Step 1: Define
Clearly and succinctly define what you want to know. Write it down, to keep it front of mind.
Step 2: New information
Don’t ask customers potentially annoying questions you already know the answer to, like how long they have been your customer. Ask questions that gain new information. This will earn their respect.
Step 3: Method
Consider whether online, point of sale, postal or interview is the best survey design for your market. Then ask, do you have the means necessary to conduct your survey of choice (telephone numbers, email or postal address)?
Step 4: Format
Which question format will better generate the information you require.
- Numerical questions (rating opinion on a scale from 1 to 10) are better for repeat surveys where you track change over time;
- Non-numerical (low / medium / high) or categorical (yes / no; male / female) are best for grouping your market.
- Open Response Questions (do you have suggestions for improvement?) are best for exploring.
Step 5: Phrasing
Ensure question phrasing is clear. Don’t ask double-barrelled questions. Make sure it only covers one thing. “Was our service easy to use and access?” is clearly two questions.
The frame of reference also needs to be clear. Do you want to know about the last time they were your customer or their whole experience of being your client?
Step 6: Wording
Ensure the question wording isn’t loaded. Feed them the answer in the question and they will simply give it back to you and you will have learnt nothing. Use the same language as your customers.
Here's an anecdote to illustrate the point:
A former colleague received a promotion that required the family to move from Sydney to Brisbane. However, his school-aged son didn’t want to move from Sydney. So, as the son of a researcher, he decided to survey his friends and take the results to his dad as proof that moving to Brisbane wouldn’t be a good idea. At school the next day he asked his friends “Do you want me to move to Brisbane?” and the response was overwhelming “No”. He triumphantly took his results to his dad and was a little deflated that his father’s reply was “Go back to school and ask your friends another question; are you looking forward to visiting me in Brisbane during the holidays?”. The next day the son asked the question posed by his father and the response was overwhelmingly “Yes”.
Step 7: Categories
Make sure all possible response categories have been covered without overlapping categories or confusion may occur.
For example, you’d think ascertaining employment status would be simple, wouldn’t you? How about yours, then? Flying solo is all about choice; while survey design is about limiting choices. Which box do you tick in the ABS standard “Employment Question”?
- Employed (full time)?
- Employed (part time)?
- Unemployed seeking work?
- Not in the paid labour force?
You only want your customers to select the one most applicable response for them, right here, right now.
If unsure of all categories include “other / not applicable”.
Step 8: Respondents
Do you know whether each survey respondent will have the necessary knowledge to answer your questions? Do the questions have the same meaning to all groups?
Step 9: Review
Have you covered all the things you wanted to know? Go back and check what you wrote down in the beginning.
Step 10: Order
Order your questions to entice completion. Just like a story needs a beginning, middle and end, so does your questionnaire: easy questions to draw them in, specific questions in the middle and personal questions at the end.
“ Consider whether online, point of sale, postal or interview is the best survey design for your market. ”