Described as everything from the ‘ultimate question’ to the more Tolkienian ‘One metric to rule them all’ it is perhaps not surprising that NPS is there to be shot at.
So why are people criticising NPS, and why is this not actually the real issue?
Let’s start with a bit of history.
In the early 2000s Fred Reicheld and a Bain team tested different questions to see how well the responses correlated with customer behaviour.
They were looking for a single question that was the best indicator of lifetime value. They discovered that the now famous "How likely are you to recommend Company X to a friend or colleague?" worked the best across a cross section of industries.
They found that:
“High scores on this question correlated strongly with repurchases, referrals, and other customer behaviours that contribute to a company's growth. In 11 of the 14 industry case studies that the team compiled, no other question was as powerful in predicting behaviour”
As NPS became increasingly popular (⅔s of the Fortune 1000 took it onboard) the debates began and they have continued ever since.
I spend a lot of time talking about understanding what customers think and feel, in fact this is how I introduce what I do; I help organisations understand what their customers think and feel so they can achieve their business objectives.
So naturally NPS and the questions around it are something I come across a lot, but to add just a little bit of rigour to proceedings I also sat down and reviewed a selection of around 20 articles evaluating (or sometimes just criticising) NPS.
Despite NPS being a widely used metric for assessing customer loyalty, it's certainly not without its critics.
Several scholars and practitioners have questioned the validity, reliability, and overall utility of the NPS.
So why is NPS being attacked?
1. Lack of Contextual Information
NPS provides a single number without contextual data, making it difficult to derive actionable insights.
2. No Standard for Interpretation Across Industries
While NPS is used across various industries, the score does not necessarily have a standard interpretation. What is considered a "good" NPS score in one industry may be "bad" in another.
3. Vulnerability to Manipulation
There are concerns that employees or departments can manipulate NPS scores for their benefit, particularly if incentives are tied to NPS performance.
4. Over-simplification of customer experience
Critics argue that reducing customer loyalty to a single number is overly simplistic and fails to capture the complex nature of customer experiences and relationships.
5. Susceptibility to Non-Response and Response Bias
The voluntary nature of NPS surveys can result in non-response bias, where only those with extremely positive or negative experiences respond.
6. It Doesn't Tell You Why
The NPS question— "How likely is it that you would recommend [our company/product/service] to your friend or colleague?"—quantifies likelihood to recommend but doesn't provide insights into why customers gave a particular score.
Without this qualitative data, it's hard to make specific improvements, and research supports the notion of additional variables (beyond NPS) being highly effective at predicting customer retention, which includes understanding the 'why' behind customer behaviour.
If it's worth doing, it’s worth doing right
The criticisms often levelled at NPS—such as ‘it doesn’t tell you why’ or being easy to manipulate are not flaws in the metric itself, but rather in its application or interpretation.
These challenges exist for virtually all customer metrics when used in isolation or without contextual insights.
Many alternative metrics are similarly limited and susceptible to the same pitfalls.
This highlights the need for a more holistic approach rather than a replacement of NPS.
From experience, for the vast majority of organisations, failure to gain value from NPS is much more likely to be due to failure to implement or execute effectively, rather than a failing of the metric itself.
So before we look at why turning away from NPS is not the real question organisations should be focused on right now, let's cover off some of the most common NPS pitfalls and how to avoid them.
Make sure you ask ‘Why?’
The first and most obvious thing, as Bain were very quick to identify is to make sure you ask why the particular score was given.
However, the devil is in the detail here, a key thing to note is that the specific wording of the ‘why?’ question can have a significant impact on the quality of the insights that you get.
As an example if you ask ‘what’s the one thing we could improve?’ you’ll probably find that most people only tell you one thing rather than multiple things or information on the entirety of their experience.
Don’t make the mistake of thinking it’s just a topline number to report on
The topline number is useful for top level reporting but NPS is much more than a headline score.
You don’t just have one type of customer so drilling down into more detail is a crucial part of getting the most out of NPS.
Combining the score and other customer variables can help unpick anomalies or create a starting point for areas of further investigation.
Working with a company that sold identical products in the USA and Canada they identified a marked difference between the two countries' NPS.
The score discrepancy obviously didn’t tell them why, but it alerted them there was an issue.
Digging into the unstructured comments from the ‘Why?’ question they uncovered a cross border shipping issue that was creating friction for their Canadian customers.
By addressing this issue they had a big impact on the NPS score for Canada bringing it up to parity with the USA.
The removal of the pain point led to reduced negative shipping comments and dissatisfaction and this was reflected in the NPS for Canada.
Beware of well intentioned but badly implemented KPIs
A number of the companies I’ve worked with have at some stage made the mistake of targeting and even bonusing employees on NPS without providing the clarity on how the employees can influence it.
