It seems some of our statements need some clarification. Yes, our critique of the NPS above is largely in reference to ‘Relationship NPS’ and the core concept of using the specific method for asking a single question – “The Ultimate Question.” Much of what you’re saying seems completely accurate and logical, but also to be a significant expansion of the NPS core concept.

From our perspective, this expansion comes down to the end users creating an internal process that requires a lot of work – preparation, planning, execution, analysis, interpretation and then team wide communication. Done effectively, it seems reasonable one could achieve the results you’ve delineated in your comment. However, that’s mostly our point. In our experience, we’ve seen most companies fail to implement this kind of NPS process correctly simply because of the time requirement. Worse, they often fail to request feedback often enough and communicate the results to the people who need them the most – the customer facing teams.

With Hively, companies can get all (if not more) of the same benefits they would get from adopting the NPS method. However, there is nothing to set up, no process to implement, no plans to make and team members get ongoing, real time feedback at the convenience of the customer.

The expansion of the NPS questions also causes problems on the customer end. While asking for additional comments on one question seems to work, asking for answers and comments to multiple questions turns into a survey. Surveys have a low engagement rate because customers typically don’t have the time to complete them.

Again, with Hively, customers can provide valuable feedback with 1 click at their convenience.

If ‘Transactional NPS’ provides ongoing, real time feedback about a product, service, item or purpose then we’d agree this is valuable – provided the questions are limited to one, specific, ongoing and also provide immediate feedback. Further, although we’d agree that measuring your personal score against your last provides a good perspective, we do think it also helps to have some kind of simple outside comparison to measure against such as – similar products and/or other team members.

Sounds as though in regard to the labels we agree but can simply disagree about the effectiveness of the (0-6, 7-8, 9-10) rating scale. We’re not disputing the results of the research, just how the scale is interpreted by the end user. We’d also suggest there is more research to be done and new methods to test.

Again, thanks for the discussion!

1) NPS is not specific enough

There are two types of NPS survey: Transactional and Relationship. You are correct that Relationship surveys work at a company wide level. However, Transaction surveys work at the transactional level and would quickly catch your Apple hard drive crash issue. They can also be linked to staff, products, touch points, etc.

Also NPS does not ask just one question. The NPS question is indeed just one question (“How likely would you be to recommend…”) but any well implemented process will also have other ways of collecting driver information. Often with additional questions but always with the addition of a free text “tell us why” request.

2) The NPS score has no value on it’s own.

The benchmark issue is only really relevant to Relationship NPS.

In Transactional NPS the most important comparison is to your own score last month/quarter/year. A good NPS score is thus one that is better than your last. In my opinion the search for external comparisons is not that useful but benchmarking yourself against yourself is critical.

3) The NPS scale is arbitrary and needs labels.

All scales are arbitrary at some level. Yes it does need labels. However, the breakpoints in the analysis of the scale (0-6, 7-8, 9-10) are not arbitrary, they are empirical. Meaning that after extensive testing they were found to work the best.

Multiple ways to get the same score.

Yes there are multiple ways to get the same score and that is a strength. One of the pieces of analysis you do when working with NPS is to look at the shape of the detractor/ neutral / promoter profile. This gives you excellent insights into how to improve your score.

4) The NPS method is biased and subjective.

NPS is not biased, not subjective and not biased by design. On the contrary is it unbiased and empirical –more so than any other system that I have seen. That is one of the key points – after examining many different questions the NPS question was found to correlate to future revenue growth better than any other. It wasn’t made up on the spur of the moment it was researched and proven correct.

Different cultures and industries do have different average NPS scores. That goes to my preferred approach of look internally for your comparisons not to external benchmarks.

5) Most importantly, NPS is not actionable.

This is not the case. Implemented correctly NPS is absolutely actionable. Root cause analysis and action planning are critical in using NPS. As I am happy to tell anyone “if you’re just going to measure NPS and do nothing –save your money.”

The additional questions to which I alluded above provide that critical information on what is important for customers: right down to individual touch points. We have had customers who implement Transactional NPS and within days they have actions plans to improve specific areas of customer dissatisfaction.

This was a general summary but I have a lot more information over at our blog:
http://genroe1to1.genroe.com/category/net-promoter-score/

Adam Ramshaw

Just wanted to drop in a comment here about Net Promoter Score. I understand that you provide a different system to NPS and I’m not going to suggest that it is perfect but many of the things that you say here about NPS are not correct.

Rather than go through point by point I’m just going to point everyone to a downloadable introduction here:
http://www.genroe.com/whitepapers/net-promoter-score-nps-an-introduction

I’m happy do to a point by point response if you’d like as well.

Adam Ramshaw