Thursday, August 02, 2012

Net Promoter Score: The Power of Simple Metrics

The goal of a corporate "dashboard" should be to provide a quick snapshot of how the business is doing.   However, far too many firms create complex dashboards full of a wide variety of metrics.   They overload managers with information, and they don't get the results that they intended to achieve.

I was reminded of the power of simple metrics, as I read this article at Knowledge @ Wharton.  It discusses the concept of "net promoter score."   The concept is straightforward: How likely would you be to recommend my company, my product or my service to your friends, your colleagues or your family members?  Fred Reichheld and Rob Markey developed this concept, and their research suggests that a firm's net promoter score is highly correlated with a number of other key measures of financial performance. 

In this exchange between Markey and Wharton Professor Peter Fader, we see a fascinating discussion about Net Promoter Score.  The bottom line: We could enhance the accuracy of the Net Promoter Score, but it may not be worth doing so.  Every firm should keep this conversation in mind as it identifies and formulates key business metrics:

Markey: The truth is that the Net Promoter Score is designed to be radically simple, not because it is statistically better, but because it is statistically fine and that simplicity appeals to frontline employees. Even CEOs can understand it.  The designations of promoter, passive and detractor are based on one question. It's a simplifying construct that helps motivate and inspire people to want to create more promoters and fewer detractors. If you really wanted a statistically robust thing that was about the statistically accurate correlations, you would always go for more questions. But what we found is that there's about a 10% or 15% improvement by adding more questions in terms of statistical accuracy, but it tremendously degrades your ability to motivate the organization to take action because then you get into these debates: Which questions are part of the index? How are they weighted? I don't know, maybe that question isn't relevant for my business. Then you end up debating the score and not actually focusing on what matters, which is getting your customers to stay longer, buy more and tell their friends. 

Fader: Indeed, what you've just described is very consistent with the academic research, which shows that a richer, multidimensional scale can be 10% to 15 % better. But this one question -- this ultimate question -- really is good enough. In the academic community, it's kind of a half-full, half-empty [situation]. I'm a half-full kind of guy, saying, "Give me a measure that is good enough, one that managers can actually appreciate, understand, implement and spread throughout the organization." It raises the whole idea of measurement and understanding customer differences to a level that we've never seen before in any organization.


Matti said...


Interesting post on an interesting topic. Pete Fader is a heavy-weight in the field and his program at Wharton can do a lot of good for customer metrics. Would be nice to see the academic research that was mentioned on the interview that shows only a 10-15% increase in accuracy from using other metrics.

The research I am aware of on this was published in the Journal of Marketing (JM) back in 2007. JM is marketing academia’s most prominent journal (and also happens to be the oldest academic business journal of any kind). There, Keiningham et al. showed that NPS did not correlate well with growth. Most importantly, they showed that the statements about statistical significance made for NPS and its potential rivals were without basis because the research findings published by Reichheld and Satmetrix were fake. Utterly fabricated. [Keiningham, Timothy L., Bruce Cooil, Tor Wallin Andreassen, and Lerzan Aksoy, A Longitudinal Examination of Net Promoter and Firm Revenue Growth, Journal of Marketing, vol. 71, no. 3 (July), pp. 39-51.]

Reichheld and his compatriots handled the ensuing situation beautifully: by not acknowledging it at all, not engaging in discussion, not explaining themselves or counter-attacking. There is no crisis of confidence unless they admit there is one. A paper in marketing’s prime journal does, of course, not make the news in any fashion. And it is hard to argue against the promotional machine of Bain Consulting or the adopters, like GE and many others, who pride themselves on using a “scientifically validated” tool.

Now, of course, this thing has been validated in the field, with hundreds or thousands of companies adopting it. What does it matter if the premise was built on sand and the tool does not do the things promised?

Other research has shown that NPS scores do not correlate with actual customer behavior. Perhaps the research in the new NPS 2.0 book where this is the major new message should, therefore, be examined with a fine-toothed comb, as well.

There are larger questions here. For me, the most interesting of those is what it says about academia, that proper research is unable to get any publicity, that academia cannot add to the popular discourse on managerial tools, and that academics themselves are more likely to buy what is sold in business magazines and trade publications rather than utilizing tested theories. NPS has entered textbooks as well, without criticism, and is being taught to business students.

Matti said...

One minor addition:

In managerial practice the problem with the NPS is that while it itself is easy to implement and interpret (on a very abstract level), the NPS score provides no guidance to what the reasons for the responses are and how customer experience could be improved – further questions are required for that. So, one ends up using more complicated tools anyhow, negating the beauty of a simple one-question method. And, therefore, using something like the extremely broadly tested ServQual metrics would be much more conducive to actual analysis, insight, and action from the get-go.

Michael Roberto said...

Thanks for that info, Matti! I was not aware of this stream of marketing research. I'm grateful that you wrote this detailed comment.

Bill Karpowicz said...
This comment has been removed by the author.
Bill Karpowicz said...

Mike and Matti: Great post and response! Matti, too bad the paper you;re referring to is from 2007. There is more information out there, and Rob presented some findings at the 2012 San Francisco Net Promoter conference. You can find the .pdf at in the Library section, or send me an email and I'll get a copy to you.


Bill Karpowicz said...

Mike and Matti:

Great post and response. The study you mention, Matti, is from 2007. Rob Markey presented at the 2012 Net Promoter Conference titled "The Evidence Linking Loyalty to Growth" You can find it at in the library section, or send me an email and I'll send you a copy. With more recent data available, NPS is a clear winner for companies.