The Key Question for Assessing Loyalty

This month's Business 2.0 heralds a "radical new research technique".  Well it's not really all that new or that radical, but it is starting to catch on.
First popularized by Frederick Reichheld, author of The Loyalty Effect, the method is relatively simple:

1. Ask your customers "would you recommend us/this product to a friend?"
2. Subtract the number of strong positive responses from the number of Negative plus neutral responses
3. The resulting number is the "net promoter" score

If you don't have a positive number, you have work to do.

Correlates with revenue growth

Reichheld's research says this number -- the net promoter score -- is strongly correlated with revenue growth. Companies with high scores, such as SouthWest Airlines, have higher revenue growth. He says most US businesses have a net promoter score of 10% or lower.

Other reasons to ask

While the net promoter score is a good metric, there are other reasons to ask the "would you recommend" question.  A good research analyst can do a lot more with that question than you might think at first glance, if it is part of a larger questionnaire.  You can run all manner of correlations and similar analyses to determine what elements of the product/experience are driving the response to that question.  You can produce those great quadrant charts that are so helpful in seeing where you stand versus competitors.  You can determine if some groups/segments are more likely than others to like or not like your product/service. 

If you do customer surveys, you really need to include a question of this type, sometimes more than one, because it is so valuable in analysis.

One number that everyone can understand

Reichheld's measure can give you a single number that everyone can understand and mobilize around. This is a big advantage when dealing with front line staff, who have a lot of information coming at them. 

If you want to use this approach, here are a few more best practices to think about:

1. Sample customers as soon as possible after they have interacted with you.  This means within days or weeks of obtaining a new car loan, for example
2. Ask them a few questions related to their experience.  A good place to start is whether they experienced the standard of service you are targetting. (ie. did anyone call you by name?  Did we thank you for your business?  etc.)  Keep it short, maybe five questions or even fewer.
3. Organize to collect this information in a systematic way.
4. When you see trends you don't understand or don't like, use other methods to find out why.  For this, nothing beats a good qualitative study. 
5. If you want to do this on a shoestring budget, develop a script and have your area managers and head office executives make the calls. 

One of the things that Damon does not really acknowledge in the article is that this one question does not solve all your problems.  If you want to know why the number is low (or why it is high), you will actually have to ask a lot more questions.  And the companies that are cited as examples, such as GE, do a LOT of other kinds of customer research.

Credits, Resources:
The neat typeface comes from Misprinted Type
The Business 2.0 article: "The Only Question that Matters", Damon Darlin, September 2005