A number of observers have eloquently weighed in on the weaknesses of Fred Reichheld’s Net Promoter Score (for an excellent discussion go to Adelino de Almeida’s Profitable Marketing blog). I’ll go one step further.
Firms should stop measuring their NPS. Why? Because the NPS:
- Doesn’t help explain WHY a customer would recommend the firm. Let’s say a bank finds out that 10% of its branches score much higher than the average on the NPS and that 10% score much lower. What has it learned? Nothing. It isn’t actionable. You might argue that it provides clues as to where to dig in… but wouldn’t it be more useful to find out the root causes in the first place?
- Measures intention, not behavior. If there’s one thing market research has taught us, it’s that consumers don’t always do what they say they’ll do. Far more important to most firms is who is actually refers the company to their friends and family — not who might do so.
- Doesn’t capture inherent consumer differences. Forrester Research’s consumer research has found that Gen X and Gen Yers are more likely to recommend a firm to their family and friends than us (cranky, crotchedy) Baby Boomers. So if your NPS increases from one year to the next, is it because you improved the products and services you deliver, or does it simply reflect an underlying change in the demographics of your customer base?
- Can incent undesirable behavior. One marketing exec told me about an interaction he had at his car dealer’s repair shop. When he arrived to pick up his car, the shop manager said “if there’s any reason you wouldn’t check off the ‘likely to recommend’ box on the customer satisfaction survey, please let me know before filling out the survey.” Do you want your firm’s personnel running around asking customers to say they’d refer the firm to friends and family — or doing the things that earn a referral?
- Uses funds better deployed elsewhere. Measurement doesn’t come for free. Firms that have built an infrastructure to measure customer satisfaction are now being encouraged to build an infrastructure around measuring NPS. It’s not worth it. There are better things to measure — like what causes a customer to refer a firm in the first place.
It’s been said that you can’t manage what you don’t measure. But a metric that doesn’t help you manage isn’t worth measuring. And the Net Promoter Score is a measure that doesn’t help you manage.
Ron Shevlin's Marketing Whims 
Great blog – ran into via Bankervision. Would love to see more bank-specific insight (which you have in spades).
I have seen in a Bank the use of NPS to drive home to local Executive across a country their need to change behaviours in a clear way. While it made the point that action was required, the issue became – what action?
Seems to me that this bank spent a lot of money just to determine that corrective action was needed. And that’s critical to my point: With all the scrutiny placed on how money is spent in an organization (what’s the ROI? what’s the ROI?)…what IS the ROI on all the money spent to measure NPS? There may be an ROI…but my bet is that there are plenty of other investments (even in metric measurement) that yield a better return.
I agree that NPS isn’t the right measure because you can’t tell what contributed to it. I’ve written hereabout my preferred alternative. But let’s be realistic: things like NPS are effective because they are a simple, single number that everyone can focus on and react to. Without that as a starting point, no one will ever take the next step of trying to improve them. As marketers, we’re supposed to be experts in motivating human behavior: how come we so often fail to apply those skills to our internal business practices?
Ron,
Great points about some of the problems with NPS. Let me add a few. 1) A “defector” is not always a defector. Just because someone doesn’t recommend, doesn’t mean they aren’t a loyal customer. At best NPS gives us a metric of postive word of mouth. It doesn’t measure negative word of mouth. 2) The data Reichheld provides to support his theory that NPS predicts growth are misleading. Look closely at his charts. He compares revenue growth over a period of time, and then looks at NPS at the end of that period. Doesn’t that imply revenue growth drives NPS? 3) we have done a lot of research, we have over 15 million survey responses to the recommend question. When you calcualte NPS from that question you get very high margin of error (usually 3 times higher than simply using the mean score of the question). We see that it is sometimes correlated with revenue growth, but never causing revenue growth. When you introduce customer satisfaction into the analysis you see that customer satisfaction will drive recommendations and customer satisfaction will drive revenue growth. Obviously this requires that you measure customer satisfaction accurately. We use the proven methodology of the American Customer Satisfaction Index to do so.
Great blog!
We use NPS as a metric but certainly not the sole metric of our customer satisfaction loyalty program and as a predictor of potential future customer behavior. I agree that it serves as an indicator but one that doesn’t provide you with the underlying causes. I compare it to the “check engine” lamp on my car dashboard. It tells me when I have a problem but it does not reveal the cause of the problem. For that I must dig deeper and hence in customer satisfaction that means we need additional measures. The other issue I see with net promoter is that it’s made up of two components-promoters and detractors. You must take into account how many promoters and how many detractors you have and consider the ratios thereof. A 25% net promoter score can be arrived from several different pathways, some good (25-0) and some not so good (50-25). I’m not a subscriber to all-or-nothing theories and I find that those in the “NPS is the only metric you need” camp a bit over zealous for me.
If NPS is not a good score, then what is? Is there any other tool that could be used instead of NPS?