NPS (Net Promoter Score or System) remain arguably the most popular marketing metric. It is widely used by managers, and I know many practitioners who love it. That said, academics are rarely in the ‘loving it’ camp. John Dawes is certainly not. In an article in the International Journal of Market Research lays out where we are with NPS and points to an overlooked problem, the overreaction of NPS to changes in the underlying responses from consumers.
Adoption Of NPS
You can’t really talk NPS without acknowledging that it has been, judging by popularity, a great success. Managers like to use it and even the C-suite gets in on the act.
Companies … appear to have included the NPS in their annual reports to a) signal to stakeholders the business is committed to being customer-oriented, b) to show that the company enjoys positive sentiment among customers, or has a plan to improve that sentiment; c) demonstrate its performance or improvement in a customer-based metric, then as a part of a managerial bonus plan.
Dawes, 2024, page 183
Dawes suggests that managers like it because it produces actionable, event provocative, scores. If NPS declines this gives impetus to act.
Classic Issues
Dawes outlines some well-known concerns with NPS. It is quite vague what it is actually meant to capture. While it is calculated by transforming a likelihood to recommend question it isn’t clear that NPS is about recommendation. Do people really mean to recommend? Do people actually recommend? In some industries it would be even a bit odd to recommend. Something might be great for you but that doesn’t mean you shout about it.
Is NPS a recommendation metric as opposed to loyalty, or intention to purchase more, or a measure of general brand strength. It makes sense that higher NPS are generally better, but academics aren’t fans when we don’t know exactly what managers are talking about. (We are more forgiving of ourselves, obviously).
The Overreaction Of NPS To Small Changes
The core focus of Dawes’ 2024 paper is that NPS is a bit hyperactive. One of the reasons it is provocative is that small changes in consumer responses can lead to what seem disproportionate changes in NPS.
This is because NPS is a transformation of a 0-10 response to the question about your likelihood to recommend. It classifies 9 and 10 as promoters and subtract from the promoter’s percentage the percentage of detractors — those answering 6 and below. Passives, 7-8 are completely ignored.
The problem is that by subtracting low scores from high score, the NPS method induces significant unwanted volatility in wave-wave results.
Dawes, 2024, page 189
| Average % wave-to-wave change (10 random samplings) | |
| Likelihood to Recommend | 2% |
| Net Promoter Score | 10% |
If you sample a population minor differences in who you interview can lead to major differences in NPS. Nothing much might be happening but your NPS might be doing all sorts of interesting and provocative things. Good if you want something to talk about at meetings, bad if you change strategy based upon the unwanted volatility.
For more on NPS see https://neilbendle.com/popular-marketing-metrics/net-promoter-score-not-as-magical-as-sometimes-suggested/, https://neilbendle.com/net-promoter/, https://neilbendle.com/managers-and-academics/, https://neilbendle.com/net-promoter-do-you-really-recommend/, https://neilbendle.com/net-promoter-and-lessons-for-academic-research/, https://neilbendle.com/recommendations-are-complex/, and https://neilbendle.com/the-benefits-of-simple/
Read: John Dawes (2024) The Net Promoter Score: What Should Managers Know?, International Journal of Market Research, 66 (2-3), pages 182-198
