Headlines: Think about the kinds of headline you want to achieve, and then work backwards to list potential questions. Questions need to be credible as well as effective.
Questions Check: Once you have your draft questions ask the polling agency to sanity check them. It may be that you need extra questions asked to achieve the type of results you are looking for.
Hard to reach Audiences: The harder it gets to reach an audience the more expensive the survey will cost. If you need to reach a business audience, do you really need C-suite respondents, or will manager level do?
Short is best: Keep questionnaires as short as possible and focus on what you really want to know. Shorter surveys are quicker to do, less expensive and limit respondent fatigue.
Open ended questions: Don’t use open ended questions for regular surveys. They require extra time and cost to code the answers, and they can be be difficult to utilise for headline generation.
Consumer Classification: Use classification questions so you can identify the groups you want to be able to talk about. Age, gender and region are standard, but also consider socio economic groups.
Time Frame: Tell the agency the date you need results by and work out a timeline from there. More difficult audiences may take longer to survey.
Results format: How do you want to see the results? Tables are the standard format, but if you have any particular analysis needs, it is best to discuss them with the agency in advance. Charts can also be produced. See output formats
Media Coverage Focus: Which media do you want to reach? If you want to prioritise local or regional press you need to structure the sample accordingly e.g. target cities or regions.
Credible Headlines: Don’t try to make findings too extreme – people invariably won’t leave their wives for a Mars bar (though sometimes it’s tempting)
Forced Answers: Don’t “railroad” people and force them to answer questions, or leave out credible options. It is good practice to allow options which reflect realistic answers.
Subtlety: Don’t be too obvious and construct a survey that directly reinforces the client’s proposition
Sample size: Although 1,000 completed interviews is good for general consumer studies you don’t need that many for specialist or B2B audiences 200 will do.
Sample size must be seen as credible by journalists: The number of interviews you need is invariably determined by what journalists want rather than the number you need to achieve statistical reliability.
Quality, not quantity: The quality of the idea is invariably more important than the number of interviews
Go international: Global studies are totally feasible, allowing country comparisons and regional coverage. Multinational surveys take longer to implement and cost more than national surveys.