Net Promoter Score Defined

If you are a good marketer you've heard of the Net Promoter Score (NPS). If you are a very good one, you know what your NPS is. If you have no idea what I am talking about, read the definition at the bottom, and then meet me at paragraph two.

NPS is a rather bold little calculation. Its inventor, Fred Reichheld (pictured above), argues that a single question and its resulting score is the only metric you need to measure satisfaction. He also claims a correlation between a high NPS and future revenue growth.

And the metric has taken off. Online forums have sprung up, full of managers keen to discuss it. High-profile chief executives have publicly praised NPS and added it to their management systems. Even finance people have taken note, with several institutional investors asking for the score as part of their due diligence. But two audiences are genuinely unhappy with NPS.

The first is market research firms, which make their money from long-winded analyses of the market that are so complex they require lots of researchers to explain the findings to befuddled clients.

Clearly, one number that your mother-in-law could compute is a direct threat to market researchers and their business. As a result, many have been hell-bent on showing up NPS as being not all that it is cracked up to be. Even Nigel Hollis, one of the biggest and fairest brains at Millward Brown, concluded: 'The jury is still out. We need to see more compelling proof that NPS actually does lead business performance before we adopt it as the sole benchmark of success.'

The group even more appalled by the apparent over-simplicity and over-exposure of NPS are marketing professors - particularly those who study customer satisfaction. Publicly outraged that sloppy statistical methods and clear research bias were being accepted at face value, privately they were probably pissed off that their complex statistical analyses, which have had next to no impact on real marketers, were being superseded by a single score and percentage from an ex-management consultant who did not even have a PhD. Reichheld enraged them further by admitting his lack of interest in statistics. He calls his method 'common sense' and claims the business leaders he targets have 'little interest in advanced statistical methods'.

The top peer-reviewed marketing journals have published a slew of anti-NPS research. Most notable was the publication in July of 2007 in The Journal of Marketing of a statistical analysis of Norwegian customer data. The paper, 'A Longitudinal Examination of Net Promoter and Firm Revenue Growth', shows that NPS fares no better than other measures of service satisfaction in predicting business growth. The paper tetchily concludes: 'We find no support for the claim that Net Promoter is the single most reliable indicator of a company's ability to grow.'

Who is right? Usually my posts here on BSI conclude that everyone bar this humble blogger is completely mistaken. But, in this rare case, everyone is right. Market researchers and academics are correct to question the lazy analysis and over-selling of NPS. But I still side with Reichheld. It may not be all it's cracked up to be, but I have seen more non-marketing executives - managers with real power - quote NPS in the past two years than all the other marketing concepts put together. This concept is making senior people think about where the money comes from. Not balance sheets, products or sales, but what really drives the business: the customers.

And anything that gets managers to think about customers and their needs is a step in the right direction. Even if that step is a little sloppy and its effectiveness exaggerated.

Net Promoter Score definition:

    * The concept was developed by Fred Reichheld, a consultant, strategist and author on loyalty. His books include The Loyalty Effect, Loyalty Rules! and The Ultimate Question.

    * Introduced in December 2003, Net Promoter is a metric derived from survey responses to a 'how likely are you to recommend...' question. Respondents who provide a rating of nine-10 are 'promoters'; those who give ratings of six or lower are 'detractors'. The NPS is found by subtracting the proportion of detractors from the proportion of promoters.

    * Reichheld's research among 4000 companies showed NPS to be 100% accurate in indicating whether a firm would grow (more consumers championed its service or product) or shrink (more were denigrating it).

    * NPS has been adopted by companies including Microsoft and American Express; many report their score to investors.

Courtesy of Marketing Magazine

Sponsored By: The Brand Positioning Workshop

Comments

[Joy Levin]

I agree that there is validity in both sides of the argument. Overall, I think NPS is a good starting point, that needs to be explored by each company more fully to understand the dynamics behind it and how it can be exploited for true customer understanding.

[Pepe Alvarado-Risoul]

The beauty of the NPS is that it's simple, easy to compute, it's a customer-based metric, and it makes sense (although still to be seen if it really works); the problem, though, is that it isn't actionable as it is, therefore you still need a mechanism to figure out the drivers of recommendation among the customer segment you are measuring. Benchmarking is not enough as Reichheld suggests in his paper "The One number You Need to Grow."

[Bruno Ribeiro]

I've used NPS on some costumer satisfaction surveys and wasn't impressed. The reliability is feeble, but on the other hand, as Joy Levin points out, is a good starting point and an easy way to explain what needs to be done to someone who isn't keen on statistics.

