The Survey Guide - How to Design a Professional Questionnaire


Questionnaires are the core of market research. With the help of a good questionnaire, exciting insights into the minds of consumers can be generated within a short time. However, in order to construct a good questionnaire, a number of things need to be considered. Even before choosing the right questions, a goal has to be set. With this goal in mind, questions are developed and answer types are selected. The choice of appropriate question techniques is not always easy. This questionnaire guide leads you through all important steps of a questionnaire construction, points out no-go areas, and helps to find the right question format.

Want to get started right away? Speak to our experts.


First of all, the aim of the survey must be clearly defined. Do I want a questionnaire about brand awareness  or customer satisfaction? Do I want to know how my advertising slogan is perceived by 20- to 30-year-olds? Am I interested in whether there is a demand in the market for a new cosmetics line? Or do I want to find out which cover image is best perceived by readers? Based on the specific objective, suitable questions are developed, from which an entire survey can then be created. A questionnaire should not be too long. Otherwise, the participants will break off the survey, or they will get bored and choose random answers to get through it more quickly. However, it must be long enough to collect all the important information. At the beginning of the survey, it is a good idea to start with general questions and then move on to the specific ones. These questions must also meet certain quality criteria.


Questions in a questionnaire consist of two parts: The actual question and the corresponding answer format. Before deciding on a question format, it is important to consider carefully what kind of data will be obtained and whether all important conclusions can be drawn from this data.


What are possible question types?


Closed Questions

Closed questions are also known as questions with bound answer format. Their main characteristic is that several possible answers are given from which a participant can choose. The possible answers should be clearly distinguishable from each other and should not overlap in content (disjunctiveness). It is also important that the answer alternatives include all possible answers that can be given to the relevant question (exhaustiveness). In case that not all possible answers can be listed as alternatives, it is advisable to include an open text field in which the participants themselves can enter an answer if none of the given ones seem to apply to them.

It is often useful to ask so-called forced choice questions. The participant is then asked to select an answer that most likely applies. Otherwise he cannot complete the survey. This avoids missing values. For forced choice questions, an additional answer alternative called "No answer" is usually offered, so that nobody can tick something that does not apply at all. It is of high importance that this alternative is an exclusive answer meaning that no other answer option can be chosen if the participant already selected the "No answer" field. There are many types of closed questions. In the next section, we will explain them in detail.


Pros and Cons of Closed Questions

  • Quick Answers
  • Simple Analysis
  • Unambiguous Coding
  • Alternative answers determine the direction of the answers
  • Creative answers rather unlikely


Dichotomous Questions

When a question is dichotomous, there are exactly two response options from which a participant can choose. These would be, for example, "Yes" or "No", "True" or "Not True", and "True" or "False". Sometimes "I don't know" is offered as a third answer option. Dichotomous questions are suitable for quickly obtaining a rough overview of consumers' opinions. However, not every question can be formulated in a way that two alternative answers provide sufficient information. Besides, the answers are not highly differentiated.


Single Selection

Questions with one single choice, also called single choice questions, are useful if a respondent is supposed to select exactly one answer alternative. As people mostly tend to choose the first answers, it is important to randomize the answer options - i.e. to arrange them in random order for each participant. In that way, we can make sure to avoid distorted results.


Multiple Selections

For questions with multiple-choice, multiple answers can be selected from a number of pre-formulated answer alternatives. They are also known as multiple-choice questions. For the same reason as for the dichotomous questions, it also makes sense here to randomize the answer options so that there are no biases due to order effects. In addition, it is important that there are not too many possible answers to minimize the risk that respondents do not read all the given answers.


Rating scales

A rating scale can be used to measure the personal attitude of a consumer. A multi-level scale is used to indicate how much the respondent agrees with a particular statement, how much he likes something, or how important it is for him. In most cases, five or seven gradations are given. Additional levels hardly provide any additional information. The middle answer option stands for a neutral answer. The different answer levels can either be numbers (1,2,3,4,5) or described verbally (e.g. agree/rather agree/undecided/rather disagree/disagree).