Organisation: “Ok Steve, we now need you to put the customer at the heart of decision making”
Steve: “Err Ok…but I have no way of understanding what the customer thinks in relation to what I do”
As with any metric incorporating NPS into KPIs effectively is reliant on ensuring that people understand what it is, what drives improvements and how they can impact it.
Some basic questions to address when building NPS into KPIs are:
- Do employees know what it is and why it is important to the business?
- Do employees know how they can impact it?
- Are they equipped to impact it?
- Can it be gamed / influenced unfairly?
Although this last point isn’t specific to NPS and is applicable to setting any KPI, experience shows where NPS is concerned it is important to have a robust process to ensure that the collection of the data is unbiased.
A common mistake is placing the responsibility for sending a survey in the hands of people who have an incentive to select customers who they believe will give positive scores.
It’s also important to be aware and honest as an organisation - the goal is to improve, the score is just that - a way of keeping score.
Have a process to act on the insights
To benefit from the deployment of NPS and use it to drive business improvement you need a repeatable process to convert insights into results.
In brief this means using insights to identify the changes to make, taking action and then measuring the impact of the actions that were made.
It may sound obvious but identifying issues is only one piece of the puzzle and even making changes to fix them is still not the whole story.
Measuring the impact is crucial to driving continuous improvement and maximising the benefits of NPS.
Organisations often fail to build these processes for improvement into their standard ways of working. This can lead to disappointing results.
But are there some organisations where NPS is less effective and should not be used?
Yes, it’s important to bear in mind the specific format of NPS can in some instances be sub optimal.
In general the NPS question is great for pushing people to think about whether they would be willing to recommend a product or service to a friend (which for most people is a higher bar than whether they would use a product or service again themselves).
However, there are some organisations for whom the specific wording of NPS creates issues, specifically when customers / users tend to take it very literally to the extent that the comments become devalued, some examples include:
- An environment where the customer has no choice - government services for example - Which other passport office am I going to use, the Australian one?
- For certain financial products “I do not give financial advice”
- For niche products or services like computer games with a very specific audience “Recommend it?! My friends would never play this game!”
And perhaps also operating systems…
It should be pretty clear whether this is the case for your organisation.
Most organisations will get a proportion of literal or sarcastic responses but as long as it is a reasonably low proportion then it certainly doesn’t prevent you from using NPS.
As always you can be led by what your data shows.
So why isn’t focusing on which survey metric you are using the thing that most organisations should be worrying about right now?
Fundamentally it all comes back to what we are trying to achieve with NPS or any other CX approach.
We want to understand what our customers think and feel as this impacts what they do, and what they do is either beneficial or detrimental to our business (purchasing again, recommending to friends, churning, making a complaint etc)
NPS can now only ever be a single piece of the puzzle
NPS was created 3 years before Twitter came into existence and 5 years before Apple launched the app store, the customer feedback landscape has evolved unrecognisably since then.
They’re only one element of the rich tapestry of feedback that customers now provide.
As such, an organisation worrying about the relative merits of a 0-10 score or a 5 point likert scale if they are not able to effectively understand customers via the other 75% of the touchpoints they have (app store reviews, Zendesk tickets, Twitter, Online reviews etc) makes no sense.
It’s very much a secondary priority.
It’s a bit like worrying about changing the grips on a bicycle’s handlebars before you’ve made sure it has 2 wheels.
Being focused on the specific metric that is used misses the much larger and more pressing challenge.
The starting point for any organisation right now is not whether to switch from NPS or not.
The place that any organisation simply has to begin is how to set up the holistic customer intelligence architecture to help them understand customers across all stages of the journey including all channels that customers choose to use.
In Summary: It's Not About the Metric, It's About the Customer
The debate around the efficacy of NPS can feel like an endless spiral, drawing attention away from what matters: understanding your customers.
While NPS has its flaws, most criticisms don't lie with the metric itself but in how it's applied, understood, or acted upon.
For those who think NPS isn't ideal for their specific circumstances, it's not the end of the world. NPS is not the be-all and end-all; it's a tool in your toolbox.
However, sweating over whether NPS is the 'right' metric is missing the forest for the trees. The real danger lies in overlooking all of the other channels and metrics that provide the complete, more nuanced view of your customers.
In today's digital age, focusing on survey metrics like NPS while ignoring other feedback channels is akin to putting lipstick on a pig; it's cosmetic, not transformative.
The real question organisations need to answer is not what score should we use, but rather how do we create a holistic understanding of our customers including all of the different channels, touchpoints and methods they choose to engage with us.
Using 0-5 Stars vs a 0-10 scale is perhaps debatable, but ensuring you genuinely understand your customers is non negotiable.
.jpg)
