Misusing Recommendations

Our clients often expect us to include a recommendation question in a survey, to get an NPS score. However, Valeria Maltoni over at Conversation Agent says that the NPS (Net Promoter Score) doesn’t tell you enough. She’s right.

Some history: looking for a tool that would best predict corporate success, Fred Reichheld examined an assortment of survey questions, and found that a positive answer to the question “Would you recommend our company to a friend or colleague?” had the highest correlation with long term revenue growth.

To help improve corporate performance, Reichheld recommended that the question be asked soon after a transaction, to quickly assess the experience, and keep employees aware of their crucial role as the company’s public face.

It was also a good way to tell which agents or group of agents were leaving the best impression.

That was it. End of story.

Except that companies immediately started misusing it, asking it annually (or more often) of all customers regardless of whether or not they’d done a transaction. At i-OP one of our own vendors sent us a survey with the NPS question twice a year, like clockwork. No reasons why, nothing qualitative.

Problem is, B2B relationships are complex – a lot of factors affect our experience, and you need to understand which of those have the highest correlation with the customer’s overall assessment. To do this analysis you need to use some form of regression-like analysis (we actually use something called Bayesian Inference) to find the drivers behind the experience.

If you just ask the question, and don’t delve into the reasons, you’re diminishing the relationship, not building it. And if you don’t understand what the drivers are behind that experience, you don’t know how to focus your efforts to improve it.

We dropped that vendor. It’s not enough to ask how you’re doing – you need to ask why, and listen deeply to the answers.

About these ads