In this series of posts, we will explore the idea of using member surveys to improve service.  Being based in the Credit Union industry, we’ll be looking at the topic from the perspective of Credit Unions who want to create their own survey.  In other words, how to use a DIY approach.  

We’ll be covering the “5 W’s” of member surveys (Who, What, When, Where, Why) and as one of my programmers points out, the “H” as well (How).  Yesterday we discussed “Why” you should survey your members.  Today, I’m covering the second “W”—Who to survey. Read the full blog here (who, what, when, where, why, and how).

Who to Survey?

Be Choosy

Once you’ve identified the reason you are opening this new communication channel with your membership, you will also need to decide who to communicate with.  The easy answer is “everyone”, but that isn’t exactly practical, and even if it were, it probably wouldn’t give you the results you’re after.  Try to survey all your members and you will risk Survey Fatigue – too many surveys and members simply won’t respond anymore.  Not to mention “comment overwhelm” – if you send out too many surveys, you will get some great numbers to look at, but what about all the member comments?  Much of the value from surveys can be found in the typed in comments that accompany most survey forms.  Can you give the thoughtful consideration of what each of the 10,000 responses means to your business?  Probably not.

Sample Size Matters

So, don’t survey everyone– instead, look for a solution that will give you the actionable information you need, but also allows you to do it continually and consider each response carefully.  This requires the introduction of a popular term in the survey world: Sample Size.  Knowing your minimum required sample size won’t get you noticed at parties, but it can help you figure out how many survey responses you need to have responses that can be considered reliable.

I won’t get into the math here (because you will stop reading if I do), but basically, the required sample size is based on two numbers: what is the size of the total membership, and what is your tolerated “margin of error”.  There is a calculator here that will allow you to calculate yours (I know, I said no math!).  For example, a credit union with 50,000 members, and a “tolerated” margin of error of +/- 6%   would need 266 responses to reach 95% confidence that the scores were accurate.  If you’re trying for a lower margin of error, your required responses will go up.  Check out the calculator, it’s pretty cool.

Wrap-up

Using the calculator linked above will tell you the exact number of responses you need to get results you can trust.  If you have accurate response rate numbers, you’ll know exactly how many surveys you need to send to reach your minimum sample size.

There, was that so bad?