During your beta test, you want to get as much useful feedback out of your testers as you can. But each tester only has a limited amount of time and energy to devote to your test. Surveys, if used properly, are a great way to get the most out of your testers and their time. Here are some quick tips to help make your surveys awesome and get the data you need.
1. Plan Your Survey
Treat every survey like a small project. This means determining what information you are trying to get, how this survey relates to the goals of your beta test, who the target population of the survey is, what you plan to do with the data, and what the questions and acceptable responses are. Finally, you’ll need to run it by all the appropriate parties in your company. Draft the survey first in Word, so that you can re-work it, pass it around to other stakeholders, and then put it into your beta software.
2. Use Consistent Grammar, Punctuation, and Capitalization
You want your testers to take your beta test seriously. Set a good example by giving your survey questions a professional appearance. Use consistent and accurate capitalization and punctuation throughout the entire survey. And of course, always check your spelling.
3. Keep Questions Short
Short questions are much easier for your testers to read, understand, and respond to. From a project management standpoint, shorter questions are easier to look at in views, filters, and reports. If your questions are longer than one line, consider rewording them or splitting them into a couple different questions.
4. Keep Words Simple
Don’t try to impress your testers with big words. This is especially important with international tests, where beta testers have vastly different levels of comprehension. If your testers are confused, it will skew your results. Added benefit: simple words often also mean shorter words.
5. Keep the Survey Quick and Focused
Remember that in most scenarios, testers are volunteering their time and energy. Respect that. Generally, 10 questions is a good survey, 15 is long but acceptable, and 20+ is only really appropriate toward the end of a beta test (since you won’t be asking for much more afterward). If you plan to survey your testers more than once a week, keep them to around 5 questions each. If you cannot reduce questions, split them into multiple small surveys, spread throughout the week. Also (see #6), questions that require a tester to write out a response take more energy to respond to. As a good rule of thumb, we consider one free-form response as equivalent to three multiple choice questions.
6. Use Selections When Possible
Before you add any question to a survey, spend a couple minutes thinking about how it relates to the goals of your beta test and what you want to do with the results. This will determine how you collect the data. Selection responses (drop-down boxes, checkboxes, radio buttons, etc.) are often a much better choice than free-form (text) responses in terms of tracking and reporting. Also, selection questions are much easier for testers to fill out, which means you can fit more questions into the survey before they get tired and move on to something else.
7. Preview Your Survey
Always make sure to review your surveys as you create them. You want to make sure you know what your testers are going to see when they fill out and the survey, and that it is the tester experience you want.
8. Limit Access
Give survey access only to those who need to fill out that particular survey. There is no sense in cluttering up the experience for those who are not supposed to complete a survey, and their responses could skew your data.
9. Use Rating Scales of 5 (Not 10)
Although common, there is no reason rating scales need to be 1 to 10. Rating scales with 5 points are much easier for both testers and your team. A 5-point rating scale allows room for very strong feelings (1 and 5), general good or bad feelings (2 and 4), as well as general indifference (3). This makes selecting choices more natural and obvious, while also making reporting easier and cleaner.
10. Label Rating Scale Values
Rating scales are useful in nearly every survey. Unfortunately, many surveys have unmarked values (such as 1 2 3 4 5) which can be interpreted differently by every tester. By giving labels to at least the first and last values (such as 1=Very Poor, 5=Excellent), testers are given a clearer picture of what the values are intended to represent.
Screenshot courtesy of Centercode