Three ways the great NPS debate rages on

Every defense of Net Promoter Score eventually lands in the same place: “NPS is fine. You just can’t use it on its own.”

With arguments, results, charts, metaphors, caveats, and healthy debate, CX appears to agree on this one thing: NPS cannot stand alone.

And that is precisely its problem.

As a metric, Net Promoter Score (NPS) was never positioned as one signal among many. It was sold as the signal. The one question to rule them all. One number. Three types of customers. One simple way to understand how customers feel and thus predict how the business will perform.

If a metric only works when surrounded by disclaimers and other data, then the issue is not misuse. The issue just might be the question itself.

Don't miss any news, updates or insider tips from CX Network by getting them delivered to your inbox. Sign up to our newsletter and join our community of experts. 

Learn More

Argument #1: “It’s just misused”

This is the most common defense of NPS, and one of the least convincing for its continued use.

Apparently, it’s widely accepted that everyone misuses it. Leaders turn it into a target, then teams game the score. Goodhart’s Law manifests itself: when a measure becomes a target, it ceases to be a good measure. Company-wide it becomes the number that rightly or wrongly many successful or unsuccessful projects will hang their hat on. The question, “what effect does that have on NPS?” gets asked. (Well-meaning but what you meant to ask was, “what impact will this have on customers?”)

“If we could just get back to its original intended use, then it’s a great metric,” some say.
NPS has been misunderstood and misused across industries, geographies, and business models for years. When a common tool continues to fail so broadly, the question is no longer “why do we keep getting it wrong?” The question becomes “why does the tool make it so easy to get wrong?”

Clearly the scale of this misuse tells us just how wildly out of control NPS really is.  
Its simplicity is its greatest strength and its biggest weakness. 

NPS is easy to explain upward. Easy to benchmark. Easy to turn into a target. Easy to tie to incentives. Those are not accidental qualities. It’s why it spread so quickly.

My argument is that a metric that encourages score chasing, simplification, and performance theatre is not being misused at all; it is behaving exactly as modern organizations intend it to.
NPS requires so many caveats and explanations, so much supplementary analysis, disclaimers, safeguards, and exceptions. Which begs the question, is a metric that requires this many guardrails actually fit for purpose? Is it even worth using if left to its own simplicity that the world over will misuse it time and time again?

We don’t use it as intended because it’s so easy not to. Its simplicity (read: vagueness) is its own detriment. 

Defending NPS by saying “you just have to use it properly” is defending a system that only works in a utopian society. We have to account for rampant human behaviour and error. Unfortunately, NPS doesn’t measure up. 

Argument #2: “NPS measures loyalty, not experience”

Another argument is that NPS was never meant to measure experience at all. It measures brand loyalty.

That is true.

But it supports a damning argument that I wonder if its defenders realize. 

If NPS is used to measure loyalty, why is it used to manage frontline service? Why is it used to judge individual interactions? Why is it used to prioritize CX investment, redesign processes, and assess the quality of day-to-day experiences?

Pointing out the obvious, while this argument is used in defense of using NPS, it is one of the greatest arguments that NPS is not a good measure for customer experience. 

As I argued in my first article, NPS is not even asking customers about their experience. We are not asking what happened, how easy it was, whether expectations were met, or where friction occurred. We are asking a question about advocacy and then retrofitting the answer to explain something else entirely.

We’re not asking the right questions. We’re asking a question about Category A and applying the answer to Category B. 

NPS is a blunt, lagging, relationship-sentiment measure. It tells you how someone feels about a brand at a point in time, not what happened to them, where the experience failed, or what should be fixed next.

We force NPS to answer questions about moments, interactions, and systems it was never designed to evaluate and then we have the gall to be disappointed by the result.

And yet, when NPS fails to explain what went wrong, we do not question the question. We question the execution. We blame follow-up. We blame governance. We blame the organization. We blame the people using it. 

We are using a loyalty question as if it were an experience measure and then act shocked when it cannot do the job we want it to. 

Argument #3: “Just add more metrics”

Another popular argument in support of NPS is that it is perfectly fine… as long as it is paired with other measures. CSat. CES. Verbatims. Segmentation. Financial overlays. Operational data. Governance frameworks.

In other words, NPS is absolutely acceptable as a metric provided it is surrounded by better data. 

But at that point, is NPS really the insight? Or is it simply the familiar headline number that remains because executives recognize it, not because it’s providing any color or significant weight.

This defense is reasonable on the surface, but just like argument #2, it concedes the point. If NPS is only useful once it is contextualized by multiple other data sources, then it is not the measure we should be using. It’s just an easy label. 

In reality, true insight comes from talking to frontline teams, customer verbatims, pain point analysis, operational data, and outcomes. NPS has been enjoying its useless throne as a summarizing artifact for far too long, often disconnected from its subjects, the actual drivers of change.

If every CX practitioner agrees that:

• NPS alone is insufficient
• It is not diagnostic
• It requires extensive scaffolding
• It should not be a target
• It should not be used transactionally

Then the original value proposition has failed.

Bain and Company did not promise us that NPS was good when you use it with better measures. They promised that it simplified complexity into something we could act on. Now we know that is not the case and what remains is a number that persists largely out of habit and perceived benchmarking. 

Despite needing to surround itself with other metrics, NPS remains the number that is tracked most closely, discussed most frequently, and escalated most urgently, even when everyone knows it can’t explain what actually happened on the ground.

The metric is no longer serving insight. It’s just a comfort blanket. Perhaps that’s the best argument for keeping it around though, familiarity?

The industry’s real problem with NPS

NPS succeeded because it made CX understandable and quantifiable to executives. It gave the discipline a much-needed number.

That success does not make it timeless. Yet it did lay the framework for what a new metric needs to achieve. Albeit in a more concrete, customer-focused way.

Today, NPS often functions as an inflated sense of customer understanding. It reassures everyone that experience is definitely being measured. It allows organizations to say they are customer-focused without being accountable for what that actually means.

At some point, continuing to defend the tool stops being pragmatism and starts ironically becoming arguments for its removal. 

The problem is not that NPS is dead. The problem is that we are still asking it to do work it was never capable of doing in the first place.

We did not simplify CX.

We oversimplified it.

And we have spent the last two decades dealing with the long term consequences.

Quick links

• Top free customer insights tools to elevate your CX in 2026
• Is NPS to blame for the decline in customer experience?
• Bringing a long-term vision to CX