Last week I read an article in the Wall Street Journal, The Dubious Management Fad Sweeping Corporate America, which said that American companies were irrationally obsessed with their net promoter scores, NPS. Last year 155 S&P 500 companies cited their NPS during earnings conference calls, none of whom said their score was declining. All of the companies believed their scores were predictors of growth ahead. The WSJ article was skeptical, and I am too. In my opinion, NPS is like Average Selling Price (ASP) and Gross Margin – interesting, but insufficient as stand-alone metrics. The devil is in the details.
I have no doubt the goal of all those companies is to make their businesses better –higher customer retention, more sales volume, and greater profit. Measuring those metrics after the fact can be too late, and companies want leading indicators. I also agree that when customers are happy with your products and services, they are more likely to continue to buy from you, and they are likely to be less price sensitive. So, if you can correctly measure customer happiness, it should provide you a leading indicator. Unfortunately, customer satisfaction surveys can be used incorrectly and single scores are really averages of a lot of variation. Those factors can make the resulting scores unreliable.
Several years ago, I wrote a blog post, Are You Misusing Customer Satisfaction Surveys? in which I discussed car dealers penalizing their workers whenever they received a score less than 9 (out of 10). Since then, each time I have taken my car for service to any dealer, my advisor has told me to expect a survey request via email, and scores below 9 are considered failure. The Journal article also said some companies’ employees tell customers their compensation is tied to the survey. Unless a customer is really angry, he or she is generally not inclined to say something that will cost the employee money.
I don’t answer car dealer customer surveys anymore. While in aggregate I am generally happy with the services I receive, there can be significant variation in my happiness. That is, some things are great, but there are often aspects of it that are not perfect. Unfortunately, the dealers will never hear about them and never have the opportunity to improve them, because I don’t want to affect their employees’ compensation. Similarly, the companies cited by the WSJ will not learn about areas they could improve and that would improve their customer loyalty.
Companies managing just the NPS remind me of trying to manage with averages of Gross Margins. Even when the average margins look acceptable, there are often customers or products that could be significantly improved. If you don’t look at the details and identify those that could be improved, you will not get any better.
Another problem with Net Promoter Scores is whether they actually predict customer behaviors. Not long ago we did some work for a B2B client who believed their high-NPS customers were more loyal. We found no correlation between the customer scores and their level of repeat purchases. The proportion of high-NPS customers who had completely stopped buying was similar to the proportion of low-NPS customers. We also evaluated price levels paid by the customers and found minimal differences in prices paid by the highly satisfied and less satisfied groups of customers. If NPS and pricing were not the reasons for customer defection, what was? Obtaining detailed customer satisfaction information would help answer that question.
This post is not intended to dump on customer satisfaction surveys or getting customer feedback. I believe those things are important, however I also believe they must provide enough detail to identify which facets of the customer experience can and should be improved.
Designing an appropriate survey can be a delicate balancing act. If it is too long, customers will not complete it; but if it is too short, companies will not capture enough detail to be actionable. Similarly, if the questions are too general, they do not allow the respondents to provide nuance to their answers or provide meaningful insight to the questioner. Open-ended free-form questions provide the respondent an option to elaborate, but those are often ignored, and they can be difficult to categorize. Modern survey tools enable you to offer more or fewer questions to the survey taker, depending on his/her response. You can also randomly divide customers into groups and offer short surveys that go into detail on a limited set of topics.
The important point in all this is details matter. It is hard work and takes time to create surveys and other feedback mechanisms that enable you to fine tune your products and services. Spending the time to get it right will be much more valuable to you than simplifying it to a single question or score.
Leave a Reply
Want to join the discussion?Feel free to contribute!