Importing Data for Optional Surveys

Often companies use online surveys to gauge some aspect of the user experience after an action has been completed. An example was a recent survey taken by a SurveyMethods staff member at the end of a medical billing transaction. The individual had to pay a brief medical bill.

They entered in their invoice number, lab code, and credit card information and clicked send, and immediately there was a link above the receipt that said "please take a moment to fill out a brief survey about your recent transaction experience." The individual clicked on the link, despite the survey having no incentive, simply because they wanted to be helpful. Immediately they were taken to a new page that introduced them to the survey. They clicked "Start Survey" and were taken to another page that asked for their invoice number and lab code.

At this point they closed the survey.

The flaw with this survey design is that it forced the individual to re-input time consuming data that they had already entered on another page. The invoice number was 11 digits long, and the lab code was another complicated mix of letters and numbers, and yet even though they had entered it immediately before in the previous transaction, the survey was asking them to once again enter in these long numbers themselves. Survey design needs to be about reducing as much work on the respondent as possible.

This was adding a great deal of work – work that could have simply been completed for the individual on the backend by importing the data or providing the user with some automated code that linked them to their invoice information. Instead, it required them to do a lot of extra work. Perhaps if the survey had some sort of incentive, this would make sense.

But the survey was completely optional, designed to help the company with no benefit to the individual (most people do not need to pay for lab work often), and asking them to do a lot of work themselves in order to provide this company with that information. This is yet another example of poor survey planning and design, and a reminder that when you are creating a survey that you hope respondents will complete, you need to be as thoughtful as possible about what you are genuinely asking the respondents to do. One could easily have used the software by SurveyMethods and have had the data inputted automatically, saving the respondent considerable time and ensuring that completing the survey was not some unnecessarily difficult task.