Write the research problem as one testable sentence before contacting suppliers. “Will people like this campaign?” leaves too much room for vague answers. “Which message makes first-time donors more likely to request information?” gives the team a defined outcome. Ask each provider to show how that question would become a sample plan, questionnaire, and analysis approach. A sample is the portion of an audience selected for study. It does not need to include everyone, but it should reflect the decision. A youth charity, for example, may need young people and parents rather than only current supporters.
Let the question determine whether the project needs quantitative or qualitative research. Surveys collect structured responses that can reveal how preferences vary across a defined group. Interviews, focus groups, and open-ended exercises can uncover the reasons beneath those responses. Suppose a retailer learns that customers reject a refill service. Survey results may show how widespread the objection is, while interviews may reveal worries about storage, hygiene, or convenience. If both breadth and explanation matter, ask whether the agency can connect the methods. A mixed project should have one logic, not two disconnected reports placed in the same slide deck.
Ask how the team will define and compare audience segments. Useful groups might include regular donors, occasional donors, lapsed supporters, and people who have never given. Labels alone do not make a segment meaningful. The provider should explain which behavior, need, or attitude separates one group from another and how that distinction will affect the message, channel, or offer. A campaign that succeeds with loyal supporters may fail with new audiences. Reviewing results by segment can prevent a team from treating a strong overall average as proof that everyone responded similarly.
Read the questionnaire before fieldwork begins. Look for leading wording that nudges people toward a preferred answer and double-barreled questions that ask whether something is both affordable and convenient. Check the response options for missing choices, inspect the question order, and ask whether a small pilot will be run. A pilot can expose confusing terms before they reach the full sample. Also ask how the provider will identify duplicate entries, incomplete submissions, and respondents who move unusually quickly. A final questionnaire and change log make later interpretation easier when wording changes during the project.
Reporting arrangements deserve attention before a contract is signed. Ask whether the analysis will show subgroup sizes, explain any weighting, and describe limitations plainly. A confidence interval expresses uncertainty around an estimate, so readers do not mistake a survey percentage for a perfectly exact measurement. The agency should explain whether its conclusions fit the sample and design. Request a report that connects evidence to an action, such as testing one message, prioritizing an audience, or conducting interviews on an unresolved issue. Charts without interpretation can leave executives and creative teams drawing different conclusions from the same result.
Compare providers by fit, transparency, and working habits rather than reputation alone. Confirm who will conduct interviews, who owns the raw data, how participant privacy will be managed, and what happens if the brief changes after recruitment. Ask for a realistic timetable covering questionnaire review, fieldwork, cleaning, analysis, and a findings session. The team should explain results in plain English to leaders, creatives, and community partners. If the work will shape public communications, look for public communications guidance that keeps evidence separate from promotion. The right partner can answer the specific question and make the result usable.
