… in 2009, Walmart surveyed its customers and asked “Would you like Walmart to be less cluttered?” They said yes. So Walmart cleared out space and reduced inventory and customer satisfaction shot up. However, same-store sales plummeted, by Phil Terry’s estimate, by $1.85 billion, and now Walmart has fired the team that put the idea into place and is spending hundreds of millions to undo what they spent hundreds of millions doing. But wait! Weren’t they listening to their customers? Why weren’t they rewarded?
It was a costly mistake that required lots of overhaul and refurbishing. 15% of inventory was removed from the stores…
Ben Popken. (18 Apr 2011). Walmart Declutters Aisles Per Customers’ Request, Then Loses $1.85 Billion In Sales. Consumerist® [“independent source of consumer news and information published by Consumer Reports.”] 101 Truman Avenue, Yonkers, NY 10703-1057. US.
Now, this hurt Walmart badly but it gives us great gain. Let’s look at tips to design better surveys.
- Limit total question number to 20: this aids survey completion (we have only 24 hours each day).
- Set a defined overall goal e.g. consumer behaviour / product satisfaction: I believe this sets the tone for clear and targeted questions.
- Do a trial run before the actual send out!
Micro Angle (Questions italicised)
- Restrict or avoid open ended questions: these are hard and time consuming to quantify; and may be skipped by respondents altogether
- Use neutral instead of leading or loaded questions:
- How fun was your last class? – assumes it was already enjoyable when the respondent may have found it horrid! Use something like: Rate the taste of the XYZ sandwich.
- Should concerned dog owners vaccinate their pets? => Do you think dogs require vaccination? – avoids emotional manipulation (of sorts)
- Where do you take your coffee? – since some may not take coffee at all, you need to modify based on your survey aims or remove it altogether
- Ask for only one thing per question
- How beautiful and durable are the furniture products? => (a) How beautiful are the furniture products? and (b) To what degree do you find the furniture line durable? – practical for answering and interpreting results
- Use direct and (if needed) jargon explained language
- Are you in command of languages like python, java or c? => Do you use computer programming languages (such as Python, Java, and C etc)?
- Do you like wine dry? => Do you prefer wine that is not sweet?
- You could do this another way => You like wine that is (a) not sweet, (b) slightly sweet, (c) averagely sweet, (d) quite sweet, (e) extremely sweet.
- Skip absolute questions
- Do you shave daily? => How many times do you shave each week? – much more useful and prevents respondents from quitting the survey
- Allow choices of ‘Not applicable’ or ‘Prefer not to answer’ where suitable – according to SurveyTown and SurveyMonkey this has caused respondents to quit surveys
- Provide a proper range of answers e.g. from ‘Great’ to ‘Terrible’ rather than ‘Good’ to ‘Fantastic’
- David Borman. (2018). Statistics 101. Adams Media. Avon, Massachusetts, US. p. 213.
- 10 Examples Of Biased Survey Questions. (April 12, 2016). SurveyTown. https://surveytown.com/10-examples-of-biased-survey-questions/.
- 5 common survey question mistakes that’ll ruin your data. (accessed 19 Jun 2019). https://www.surveymonkey.com/mp/5-common-survey-mistakes-ruin-your-data/.
- Writing Good Survey Questions. (accessed 19 Jun 2019). https://www.surveymonkey.com/mp/writing-survey-questions/.
- Darice. (9 Jun 2014). 5 Ways to Craft Effective Survey Questions [Infographic]
Zoho Corporation. https://www.zoho.com/survey/blog/5-ways-to-craft-effective-survey-questions-infographic.html.