July 18, 2011
Many of our clients work at companies that use Net Promoter. I recently had dinner with two of them at Forrester’s Customer Experience Forum, 2011. Both are senior people at companies that have been recognized as customer experience leaders in their respective industries.
When a third guest (Forrester’s CEO) asked them why and how they use Net Promoter Score (NPS), they gave remarkably consistent answers. In brief, they use it as a simple, easy-to-understand metric — one number — for aligning the business. Its main appeal is that busy executives don’t need to spend hours studying tables and spreadsheets to get a sense of how their firms are doing. Similarly, frontline employees down to the lowest levels of the organization find that NPS makes intuitive sense.
But there’s a next big (and obvious) question for people like our dinner guests who work to improve the customer experience at their companies: Does improving customer experience raise NPS? Because let’s face it, if your firm ties its overall health back to NPS, then you better be able to connect the dots for what you do, or you won’t seem to matter.
We’ve been wondering about this issue ourselves. So much so that late last year when we ran the big consumer survey that drives our Customer Experience Index, we included the Net Promoter question for two of the 13 industries in our study: banks and retailers. We were looking for a correlation between how people rate the customer experience at a company they do business with and how likely they are to recommend that company to a friend or colleague.
Without getting bogged down in details of the Customer Experience Index (CxPi) and NPS methodologies, here’s a two-sentence overview. The CxPi asks customers to rate how well a company met their needs, how easy the company was to do business with, and how enjoyable the company was to do business with and then averages the scores for the three questions into a single number. Net Promoter asks customers to answer one question (“How likely is it that you would recommend [this company] to a friend or colleague?”) on a scale from 0 to 10, it then subtracts the percentage of customers who answered 0 thru 6 from the percentage who answered 9 or 10.
Now comes the interesting part. We analyzed the correlation between how consumers rated their experience with the banks and retailers in our CxPi and their answer to the NPS question. We found that there is a very solid connection between customer experience and NPS for both banks and retailers.
To break it down without getting too math-wonky, a correlation between two variables is typically considered high when the correlation coefficient is 0.5 or above. The correlation between CxPi and NPS easily cleared that bar. The correlation coefficient for CxPi and NPS for banks was 0.679, and for retailers, it was 0.618. For practical purposes, that's about as good as it gets because correlations above 0.8 sometimes mean that you actually measured the same thing two different ways.
Of course if you are a math wonk, you’ll point out that just because two things correlate doesn't mean that one thing causes the other. And you’d be right, that’s not the way the math works (we took statistics classes, too).
However, when we look at the possible real-world explanations for these findings (the ones that business leaders like our dinner guests care about), we conclude that more often than not an improvement in customer experience will drive higher NPS. Other explanations, like the idea that people randomly decide to recommend a firm and then retroactively change their perceptions, aren’t ones that you’d want to put in front of your board of directors.
What action items should you take away from these findings? Our top recommendation for customer experience professionals, or anyone working to improve customer experience, is to calculate the correlation between customer experience and NPS for your firm's customers.
First make sure that you track direct measures of customer experience quality and NPS separately. Whether you use something like the CxPi or ACSI — or your own unique customer satisfaction measure — to grade customer experience, you don’t want to confuse that type of metric with any loyalty metric, whether it's NPS or plain old purchase intent.
Second, calculate the strength of the relationship between the two types of metrics. Forrester used a Pearson correlation to determine the correlation between CxPi and NPS, but you should use whatever metrics and math have already taken root at your company. Why? Because you want to prove that what you do matters in terms that your senior management already buys into.
We believe that what you do matters and that this type of math will help you make the case.
As always, we look forward to your comments.