In May 2019, The Wall Street Journal published a comprehensive article on the use – and misuse –of customer satisfaction programs that heavily rely on a single measure known as the “Net Promoter Score” or NPS*. Titled “The Dubious Management Fad Sweeping Corporate America”, the authors shared a review of 40,000 earnings call transcripts for 688 companies that are or were in the S&P 500 from 2003-2018. Last year (2018) the words “net promoter” or “NPS” were cited more than 150 times in conference calls by 50 S&P 500 companies. That’s four times as many mentions, and three times as many companies, from five years earlier. Last year, “net promoter score” was mentioned in 56 proxy filings. Some companies, including American Express, Best Buy, and Citigroup list NPS as a compensation metric, alongside share price and EPS. So what is really going on here?

 

Misuse of NPS

 

Regrettably, the Net Promoter Score or NPS has morphed into something well beyond its scope. Originally intended to foster product and service improvement, many corporate leaders now treat NPS as a pseudo-financial metric. Management seems to believe that they must follow peer group practices, including the use of benchmarks collected by their own competitors. Misinformed stock analysts add pressure by implying that companies who do not use NPS must be out of touch with their customers.

 

The companies that continue to use NPS as a mantra are apparently oblivious to its corrosive effects. Much like “teaching to the test”, they ignore the fact that the measure can be manipulated. For example, a recent trip to my local Honda service center produced a text message from the Service Manager (see inset). He implored me to give him high ratings. What larger goal does this accomplish? Knowing that a survey is coming, management is unwittingly diverting his time and creating anxiety – and missing valuable insight into my service experience. And what if the initial technician who took my car had an attitude? Do I tag the Service Manager with that? A far better approach would have been to ask: “What aspects of your visit to Honda need to be changed or improved?” Or, more specifically, probe on multiple dimensions of my visit so that the employee is not associated with my evaluation of the entire experience.

Of course, the error can go in the opposite direction: companies can collect feedback that is too narrow and not actionable. Another example: I recently had to ship something, so I went to Staples and bought some bubble wrap to protect my item. Staples sent me a customer satisfaction survey asking me about my “bubble wrap experience”! Seriously?

NPS May Capture Intent, But Not Actual Behavior

Regrettably, management teams incorrectly believe that they need just one measure (NPS) to predict sales growth, profitability, and retention. This misperception also diminishes the role of insights professionals and research departments who are often closest to the customer. Many of you may recall that the NPS buzz began with Fred Reichheld’s 2003 HBR article “The One Measure You Need To Grow”. His quote: “The only path to profitable growth may lie in a company’s ability to get its loyal customers to become, in effect, its marketing department” became gospel. A “willingness to recommend” formed the basis of the NPS question, and an entire industry was born. The assumption was that a recommendation can drive both growth and retention.

This is certainly plausible — if such recommendations are ever made — but recommendations are, in fact, rare and intent is not the same as action. Other self-reported measures of behavior, such as purchase intent, have long been known to be directionally associated with in-market behavior, but the magnitude varies tremendously and is highly sample-dependent. Over the last 15 years, research has also failed to prove a clear connection with one’s willingness to recommend and profitability, growth, or retention. And the assumption that any of this can be accomplished at near-zero cost is also dubious.

New, Improved NPS!

Common sense must be used, and adjustments made, as companies learn more about their customers and business performance. Virtually all businesses require more than a single measure to understand customer needs. As I have previously written (“Rethinking NPS”), customer feedback is essential; it is NPS that is inadequate.

It seems that nothing will be able to stop the NPS locomotive. According to the WSJ, J.D. Power has signed an agreement with Bain & Company to become the officially recognized authority for benchmarking the “Net Promoter Score”. Much like J.D. Power’s customer satisfaction awards in automobiles and other categories, can “NPS awards” be far behind? In its next life, NPS will no doubt morph into support for advertising claims and other puffery. For devotees of W. Edwards Deming, this should feel familiar: NPS will become a slogan, and more manipulated than ever.

Have Courage: Use Measures That Matter

But what to do? If you are in leadership, focus on how internal processes can be improved, rather than (mis)placing the responsibility on individuals or your front line staff. An employee is responsible for bringing problems to leadership’s attention, but your job is to improve the process itself. Be wary of salespeople claiming to offer magical and simplistic answers to difficult problems. Look at the big picture and choose from appropriate (multiple) measures linked to customer happiness. Build your own normative database to best meet the needs of your unique business. One size does not always fit all, and it will cost money to truly understand what is going on with your customers and your business. But at least you will have measures that make sense.

At Surveys & Forecasts, LLC, we have been doing just that for 25 years.

* [1] Net promoter score or NPS is a registered trademark of Satmetrix Systems, Inc. a privately held experience management company headquartered in San Mateo, CA. Along with the Bain & Company and Fred Reichheld, Satmetrix Systems co-developed the Net Promoter Score (NPS).
Surveys & Forecasts, LLC