I think we'll need more studies on its validity and reliability before claiming it as the Holy Grail of marketing research or dismiss it purely based on its simplicity.

[Rian]

I think you leave out one group nay-sayers in your article -- those of us who actually use it in real-world businesses. In my opinion, the biggest problem with NPS is that it is being misused.

NPS is a *loyalty assessment* metric. Both of those words are essential to understand NPS, and the key to its misuse:

* Loyalty -- it is an overall metric of customer sentiment, and should not be tied to specific initiatives your company is doing, because there's no way of knowing which change among multiple changes affected the change.

* Assessment -- it is a backward-looking metric. It should *not* be used for product planning or future direction, even if you "mine the verbatims."

The problem is not the oversimplification, I think NPS is a good metric to measure the progress companies are making on loyalty. But when it gets used to drive strategic direction, the train derails, in my opinion.

[Denise Wymore]

This is one of the best blog posts I've ever read on the subject and the comments prove your point - they're both right.

Market research techniques are like religions to those who practice them - NPS is a new religion. It's not for everyone. It's kind of new-agey and talks about feelings. It focuses less on the stats and more on the feedback.

And like religion - there is no one true God (that outta spark some comments) so we argue over who's praying to the right one.

I like NPS cuz when I walk out of "church", I feel good - like I helped someone - the customer.

PS - I do believe in God.

[Wim Rampen]

I think there can be an advantage of NPS in achieving better customer focus in a company. On the other side of the coin, so can CSAT/ACSI, or internal measurement of retention levels and developments.

If you like to see a good read on NPS and video from Zappos CEO on how the improve NPS, please take a look at my blog: http://is.gd/1URYs

[Charlie Baggett]

Wow, maybe NPS is just too simple? Maybe simple, is just too simple? I say the proof is in the pudding and if you don’t like or want to try pudding, don’t! Your competition will love you for it!!!!

[Neil Hartley]

I think 1) NPS is being mis-used in some situations (e.g. different scoring schemes), and 2) the business needs to expand on NPS and perform driver analysis to understand why customers are promoters and why they are detractors.

[Esteban Kolsky]

There are a couple of things here that I have to comment on.

1. NPS is supposed to measure loyalty (not satisfaction) - but it makes no difference as neither one is possible to be measured.

2. I have written plenty against NPS and not because I am a market research person (I am not), but because I see executives using it to mismanage their organizations. There are more cases from uninterested parties explaining why NPS does not work than there are case studies from interested parties (BAH and Satmetrix specifically) on why it does work.

There is a certain value to having a single metric to understand what customers are doing. Executives tend to get lost in deep data analysis that would show correlation between valid metrics, and the NPS is a simple concept to understand: you either have promoters or detractors (there is also neutrals) and you have to focus on your promoters because those are the ones that make you money. Anyone can understand that. It it also supposedly to allow comparisons between companies from the customer perspective (something all executives love to show is that they are better than others). This is all supposed to be good.

Except that is not the way it works. NPS is supposed to ask only one question (two actually, propensity to recommend and propensity to return for more business), but most organizations simply implement it as part of a long plethora of questions - thus losing the value of being the golden question. It would make more sense as a single question, but asking it as one more of a bunch makes it lose the context.

NPS is supposed to predict customer behavior (they are likely to return or to recommend), but instead ends up being a magnificent tool (looks at what happened to justify the previous score, not at what could be done to improve it or preserve it).

Finally, NPS is no more useful to any organization than any other collection of correlated metrics. You could measure profitability by segment (you do segment - right?) and get more insight than by tracking NPS. You could measure cause-effect results for any program you implement (you do track end-to-end interactions and marketing campaigns - right?) and get more value and insight that if you use NPS. You could even mix the two above and measure profitability tied to segments in a cause-effect correlation to a campaign and know so much more than NPS can tell you about your customers.

Bottom line is that NPS does not do what is supposed to do. You can read as many reports from management consultants (which I am one of them) regarding how NPS was brilliant and allowed them to measure correlation between promoters and detractors, but that won't hide the simple fact that flaws NPS: it attempts to measure loyalty -- which is not possible to measure (how do you measure a historically predisposition to feel a certain way in one frozen moment of time?). That flaw, shared by customer satisfaction measurements, is the reason we don't see companies progressing on their relationships with their customers.

There is indeed a simple golden question to ask customers to make sure they are satisfied and remain loyal: did you get what you needed in a timely manner?

Anyways, I will get off my soap box now. My rant is done, and feel free to ignore it -- but remember that NPS (or CSAT for that matter) don't work as a true measure of customer feelings. That can never be measured properly.