OPINION24 October 2017

Diagnosis: NPS myopia

Data analytics Opinion UK

Net Promoter Score (NPS) is a victim of its own success, but there are six key ways to prevent NPS myopia, says Ryan Howard.

Net promoter score_crop

The Net Promoter Score (NPS) question segments respondents into Promoters, Passives and Detractors; and is then expressed as a score. We fancy more people saying nice things about us and less people saying not nice things. Sod the rest who keep schtum. This is the NPS. Boom. Brilliant. It sits aloft and alone on a throne of universal acceptance, carrying gravitas into boardrooms, making research commercially relevant.

It is also a victim of its own success, for with simplicity come unintended side effects. So much so that businesses are increasingly frustrated with unresponsive or wayward tracking.  With all eyes focused on this single figure, unexpected scores can progress rapidly into a textbook case of NPS myopia. Not only does this contagion trigger anxiety, it discredits good research.

Should you sense onset:

Gather context

Investing in functional aspects of your product or service may have little or no traction on brand trust. Unlike satisfaction, recommendation is social and emotional. Here, personal credibility and a sense of self are at play. So if recommendation is about how a brand makes one feel then this net widens to include brand image – perhaps just as much influenced by media spend and messaging than the qualities of your offering. A recommendation does not even require prior experience.

A strong recommendation is also less relevant in some sectors than in others where the first rule of Fight Club may apply. How likely is it is that you would recommend this rash cream to your friends and colleagues? No, never. Not cool!

Ignore comparisons

Unless you operate in a highly commoditised sector, those advocating your brand above others do so with good reason. Exceptional customer experience cannot be driven by relating your score to an average, comprised of a fuzzy assortment of outdated scores. Brow beating a benchmark or tanking against it, are red herrings. Similarly, the same NPS does not work well across culture where the same response is poor in Japan but awesome in Germany.

Be uncertain

When we calculate an NPS, we reduce our data, from a granular 11-point to a three-point scale. In doing so we create more statistical uncertainty – we need a lot more sample to achieve the same level of precision. This means an NPS arrives with much wider confidence intervals than we would ordinarily expect. Moreover, a derived score (a subtraction of two percentages) offends the rules of classical measurement theory and isn’t subject to standard significance testing.

Step back

Changes in the distribution of raw data may reveal great shifts beneath a stagnant NPS. An inspired improvement may well boost a large proportion of your customers off the bottom end of the scale. Though far less likely to cry foul to friends and family, their NPS segment remains unchanged.

Dig deeper

An initiative can delight some while upsetting others. Spend groups may swap segments, so it is entirely possible that revenue or customer lifetime value rockets while the NPS straight-lines or nosedives.

Formulate the cure

All things remaining equal, the NPS is a telling read of trajectory and a great deal of thought has gone into it. There is nothing wrong with the measure itself – it is all down to interpretation. In absence of insight, marrying brand strategy to a summary score that attempts to encompass so much nuance is inevitably problematic. The soaring influence of word of mouth demands more.

Do 10 customers singing your praise outweigh the efforts of a single ranting disparager? Does a low NPS equate to people talking poorly about you or not at all? Might a response of six negate another of ten? Does online have the same dynamic as off-line? Is a ‘given’ the same as a ‘received’ recommendation and would you report the NPS in isolation, devoid of interpretation, diagnostics and depth? No, never. Not cool.

Ryan Howard is director of advanced analytics at Simpson Carpenter


*Net promoter, net promoter score and NPS are registered trademarks of Bain & Company, Inc., Satmetrix Systems, Inc., and Fred Reichheld.