Dial Down The Endless Surveys And Really Listen To Customers

Business leaders’ concern about customer opinions is a good thing. But the way many companies collect data does more harm than good. Keep two principles in mind: bad data does not foster good decisions, and customers want you to listen.

Businesses of all size seem to be obsessed with surveying their customers. For example, every time I fly on United, and every time I log out of my Bank of America HSA, I am asked to complete a web survey. A lot of Amazon vendors (usually small businesses) send a survey after every purchase, too.

I think this is this is driven partly by the popularity of Net Promoter Score (NPS). NPS has become a phenomenon. Companies can easily compare their NPS score to leading brands such as Apple. Consultants argue NPS correlates with revenue growth. Tracking NPS requires frequent customer surveys.

My portfolio companies all measure NPS now and put the data in their board of directors materials and investor pitches. Simple and useful ideas like NPS often wash through the business world, doing a lot of good, but also some harm, because the devil lurks in the details.

There’s no way a typical customer can respond to all the surveys to which s/he is subjected currently. They take too much time and they interrupt the user’s flow. I expect that the typical response rate is low, perhaps single digits. And, many of these surveys are not mobile-friendly.

Who is responding? No doubt a typical customer occasionally responds when s/he is not busy. Probably a high proportion of responders are outlier customers: older people with more free time, or customers with an agenda such as frustration with a recent service experience. Millennials and other mobile-centric groups are probably under-represented.

Data as described above is tricky. If you gather and track this data regularly and see a spike in complaints, you should look for a problem in your product or service. But, this data probably does not tell you what your core customers are thinking or how loyal they are to your brand. Non-representative response is a big problem: core customers’ voices can be drowned out by the outliers who are over-represented in the response. Statistical tools can mitigate this problem, but they require significant background data on each survey respondent, which is usually not part of the survey. Big companies probably have this data and may be able to connect it to the survey response and use it to improve the data. Small companies often lack the skills to perform this level of analysis.

And, there is no “average” customer, because everyone is unique in some way, and the customers who make up the core of your franchise can fall into two or more distinct groups. One of my portfolio companies performed customer research for a high-fiber cereal brand. They learned quickly that this brand has two very different core customer groups: young, mostly-female adults who care about controlling weight, and older adults concerned about intestinal health. The average customer for this brand, a forty-something androgynous person, is not a helpful construct.

Companies that want to survey their customers after every touch point often don’t want to listen to customers. Suppose you are a typical and good-value customer for United Airlines or Bank of America, something goes badly wrong, and you want to call them up, talk to a person with a reasonable level of knowledge and authority, and discuss the problem. I searched the United web site for the word “complaint”, hoping to find a way to discuss a complaint with a United staffer. The best thing I could find was a web form to fill out for customer service, amidst many documents declaring United’s operational policies. If you call Bank of America, you will spend a long time keying in identifying numbers and listening to recordings that steer you to a robotic response, through two or three tiers of menus, before you get the opportunity to wait on hold to talk with a junior customer service rep. That does not feel like “listening to customers” to me. Small businesses are often better at this, because they lack the technology to fend off customer calls. And if you are a very high status (“1K” or “global service”) United customer, you can call a passenger service desk staffed by old pros who do a good job with flight and reservation issues.

When a customer is unhappy, they usually want to talk with someone who can make them feel heard and respected and help with the problem to some degree. I recently heard a friend call housekeeping at a resort about a problem. The person at the other end listened politely and offered a step to mitigate the problem. After a few minutes, my friend said: “That [the mitigation] will be fine. I appreciate the chance to tell you how I feel about this. That’s all I need.” My friend felt fairly treated and will be back.

Web surveys are appealing because they are “scalable”: they are done with software that collects data and boils it down to a dataset executives can analyze. This produces good PowerPoint backed by “real data”. The data may be real, but if it is unrepresentative, the results are suspect: “Garbage In –> Garbage Out”.

Keep these two things in mind as you work on better rapport with your customers. Pay close attention to the quality of the data you collect and analyze. There are many pitfalls in customer survey data. Both professional help and caution interpreting data are often warranted. And still there can be surprises, as happened to the best pollsters with the 2016 election.

Listen to customers the old-fashioned way. This is expensive, non-scalable, and often tedious, as there will always be cranks and bullies. However, in business, as in most other parts of life, if you want to have strong relationships, there is no substitute for listening.

First posted @ blogs.forbes.com/toddhixon on March 17, 2017.

Show your support

Clapping shows how much you appreciated Todd Hixon’s story.