Preparation:
Before we start building a survey we will create the outline with our customers first. This enables the project team to think carefully about each question any misinterpretations that may occur and what responses are likely to come back. Designing the structure of any survey is key to its usability, especially if we are using features like conditional questioning or branching.
Structure into logical bite size chunks:
Giving structure to the user is important, always open with a welcome page which instructs the user to the aim of the survey and instructions on its use. Breaking up your survey into sensible well-ordered pages helps group specific questions sets together, making the survey more user friendly and less likely to put off respondents.
Survey length:
The longer your survey the less likely respondents will want to complete it. Ask the important questions that you really want answers to, and remove questions that are less likely to make an impact on your actions following the results
Interactive questioning:
Conditional logic/branching is when you ask a question or a set of questions, following the response to a previous question. For example, if you ask “do you own a car” and the respondent answers “Yes” you can move on and ask further questions like “what colour is your car?” if your respondent answers “No” in response to the question it’s unnecessary to display the car colour question
Design and feedback:
When we design surveys using brand colours and logos are far more effective than cluttering the page with lots of imagery which will detract from the content on the page. Clear headers, easy to read fonts and as previously mentioned a well structured layout are all important.
Remember your respondents are often not obliged to complete your survey. We recommend a clear and concise welcome page at the start of a survey that is brief, and explains the reasons for the survey and, possibly, where results will become available. If personal information is being gathered, give peace of mind as to your privacy policy. Include a thank-you message and if you are gathering email information for future communication, put a short opt-in to ensure they are prepared to take part.
01273 669 710
In terms of survey lengths how many questions would be considered too many and turn people off to perhaps starting or finishing a survey? Can you recommend an optimum number or give an indication as guidance that could improve response rates?
Hi Jane, Good question! Getting the length of a questionnaire right is quite challenging as you have to balance a number of conflicting requirements. There is always pressure to add more questions in order to get a more comprehensive response but if a questionnaire is too long then respondents will exit before the end is reached – and this results in distorted or incomplete response data in the survey database.
The number of questions a respondent will tolerate varies depending on the type of survey that is being undertaken and the relationship between the respondent and the organisation conducting the survey. Internal surveys, such as staff-satisfaction surveys, or employee attitude surveys, tend to have a large number of questions as the respondents are ‘hostages’ and also typically they are motivated to give feedback to their employer. For such surveys 60 questions can be satisfactory.
Surveys where the respondent has no obligation to the survey organiser (e.g. customer feedback surveys and marketing surveys) must have many fewer questions unless there is a tangible reward (such as a prize draw) at the end of the questionnaire. In this type of survey the motivation of the respondent is much less and 20 questions is a good number to aim for. It is hard to make firm statements but, unless brand loyalty is very strong, go for a lower number.
For telephone surveys and interview surveys the number of questions will again be small – particularly if the respondents are to be ‘captured’ by surprise in a public place.
So, in summary, the number of questions in a survey depends on a number of factors that are determined primarily by the relationship of the survey organiser with the respondent.
Hi, I appreciate encouraging staff to participate in surveys is task in itself and a lot seem to have the mind set that its not worth them completing a survey as they won’t see any results or changes from it. I am in the planning stages of a survey but I would like to be able to share the results with the staff so they can see in real time feedback from their input. I dont mean a detailed break down but some graphs / charts / percentages something real in exchange for their time, it that possible?
Hi Michelle, thank you for your post.
We are regularly asked to provide on-line surveys that automatically give respondents some post-survey real-time feedback. This is an area that is not considered very much and I though that I would write about what we are doing. Real-time on-line analysis and feedback can add huge value to an on-line survey. It can give you the ability to offer a big juicy carrot to engage respondents who might not normally see any value in spending their time on your survey.
1 What do we mean by real-time analysis
You have completed an on-line questionnaire and get to the end. Before you exit you are given the opportunity to see feedback on your responses. The feedback is presented normally in a format that includes charts, graphics and text and there can be options to print and download as a pdf.
We have found, increasingly, that survey sponsors want to be able to give something back to respondents – to thank them for their co-operation. This is a much healthier way of rewarding participation than prizes and rewards and is much more likely to encourage accurate responses. Survey feedback can be an analysis of an individual’s responses or it can be of the survey as a whole. It can also be used to educate, inform warn and it can communicate messages in a way that is very specific to the individual respondent.
2 But how is it done…
Our industrial-strength survey system has a very sophisticated database behind it and we are able to detect survey events, such as someone clicking on the finish button, and to automatically execute processes based on that event. A typical requirement is to automatically analyse the respondents responses to the questionnaire and then present the analysis to them as part of the same survey dialogue. We have recently done just this for a Government agency that wanted their respondents to receive informative feedback and guidance after they had completed an on-line assessment.
3 Is it complicated to do?
No, it is all quite easy. We expect survey sponsors to be able to provide us with the rules on which the analysis is to be based and the format that they would like the feedback to be in. We have many output formats, from bar-charts to spider diagrams, and we can help choose which is most appropriate for the project. We will normally present specific textual elements to supplement the graphical output and we can deliver video and audio as well.
4 What does the feedback look like
Feedback will often commence with an explanation – putting the survey feedback in context and providing some guidance as to its scope and usage. Sometimes liability statements also need to be presented – particularly if the topic of the survey (or the feedback) has been something that will influence the subsequent behaviour of the respondent. This all goes into the introduction. There might also be contact information or links to a help facility.
The feedback itself will normally include histograms and textual explanations and may consist of many pages. There will normally also be some end messages prior to exit.
5 Sounds Expensive
Thankfully, for our system this is not the case. We have an on-line framework which we customise for each project. As it is already live and tested the customisation is quite quick and is actually exceptional value for money.
Hi johna,
Ive just been reading your last comment on how your surveys are customisable and I had a question I was hoping you could answer? We have many different roles within our organisation from admininstrators to scientists and I want make our survey accessible to everyone. Is it more practical for a survey to be a “one size fits all” to keep costs down and set up quicker or should it vary depending on the audience and be customised to keep the content more relevant to a specific audience? A lot of questions won’t be applicable to certain staff but should these questions be left in anyway with a note at the beginning advising who should answer what so that the survey is transparent to all?
Any advice you could give would be great. Kind regards