Now that I have your attention (presumably), the first words of wisdom I have on the subject are that the perfect B2B survey is unlikely ever to be written. But by following very simple “best practices” in creating surveys, their effectiveness and value can be maximized – and the impact of preconceived notions that might skew the results can be minimized.

The heart and soul of any survey, and therefore the area in which preconceived notions are most harmful, is the style in which survey questions are worded. Neutrality can be surprisingly difficult to achieve in a survey question, even if the researcher isn’t hoping for a certain result and allowing that hope to creep into the question.

The classic example of a preconceived notion affecting a survey is the leading question. For example, if a manufacturer redesigns their website with the objective of making it easier for distributors to use, they may perform a survey to see how well their efforts worked. Too often, the desired outcome creeps into a survey question such as:

• We recently redesigned our website to add functionality and make it easier and faster to use. What are your thoughts on the new site?

Here we see a question that would be somewhat difficult for a respondent to answer in the negative, regardless of their actual user experience. Which is, perhaps unintentionally, what the question writer is hoping for. After all, of course they’re going to like the new site! It has to be better! We’ve added functionality!

Much more effective would be something like:

• What has been your experience using our new website since it was redesigned?

This style, more specific to user experience and more objective, is much more likely to produce useful results -- particularly if there is also a question about “How many times have you accessed our website since it was redesigned on (date)?”

Avoiding leading questions, and building in objectivity, might well be the most important way to ensure useful results in a B2B survey. But there are a number of other best practices to follow in achieving this goal, including

• Use simple, non-jargon words and phrases whenever possible
• Be consistent in the use of all technical and official terms, avoiding switching back and forth between, say, “human-machine interface” and “machine display terminal”
• Avoid assuming that respondents are familiar with all technical terms
• In responses, be cautious about using absolute words and phrases such as “never,” “always,” “unsuccessful,” etc., especially when they express negative concepts -- limit them to objective concepts (“How often do you use Function X on product Y”) rather than subjective (“how often does a Function X make you want to buy a product Y”)
• Assign one question to each concept you want to have respondents evaluate. For example, if you want to assess the value of an instrument panel as well as the LCD display in the middle of it, create one question for each item
• Make answer choices appear in random order, so that the “preferred” answer appears in a different spot on the list each time. Some professional survey organizations such as Survey Monkey offer a “Randomized Answers” option to their subscribers

For additional tips on creating survey questions, check out this guide to better surveys.

All in all, creating the right survey questions for objectivity and maximum value is part science and part art. In my next post, I’ll talk about some ways to combine the two.