Rating scales are particularly suitable if you want to find out quickly how differentiated consumers' views are about an idea or a product. The evaluation of the answers is simple and the statements are easily comparable between consumers. One disadvantage of rating scales is that not everyone interprets the gradations in the same way. It is also possible that the answers are distorted towards the middle of the scale because people tend to avoid extremes. Rating scales can be divided into unipolar and bipolar scales. Bipolar scales have a positive and a negative pole.

For example: How often do you pay with payment method X?

With unipolar scales, on the other hand, there is a zero point which marks the lowest possible level of agreement. The level of agreement increases from this point in only one direction.



Analogue Scales

With an analogue scale, respondents can indicate how much they agree with a statement or how much they like something, similar to a rating scale. The difference is that here they use a slider to specify a precise point that corresponds to their opinion. For example, one could ask how much a consumer is interested in product X and give him a scale from 0 to 100 on which he can choose any number between the two poles. The increments of the scale can be determined in advance. For example, the 100 can be reached in steps of 1, or in steps of 5 or 10 (10, 20, 30, etc.). The smaller the steps, the more differentiated the answers.



A ranking can be used to check whether there is one specific product or slogan that consumers prefer. Particpants will be asked to rank certain response options. Points are distributed depending on the position of the answer option. If slogan A is placed first by a consumer, the slogan gets 4 points. If slogan C is placed on position 2, it gets 3 points and so on. The slogan that has the most points at the end of a survey is the most popular one.


Picture Selections

Creating questions using a selection of images is particularly useful for finding out how customers like the design of a product. It could also be used to find out which advertising slogan fits a product better or how the design of an app or website is received. Image selection is useful when it is difficult to express questions or viewpoints in words.


Open-Ended Questions

Open-ended questions should be chosen when a company wants to gain information about topics they might not have been able to come up with. In general, open-ended questions are important for gathering impressions and ideas. However, it is important to ensure that not too many open questions are asked within a questionnaire so that participants do not lose their motivation to answer them in detail. Open questions are extremely suitable for text input and for entering numbers. For example, it could be asked what price consumers would be willing to pay for a specific product or service.

Open questions also include text completion tasks. If you don't need a very detailed answer, or one word is sufficient, you could simply ask participants to complete a certain sentence.


Pros and Cons of Open-Ended Questions

  • Answers are not influenced by specifications of the question
  • Creative Answers
  • Detailed Insights
  • Limited quality when respondents are not motivated/ not in a good mood
  • Difficult to compare / interpret answers
  • Capacity of participants to express themselves is varying



Atypical Questions

Heat maps

A heat map can be used when consumers are asked to answer a question by clicking on an image. For example, they could be asked what they first look at in an advertising graphic, what they find most striking on it or in which part of a photo they think the advertised product is located. A heat map can also be based on a map.


Photo / Video / Audio Recordings

For market research, it is often interesting to see how consumers live. How do they furnish their kitchen? What is the refrigerator filled with? What furniture is in the bedroom? Insights into consumers' homes can be gained easily with the help of digital market research. A questionnaire can include questions that ask consumers to take pictures of certain things or places. One can even ask them to upload a video recording. If it is interesting to hear what consumers have to say about a question or how they pronounce specific words for example, they can also be asked to provide an audio recording.

More questions? Our experts are happy to help.

Open live chat


10 No-Go's when Creating a Questionnaire

  1. Double negation
  2. Complicated sentence constructions
  3. Unclear abbreviations
  4. Technical terms
  5. Ambiguity
  6. Double questions or linked statements
  7. Moral valuations
  8. Suggestion or leading questions
  9. Missing answer categories
  10. Query unclear time periods ("recently" etc.)


You can call this via showToast(message, { variant: 'normal' | 'error' }) function