Savvy B2B Marketing

Good questions! Ah, I could write forever about this (but alas, with only a few hundred words, one must be judicious when offering tips!)

1. Usable format

Think a moment, for example, about pie charts. They are, by design supposed to add to 100%. I have seen several surveys which have been fielded only to discover the answers "don't know" or "not applicable" are left off the roster of options. This makes your pie chart incomplete, and even more troubling, the data collected, incorrect! Always give respondents an option to tell you the truth -- don't force them into giving a wrong answer just to move forward.

Here's another. Have you ever needed to calculate an average price? You can either ask for a dollar amount (fill in the blank) or you can give ranges that respondents can choose from. In the first case, you can set the format on the software to add in decimal points...but you should always set a reasonable cap, otherwise you can wind up with people who "accidentally" type in too many digits. In the second scenario, you need to make sure your ranges are appropriate for what you are measuring. Then you can use the midpoint of the range to calculate the average.

So for example:

Q. How many widgets do you anticipate purchasing this year?
A. none
B. one
C. two to four
D. five to eight
E. nine or more

You can easily calculate an average of widgets that will be purchased by multiplying the number of respondents by the midpoint of the range. To play this out.... leaving out the "none" , B=1, C= 3, D=7, E=10. Your understanding of how often your product is purchased helps you set that upper range of the last option.

Another area where "bad data" creeps into the picture is in the ratings area. A scale from 1-5 or 1-7 is great to use, but you need to be consistent and clear when you decide to present the options. 1 should always be low or negative and 7 should be high or positive throughout the whole survey. Why? People get trained quickly as to where to mark a good rating vs a bad one. You may think you're being clever by switching the scales but that can cause respondent confusion.

Finally, setting a reasonable time frame for memory recall is crucial. You should aim to set a time frame that is recent and appropriate for the product you are measuring. In the media world, when we ask people to recall what they've seen, watched or listened to, we talk about “yesterday”. With grocery shopping, it’s the past 7 days. In B2B, the standard might be “this past quarter”, “within 6 months” or “this past fiscal year.” Start with the smallest unit you can measure…you can always multiply and extrapolate to estimate larger numbers.

Response rates:

Once, I was at a seminar given by a company that claimed that they always get 100% response rates when they conduct their B2B surveys. We attendees all shook our heads and compared notes. No one in the room ever got 100% response rates.

When we pressed them further, we found that this company prescreened their samples by phone, offered large incentives and scheduled time to take the survey on the internet with the respondents. Oh, and they were only talking with about 25 people for any given survey.

If you can afford to hold hands like that then maybe you, too, will get that kind of response rate!

The rest of us acheived anywhere from 2% to 25%. While it is easy to generalize, the truth is that there are a number of variables that affect response rate.
- Source of respondent list - (Are they your customers, readers of a trade magazine, names purchased from D&B or Hoovers?) If you are well-known to your respondents, then your response rate is traditionally higher than if they do not know who you are.
- The methodology – phone, mail, intercept or internet survey – each has its own strengths and weaknesses. We have found that soliciting by mail at least 3 times (using good, clever creative) can increase the response rate to internet surveys dramatically.
- The survey instrument – Respondent experience is a big deal! If the survey is too long, if it isn’t pleasing to the eye, if it isn’t well-written, you may have good response, but bad “completion” rates. Completion is the percentage of people who make it through your screens (the questions that qualify them to take your survey) and finish the survey to the end. Engaging the respondent is key to getting a high completion rate (90%+).
- Incentive – What you give away is really important. Just an entry to a $500 sweepstakes doesn’t work in B2B surveys. The more appropriate the incentive to the industry, the more likely you’ll get people to give you their time to take your survey. We’ve given away pens, shop aprons, mugs and had a sweepstakes for an expensive commercial printer. These incentives created buzz around the survey itself within the industry. And the promise of sharing top line information with respondents let them learn more about how they measured up compared to their peers.

Whew!

Ask a straightforward question and see what happens?

In my opinion, survey writing, deployment and analysis are real skills that constantly need to be honed as new methodologies come into play. I have heard it said that traditional market research will be supplanted by real transactional data in the coming years. While internet sites can offer insight into behaviors (measuring clicks), the resulting data aren’t always representative or projectable. For that reason, sampling, questioning and interpreting will always be in style for companies that seek customer insight.

Happy crunching!

Wendy

PS My favorite quote is from the master himself, Einstein. He said: "If we knew what we were doing, it wouldn't be called research." Research is a fascinating business because we are always learning, always exploring, always pushing to find connections.