I recently received an email inviting me to fill out a customer satisfaction survey from one of my favorite online outdoor equipment retailers. I freely admit that I only opened it out of competitive curiosity. When I opened the email, and it promised to take only two minutes and ask two simple questions, I exclaimed, ‘An NPS survey! ROCK ON!’
But that was me speaking as the survey taker. As a data scientist and survey analyst, I recognized the shortcomings of an NPS-style survey. There’s no way to control the data points in the sample, so this equipment retailer wouldn’t be getting the data required to slice and dice my results in interesting and meaningful ways.
For example, they wouldn’t be able to narrow results down to a particular type of service, a particular shift, or a distinct product. They wouldn’t even be able to do a basic regression analysis to identify key drivers. On the other hand, is it really worth it to swing the opposite way and create an experience-killing 47-question survey? Probably not.
Which brings up one of the major conundrums faced by VoC practitioners today:
There is a fundamental disconnect between the way customers want to share their experiences and the way researchers want to control the scientific sampling of information. The result is friction in the feedback process.
No More “Us Vs. Them”
Consumers want fewer questions. Companies want robust analysis. The smart companies are doing what the consumer wants—but the even smarter companies are finding ways to satisfy the interests of both parties.
Today, there are ways to begin cutting the length and overall friction from your satisfaction surveys today, without sacrificing your useful back-end insights. Here are three things I would recommend to any forward-thinking customer experience practitioner trying to find the fabled frictionless feedback:
1. Use the Power of Text Analytics
Shorter surveys don’t have to mean smaller data sets, as long as you’re asking the right questions and taking advantage of powerful natural language processing engines. Mindshare Monitor™ and Mindshare Discover™ are both supercomputing solutions that can extract profound structured meaning from free-form text.
Learn all you can about using this powerful technology, experiment with it, and find out which of your survey questions have become redundant.
2. Separate Your Feedback Channels
Just because you’re favoring shorter surveys doesn’t mean you have to completely give up on longer market research-style surveys. You can dedicate separate feedback channels to administering two separate surveys simultaneously:
1. A short review-style survey that any customer can comfortably complete.
2. A longer research-style survey providing granular details for analysing new product introductions or marketing research.
Your goal should be to offer your customers a good feedback experience while still collecting enough data to power other types of analytics.
3. Shorten Surveys
Identify which data points are the most important to your goals—and ruthlessly eliminate the rest. Not only will this provide a better experience to your customers, it will drastically simplify analysis. If you’re collecting more data points than you need, you might be creating more questions than you are answers. If you can’t arrive at a decisive action after asking yourself a “so what” question, consider eliminating the metric.
For example, a restaurant survey may have a question inviting the customer to “rate our tabletop displays.” You learn you have an average rating of 73 out of 100. So what? Well, it means your displays should be better. So what? Well, a 10% increase in those ratings should mean people learned more about your product. So what? You get the picture.
Think Forward and Implement Now
Customer tastes and preferences around survey-taking behavior are changing. The forward-thinking customer experience practitioner will see this as an opportunity to start implementing the most advanced review-oriented survey techniques to provide a great frictionless feedback experience to customers—while collecting an even broader data set.