An agency may provide research design, participant recruitment, analysis and presentation, although the exact service mix varies. When comparing a uk market research agency, ask who will write the questionnaire, check the sample and interpret the results. Quantitative research collects structured answers that can be counted, such as purchase intent or awareness. Qualitative research uses interviews or moderated groups to explore language, motives and objections. Neither method automatically replaces the other. A survey can show that a problem is widespread, while interviews may reveal that the wording or process causing it is misunderstood.
The sample is the group invited to take part, and representativeness means it resembles the wider audience you want to understand. A sample of existing customers may be useful for improving retention, but it cannot automatically describe potential customers who have never bought from you. Recruitment should therefore reflect relevant characteristics such as age, region, usage or organisation size, where those characteristics affect the decision. An agency should also explain exclusions. For example, researching a national charity campaign only among regular supporters could overstate familiarity and miss the concerns of occasional donors or people who have not engaged before.
Question design deserves close attention because small wording choices can influence answers. A leading question suggests the response the researcher expects, while a double-barrelled question asks about two issues at once, such as price and convenience, without allowing separate answers. Segmentation means dividing respondents into meaningful groups so differences can be examined rather than hidden in an average. For a charity testing a new appeal, supporters, volunteers and lapsed donors might need separate analysis. The trade-off is cost: more segments require more participants and careful interpretation, or the apparent differences may be too thin to trust.
Ask how uncertainty and bias will be handled. A confidence interval is a calculated range showing how much a survey estimate may vary because only a sample was questioned; it is not a guarantee that the true result sits inside that range. Weighting adjusts the influence of responses when the achieved sample differs from the intended audience, but it cannot repair every recruitment problem. Response bias occurs when those who answer differ systematically from those who do not. A credible report should state fieldwork dates, sample source, exclusions, weighting and question wording rather than presenting a single percentage without context.
The final deliverable should make decisions easier, not merely archive research. Request a clear distinction between observed findings, interpretation and recommendation. Charts should show the base size, while quotations from interviews should be labelled as examples rather than treated as proof of scale. If the work supports public messaging, the research team and communications team should agree how findings can be used without overstating them. Practical charity communications insight may be useful when a sensitive issue needs clear public explanation, but research evidence still needs to remain separate from campaign promotion.
Price should be assessed alongside method, access and usability. A low-cost survey may suit a narrow operational question, whereas a high-stakes repositioning decision could justify interviews, testing and a larger sample. Ask for a staged proposal if the full programme feels disproportionate: an exploratory phase can refine the language before a broader survey. Check ownership of raw data, privacy responsibilities, respondent consent and the timetable for revisions. Most importantly, agree in advance what finding would change the plan. If no possible result could alter the decision, research may be offering reassurance rather than useful evidence.
