aharish.xyz

Creating A Survey Questionnaire

4 min read

The central and most crucial aspect of surveys is the questions you ask. These questions need to be framed and arranged with care so that you get the most reliable and accurate data. There are also certain things you need to keep in mind while conducting a survey.

Types of questions

  1. Categorical Questions - looking for a simple count using yes or no questions. You can also use MCQs. You can also use checkboxes where they can select more than one
  2. Ordinal Questions - when questions have a clear order. For example, age dropdowns. You can also use rank. For example - how would you rank the following: juice, tea, milk
  3. Interval/ratio questions - Allows you to conduct analysis to find average, testing corelations, Likert scale, regression models and so on
  4. Ranking Scale - On a scale of 1 to 5 how would you rate your ticket booking experience
  5. Matrix: multiple rating
  6. Textbox: For personal questions like - how many apps do you have on your phone.

Do’s and Don’ts for Surveys

  • Use simple direct language - avoid jargon
  • Be specific - things can mean different things to different people
  • Break down big topics
  • Avoid leading questions - “Do you think Indians who prefer paying by cash use an online wallet?” instead of “Would Indians use an online wallet?”
  • Avoid complex questions - “Do you drink milk and juice on a regular basis?” - split it into two
  • Use interval questions - scale of 5 or 7 with labels for each like “Strongly agree, agree, Strongly disagree”
  • Avoid biases - asking for gender, race, or income will lead to influence
  • Provide complete options - when there is no neutral option, they tend to get annoyed. The options should be mutually exclusive and collectively exhaustive other wise they will be confused.

Important

Surveys like NPS are only conducted for the most valued customers where as customer satisfaction surveys are for anyone that conducts a transaction on the product

Cohorts

Cohorts are based on user behaviour. On Practo, for example, anyone that comes and searches for doctors is one type of user. But if he also ends up booking an appointment, then he becomes a slightly more valuable user to you since he has done a transaction with you, so he falls into a different bucket.

By building these buckets, you can ensure you are targetting these users through different types of surveys. For example, anyone that has created a transaction can be sent a customer satisfaction survey. But for someone who do multiple things, then these people become valuable customers and for them you would want to do a NPS every 90 days.

The types of questions you can ask in a survey are:

  1. Categorical questions: These include yes/no questions and MCQs
  2. Ordinal questions: Such questions are used for elements that have a proper order, and you can get counts or percentages to define this order
  3. Interval/Ratio questions: It can be the ‘Likert scale’ (‘strongly agree’ to ‘strongly disagree’) or the ‘ranking scale’ 
  4. Matrix questions: It’s a grid where multiple items can be rated on one scale
  5. Textbox questions: This can be used to gather respondent-specific data

The essential points you need to bear in mind while conducting a survey are:

  • Use simple and direct language
  • Ask specific questions
  • Break down larger topics into smaller units
  • Don’t ask leading questions, do not ask questions that are based on your opinions. 
  • Avoid asking complex questions
  • Avoid asking biased questions
  • Provide complete options

To summarise, conducting a survey is a 5-step process:

  1. You identify your target audience. Who should be asked to fill up the survey?
  2. Keeping your audience and objectives in mind, you design the survey. The questions must be designed in such a way that the results of the survey yield actionable results for the brand manager. 
  3. The third step is to distribute the survey - online or offline. Your distribution strategy would also be defined by the required reach and the channels most relevant to the audience whom you want to reach. 
  4. The fourth step is about getting responses. Generally, incentivising the audience to fill out the surveys and allowing multiple follow-ups can help get the desired sample size. 
  5. The final step in the process of quantitative research is data analysis. Your survey data will give you numbers and percentages. They must be analysed to do 2 things:
    1. Validate or reject the hypotheses
    2. And be capable of being translated into actionable for the brand.

Graph View

Backlinks