and the trade-offs between them. A study that has to report before a launch date cannot wait for a representative sample; a study that will set a multi-year strategy can. The research design guide turns these decisions into a brief, which is the document that keeps a study honest when an agency or a colleague runs it. Further guides in the section cover the work around a study: participant information and consent, whether a survey is anonymous, confidential or attributable, commissioning a research agency or freelancer, reporting research findings and building a research repository.
Two things this hub deliberately leaves to other sections. The statistical mechanics of a controlled experiment (sample sizes, significance, how to read an A/B TestControlled comparison of two variants.) belong to the CRO testing section, and research aimed specifically at why a page does not convert belongs to the conversion research methods guide. Marketing research borrows from both and does not repeat them.
Market and competitor research
The market and competitor section takes the outside-in view: how large the opportunity is, which way it is moving, and who else is competing for it. Market sizing is the question most often asked and most often answered badly, usually with a single large figure lifted from a report and divided by a hopeful percentage. The market sizing guide shows how to define total, serviceable and obtainable markets so the definitions hold, how to build a bottom-up estimate that can be argued with, and how to present a range instead of a false point.
Competitor research in this hub is the ongoing programme rather than the one-off analysis. The competitor analysis framework in the strategy section covers the single structured pass; the competitive intelligence guide covers what to watch after that pass, from ad libraries and job adverts to Share of VoiceYour share of market visibility. and search visibility, on what cadence, and where the ethical and legal limits sit. The frameworks that organise the outside-in view, PESTEL for the forces acting on the market and SWOT for the position within it, are only as good as the research that feeds them. Further guides in the section cover running a desk research pass before any fieldwork, interviewing industry experts and the trade about how a market is structured, and estimating your market share against the sizing definitions.
Customer research
The customer research section covers talking to customers directly: interviews and observation for depth, surveys and questionnaires for breadth, and the work of recruiting, incentivising and synthesising that both depend on. The customer interviews tutorial walks through a small study step by step, from the learning goal to the decision; the survey design guide covers the craft of questions, scales, order and sampling that separates a usable survey from a pile of opinions. Further guides in the section cover recruiting participants and writing a screener, building a research participant pool, turning interview findings into survey questions, drawing a survey sample from a customer list, pretesting a questionnaire, cleaning and tabulating survey responses, coding open-text answers by hand, running a focus group, running a diary study and researching business buyers.
Customer research in this hub is about evidence. What the business does with that evidence, segmenting the marketDividing a market or customer list into groups that need different offers or messages. and choosing where to compete, is the strategy section's job: the segmentation guide turns research into segments, and the audience research guide covers the lighter, continuous version of customer understanding that draws on data the business already holds. The STP framework sits across the boundary, and jobs-to-be-done is the lens most customer interviews in this hub are built around, because asking what someone was trying to accomplish produces better evidence than asking what they want.
Ethics, consent and data
Research involves people, and the professional codes exist because people can be harmed by it. The ESOMAR and MRS codes share a few principles that any team can apply without belonging to either body: participation is voluntary and informed; participants know who is collecting the data and why; research is never used as a disguise for selling; children and vulnerable people get extra protection; and findings are reported honestly, including the limits of the method. A team that follows these is already most of the way to lawful handling of personal data, but not all of the way.
Data protection law applies whenever a study records something that identifies a person, and in most markets that includes recordings, email addresses and free-text answers that name people. The general direction is clear even where the detail differs by jurisdiction: have a lawful basis for collecting the data, tell people what you will do with it, keep it no longer than needed, and separate identifying details from the answers as early as possible. Research invitations sent to customers also sit close to the line between research and marketing; an invitation that promotes a product is marketing, whatever the subject line says. Legal counsel, not a marketing guide, settles the specifics for a given organisation. The participant information and consent tutorial prepares what people are told, a consent statement for each use, withdrawal and deletion before the first invitation, and leaves the lawful basis and retention to whoever owns data protection.
A first research plan
A team that has never run a study does not need a programme. It needs one decision, one question and one small method, done properly, so the habit forms.
A meal-kit company's first study
Hearthbox, a meal-kit subscription with around 6,000 active subscribers, is deciding whether to add a low-cost "two meals a week" plan. The marketing lead writes the decision down, then the question: would a smaller plan win new subscribers, or would it mostly pull existing ones down to a cheaper tier? She spends two days on secondary work, reading competitor plan pages and the company's own churn survey, which shows that "too expensive" was the stated reason for about 35% of cancellations in the past year. She then runs ten interviews, five with recent cancellers and five with people who browsed the pricing page but never subscribed, and follows with a short survey to 1,200 lapsed customers that gets 180 responses. The interviews reveal that cancellers mostly wanted flexibility rather than a lower price; the survey finds that 62% of lapsed customers would consider returning for a plan they could pause freely, against 28% for a cheaper fixed plan. The company launches a pause feature instead of the cheap tier, and the study costs under £900 in incentives and a fortnight of one person's time.
The pattern generalises. Write the decision. Write the one or two questions that would settle it. Spend a day or two checking what is already known. Choose the smallest method that could answer the question, usually a handful of interviews followed by a short survey. Agree before fieldwork what finding would change the decision, so the result cannot be argued away afterwards. Record what was found, what it does not cover, and what was decided. Then, when the decision plays out, go back and compare. Marketing research improves the same way a FunnelModel of the customer journey stages. or a KPIA metric tied directly to an objective. dashboard improves: by being used, checked against outcomes and corrected.
Common marketing research problems
Most research failures show up as one of a few symptoms. The table maps each to the guide that diagnoses it. When a study is already in trouble, with a survey few people finish, results a reader says are not representative or a market size under challenge, the marketing research troubleshooting section works out whether design, fieldwork or reporting failed and gives each cause its fix.
| Problem | Usual cause | Guide |
|---|---|---|
| The report is finished and nobody knows what to do with it | No decision was named before the study; the questions were topics, not questions | Research design and the brief, Reporting research findings |
| The study will report after the decision it was meant to inform | No decision date in the brief, a fuller sequence than the decision needs, or recruitment started from scratch | Research taking too long |
| The survey confirmed what the sponsor already believed | Leading questions, a sample of friendly customers, or no agreed threshold for changing course | Survey design |
| A huge market size that nobody can defend | Top-down figure from a report with no bottom-up check and no stated range | Market sizing |
| Interviews produced anecdotes, not findings | No discussion guide, no synthesis step, participants chosen for convenience | Customer interviews |
| The team was surprised by a competitor's move | Competitor analysis done once and never updated | Competitive intelligence |
| The survey went out before anyone knew what to ask | Quantitative work run without a qualitative stage to find the right questions | Qualitative vs quantitative research |
| A survey went out with options written from the team's labels, not from what interviewees said | Findings left behind between the interview synthesis and the first draft of the questions | Turning interview findings into survey questions |
| Findings about why a page loses visitors do not match the analytics | The wrong family of research; conversion problems need behavioural evidence | Conversion research methods |
| Research findings never became segments or messages | Synthesis stopped at the report instead of feeding strategy | Segmentation, Audience research |
| The same question is researched again because nobody can find the last study | Findings left in decks and folders, with no shared tags and no record of the decisions they informed | Building a research repository |
