Net Promoter Score is the most widely used customer metric in business. It is also one of the most poorly matched tools to the context it is most commonly applied in.
NPS was developed by Fred Reichheld and Bain & Company in the early 2000s, primarily validated in consumer markets where purchase decisions are frequent, individual and relatively low-stakes. The logic was elegant: distill customer sentiment into a single number, track it over time and use it as a proxy for growth. In consumer contexts, where word-of-mouth drives acquisition and the link between recommendation and revenue is traceable, that logic holds reasonably well.
In B2B it breaks down almost immediately.
The wrong question for the wrong context
The NPS question – how likely are you to recommend us to a colleague or friend – assumes that recommendation behaviour is commercially relevant. In consumer markets it often is. In B2B it rarely is.
Purchasing decisions in B2B involve procurement processes, vendor panels, contract negotiations and multiple stakeholders with different priorities. A satisfied contact at a client company recommending your firm to a colleague has almost no bearing on whether that colleague’s company will actually buy from you. The decision mechanism is entirely different.
More importantly, the question has no mechanical relationship with renewal. A customer can be willing to recommend you and still not renew. A customer can be unwilling to recommend you for reasons entirely unrelated to their own experience (company policy, competitive sensitivity, personal preference) and still be a loyal, long-term account.
Measuring recommendation intent to manage retention is measuring the wrong thing.
The aggregation problem
Even if the question were well-chosen, the way NPS is reported creates a second problem. The score is an aggregate: a net figure derived from the difference between promoters and detractors across your entire customer base. That aggregation destroys the information that would actually be useful.
A score of 32 tells you nothing about which segments are driving it, which drivers are underperforming, which accounts are at risk or what would need to change to move it. It’s a single number that summarises a complex, multidimensional reality into a figure that can be reported on a dashboard and then largely ignored until the next measurement cycle.
In B2B, where customer bases are often small, segmented and strategically important at the individual account level, aggregate scores are particularly misleading. Losing one major account can represent 20% of revenue. NPS won’t tell you it was coming.
The action problem
The most practical limitation of NPS in B2B is that it doesn’t tell you what to do.
A score goes up or down. You don’t know why. You commission follow-up research to find out why. That research reveals some themes. Those themes get discussed in a quarterly review. Initiatives are proposed. By the time anything changes, the next measurement cycle is underway and the connection between the action and the outcome is impossible to establish.
Driver modelling inverts this sequence. It starts with the structural question: what factors actually influence loyalty in this customer base and how much? The measurement is designed around that structure. The output is a ranked map of where to act, not a number to explain.
Why it persists
NPS persists in B2B for the same reason most legacy metrics persist: it’s established, benchmarkable and easy to explain to a board. Those are real advantages. The problem is that they’re organisational advantages, not analytical ones. NPS is easy to report, not easy to act on.
The companies moving away from it aren’t doing so because they’ve found a simpler metric. They’re doing so because they’ve recognised that simplicity in measurement comes at the cost of utility in decision-making. A metric that everyone understands but nobody can act on isn’t serving its purpose.
The alternative
The case against NPS isn’t a case for complexity. It’s a case for measurement that is designed around the decisions commercial leaders actually need to make.
In B2B those decisions are: which accounts are at risk, what is driving the risk and where should we invest to change the outcome. A structural loyalty driver model is built to answer those questions. NPS isn’t.
CLPS replaces NPS as the primary loyalty metric with a validated driver model that measures the factors that actually determine whether B2B customers stay, expand or leave. The output is a diagnostic, not a score.