Commissioning a survey
Good survey work starts before the first question is asked. If the brief is clear from the start, the findings are easier to trust, easier to explain and more useful when decisions need to be made.
Start with the decision, not the questionnaire
Before you commission a survey, decide what decision the research must support. A survey can test public opinion, compare audience groups, track change over time or check how people respond to a service, product or message. It should not be used as a substitute for a vague brief.
A useful starting point is to finish this sentence: “At the end of this project, we need to decide whether to...” That keeps the work focused on an action rather than a stack of charts. If several teams are involved, get agreement early on what the main research objective is, because late changes often create avoidable delays.
- Define the main decision the research must inform.
- Separate primary questions from nice-to-have questions.
- List what you already know and what still needs evidence.
- Decide whether you need a once-off survey or ongoing tracking.
For organisations working across South Africa, it also helps to decide whether the issue is national, provincial or local. A Johannesburg communications team may want a national view, while an association dealing with service delivery or regulation may need differences by province called out clearly.
Choose who must be heard
One of the most important decisions is the target population. In plain terms, whose views count for this project? The answer might be adults nationwide, registered members of an organisation, current customers, employees, business decision-makers or residents in a specific area. If this part is not defined properly, even a well-run fieldwork process can answer the wrong question.
South African research often needs careful attention to geography, language and access. Urban and rural areas can behave differently. Provinces differ in media habits, transport patterns and service conditions. In some projects, online responses may be sufficient. In others, phone interviews or face-to-face interviewing are needed to reach people who are less likely to respond online. This matters a lot if you need a balanced view across all nine provinces.
- Define exactly who is in scope and who is out of scope.
- Decide whether the study must represent national, provincial or community-level views.
- Consider language needs for respondents across regions.
- Think about whether online-only research will miss important groups.
If the audience is hard to reach, say so at the briefing stage. That allows the research plan to match the reality on the ground, rather than forcing a method that looks tidy on paper but is not a good fit in practice.
Match the method to the job
Different research questions need different methods. A quick online study may work for fast feedback on wording, awareness or message recall. Phone interviewing can help where broader reach is needed and the questionnaire is still manageable in length. Face-to-face fieldwork is often the better option where internet access, language flexibility or trust in the interview process may affect response quality.
Some briefs also benefit from a mixed approach. For example, a quantitative survey can measure patterns, while focus groups help explain why those patterns appear. The key is to avoid choosing a method simply because it is familiar or convenient. The method should suit the audience, the topic and the level of confidence you need in the findings.
Timing matters too. Public holidays, school terms, month-end pressure, travel periods and load-shedding can all affect response rates and interview quality. In coastal areas, heavy rain and humidity can also disrupt planned field schedules. If the topic is sensitive, allow time for proper testing of question wording before full rollout.
Write a brief that answers practical questions
A strong research brief saves time later. It gives the research team enough detail to recommend the right approach and flag risks early. If your organisation has internal approval steps, include those as well so that timelines are realistic from the beginning.
- What is the research objective?
- Who is the audience or population to be studied?
- Which provinces, towns or communities matter most?
- What decisions will be made from the results?
- When are findings needed and what is driving that deadline?
- Are there language requirements for interviewing or reporting?
- Will legal, policy or reputation risks require careful question review?
- Who must approve the questionnaire, fieldwork plan and final report?
If you already have draft questions, treat them as a starting point rather than a finished instrument. Good questionnaire design usually requires editing, ordering and testing. Small wording changes can affect how people understand a question, and a long survey often performs worse than a tighter one. It is better to ask fewer, clearer questions than to cram everything in and end up with weak data.
Ask about sampling, weighting and quality control
Not every client needs technical detail, but every client should ask how the study will be structured and checked. This is especially important if findings will be discussed internally by leadership, boards, policy teams or communications teams. Method notes should explain, in plain language, how respondents were selected, how fieldwork was conducted and how the data was reviewed before reporting.
In South Africa, quality control needs to take field realities seriously. Interviewer training, language matching, call-backs, supervision and back-checks all matter. For online work, screening and response verification are important. For face-to-face projects, route planning and local conditions can affect productivity and should be built into the field plan not added as an afterthought.
- How will respondents be selected?
- Will results be weighted, and if so why?
- How will interviews be monitored or checked?
- What method notes will be included in the report?
- How will sensitive or confidential client information be handled?
If a project touches on public opinion or policy issues, ask for neutral question wording and a clear statement of limits. A non-partisan research process should measure views fairly rather than push respondents toward a preferred answer.
Plan the outputs before fieldwork starts
Many survey projects run into trouble because the reporting expectations were never settled at the start. Decide early what you need at the end of the project. Some teams want a concise management summary. Others need a fuller report with charts, commentary and method notes. If the findings must be compared across regions or audience segments, say so early so that the questionnaire and reporting structure can support that.
It also helps to agree who the report is for. A board pack needs different language from an operational briefing. A policy team may need greater detail on method and wording. A marketing team may be more focused on audience differences and trends. Clarify whether you will need a presentation session, written recommendations or a clean summary for internal circulation.
Finally, decide how results will be governed. Client data should remain confidential, and internal teams should know who may access findings and who may share them. If the project is politically sensitive or public-facing, this is not a small admin point, its part of the research plan. A well-commissioned survey is not only about fieldwork. It is about asking the right people the right questions in the right way, then reporting the answers clearly and honestly.
What our Opinion Polling customers say
Verified Google reviews
Questions we get asked
How long should a questionnaire be?
Long enough to answer the brief properly, but short enough to keep respondents engaged. If a questionnaire is overloaded, response quality can drop. It is usually better to prioritise the most important questions and remove repetition.
Should we choose online, phone or face-to-face interviewing?
That depends on who you need to reach and how important representation is across different groups and places. Online can be fast, phone can extend reach, and face-to-face can be the best option where access, language or trust may affect participation.
What should we prepare before sending an enquiry?
Prepare a short brief covering your objective, target audience, geography, timing, likely stakeholders and any draft topics or questions. It also helps to note whether the work is confidential and whether findings will inform public communication, policy or internal decisions.
Also worth a look
Tell us about the job
Send sizes, materials and where you are through our contact form, and we will come back with the practical options.
Open the contact form