Concerns with The Net Promoter Score (NPS)

Brief History of the Net Promoter Score (NPS)

The NPS was developed by management consultant Fred Reichfeld (associated with consulting firm Bain and Co.) and presented in a Harvard Business Review article in 2003.

Since that time the metric has become widely adopted across many industries and is estimated to be in use in around 2/3rds of Fortune 1000 firms. Part of its appeal is that it is perceived as a simple and logical measure of overall customer value and advocacy, which can be encapsulated in a single question.

The other main factor driving its widespread adoption is the hype from agencies and research firms that their clients must measure NPS on an ongoing basis (and thereby creating a long-term income stream for the advising firm).

Original Article and Subsequent Academic Support

The original Reichfeld article was not peer-reviewed and was not published as a journal article. In 2007 Keiningham et al. replicated the research design and analysis undertaken by Reichfeld, but they were unable to find a relationship between attitudinal loyalty (NPS) and a subsequent change in revenue. Their study was published in the Journal of Marketing.

In 2013, Pollack and Alexandrov (with half of their study based on banking services) supported the Keiningham study and stated that “given the present state of evidence, it cannot be recommended to use the NPI (net promoter index) as a predictor of growth nor financial performance”.

Therefore, since publication the NPS paper was first published in the early 2000’s, it has struggled to attract widespread academic support for its claims to a link between NPS and financial performance.

And while it appears that while the Reichfeld article was able to show some correlation in some industries, there was no NPS-revenue causation demonstrated. This would need to be analyzed and demonstrated within a firm’s own situation before usage and reliance.

Why would there be a limited NPS-Revenue Relationship?

Indeed the NPS concept appears to have some logical basis – more happy customers = more word-of-mouth = more new customers = more revenue!

But the original paper was based on  simplistic industries that sell one service (e.g car rentals and airlines).

But in most firms, there are so many drivers and influencers of revenue and word-of-mouth and attitudinal loyalty are just two factors among many aspects.

As a example, when Vodafone in Australia had a dramatic reduction in customer numbers (due to network coverage issues) their NPS rose significantly as the detractors left, indicating an inverse relationship between NPS and revenue.

And in some industries, it would be very unlikely that a NPS score would have any significant revenue impact for the following reasons:

  • Corporate revenue is usually driven by multiple and complex factors including product lines, economic cycles, legislative, pricing, supplier and other partner deals, customer relationships, brand equity, promotional success, and so on.
  • Detractors (in the NPS sense) may also likely to remain loyal and profitable customers in some industries due to in-built switching barriers – which means that detractors will continue to contribute to revenue.
  • Personal word-of-mouth sources (promoters) may also be minor influences in some industries, depending upon the strength of retailers, competing brands, mainstream influencers, ease and convenience factors, and habitual purchasing behavior.

It is Best to Use NPS as a Diagnostic Tool Only (not as an output)

Possibility due to the simplicity of the concept, many firms have mistakenly transitioned NPS from an input mechanism to a key outcome. NPS has some value in service recovery provided the root causes are identified, as well as validating process and service improvements as they are implemented over time. However, it is relatively meaningless as an output metric given its lack of validation and the ease of measuring true behavioral loyalty.

Many service firms seem to rely upon NPS as a quasi-loyalty metric in the absence of any hard data. But firms with access to a customer database will have accurate and precise loyalty data, which can be analyzed by various segments, product relationships, time, acquisition source, and so on – making customer database analysis an extremely valuable tool. And because customer relationships evolve over time, database analysis can also provide a predictive assessment of likely future loyalty based on actual (not intended) behavior.

Research Approach and ROI

One danger with NPS questionnaires is that they operate on a self-selection basis where respondents choose to be involved. This results in a non-representative sample, which means that the survey results will bounce up and down over time due to survey errors.

The only solution is to implement a formal and structured survey approach where a representative sample of customers is proactively surveyed on a regular basis. However, this would require the use of a professional market research company to help develop and execute, most likely at a reasonable (six figure investment), which then highlights the need to justify the “investment” on an ROI basis.

A set of related questions would also need to be developed for the different categories of respondents (e.g. promoters and detractors) to delve down and identify the factors and encounters driving their viewpoints. This is more difficult to do in a quantitative survey and may be more suited to qualitative (focus groups) research in a smaller firm. This extra information is required as a NPS score by itself, without the underlying causes, gives the organization nothing to work with and/or improve.

The Case Against NPS in a Nutshell

  • NPS was developed as a commercial market research product
  • Its justification was based upon correlation, not causation
  • Since its development, there has been no widespread academic or research support for the relationship between past NPS and future revenues
  • It was developed in 2003, prior to the main social media platforms, such as Facebook and YouTube, which have changed how consumers interrelate and doesn’t factor in the power of media influencers
  • There are too many revenue drivers and influencers and WOM uplift is just one of many of these
  • There is a danger of misuse of marketing dollars to overly focus on WOM if it is a minor or insignificant factors, and will lead to a negative ROI
  • Some industries, such as banking, are even more unsuitable for NPS usage, given its switching barriers, customer inertia, dominance of home loan brokers, strong influencing role of comparison sites, and the importance of price (interest rates) to consumers
  • NPS may have some minor value as a diagnostic tool for service areas, provided that it is broken down and reviewed on a segment, product and acquisition source basis
  • It is recommended that KPIs should not be set against NPS, as it is likely to encourage poor or inefficient marketing and CRM practices (where customers are not all equal as they have different $ value)
  • NPS is best used as a input tool to help improve the customer experience
  • NPS surveys, to construct correctly, are expensive and will probably not have a ROI payback to a small firm – focus groups would be a better source of information
  • Some customer databases are better able to provide both actual and predictive behavioral loyalty along with tangible financial results