NPS is not the problem and replacing it is not the answer. The problem is that most CX programmes build one ranking of their themes when the argument requires two.
Every CX leader has had the same meeting. The NPS trend is up. The slide is clear. Someone senior nods, and then asks the question that ends the good part of the meeting: “And what is that worth?”
The honest answer is usually a pause. Not because the work has no value. Because the measurement system was never built to produce a money number, and no amount of extra reporting will make it produce one.
NPS answers one question, and answers it well
It is worth saying this first, because the internet is full of people selling replacements.
The Net Promoter Score is comparable across time, across sites and across competitors, which very few experience measures are. It is also understood by everyone in the business without a briefing note, and that shared understanding took years to build.
It is close to universal, too. More than 90% of organisations collect customer feedback through surveys (Forrester, The State of VoC and CX Measurement, 2023), which is why the score is table stakes rather than an advantage.
Keep it. Nothing that follows argues for dropping it.
The two questions it was never built to answer
The score tells you how many promoters and detractors you have. It does not tell you which experiences moved the number, or what those experiences are worth.
The first gap is familiar and most teams have partly closed it. Driver analysis, theme tagging on verbatims: all of that gets you to correlation, which is where most VoC programmes stop.
The second gap is the expensive one. Forrester’s 2023 research into voice of the customer and CX measurement found that only 18% of CX leaders are effective at showing the relationship between CX metrics and business outcomes. Almost every organisation is measuring. Fewer than one in five can say what the measurement is worth.
That is not one team’s reporting failure. Survey platforms were built to collect and compare scores. Converting a score movement into a financial consequence was never in the specification.
Rank your themes twice
Here is the exercise that shows the gap in your own data, and it needs nothing you do not already hold.
Take your last quarter of open-text feedback. Identify the top five themes. Then rank those five themes twice.
Ranking one, by volume. How many customers raised each theme. Your dashboard almost certainly does this already, and it is probably the default view.
Ranking two, by value. The revenue attached to the customers who raised each theme. Account value, average order value, annual premium, lifetime value, whatever your business actually holds.
The two lists rarely match. A worked example, with round numbers so the shape is visible:
| Theme | Customers raising it | Volume rank | Revenue of those customers | Value rank |
|---|---|---|---|---|
| App login failures | 412 | 1 | £1.9m | 3 |
| Delivery delays | 388 | 2 | £1.4m | 4 |
| Renewal pricing | 96 | 4 | £4.6m | 1 |
| Onboarding confusion | 210 | 3 | £2.4m | 2 |
| Returns process | 74 | 5 | £0.6m | 5 |
Renewal pricing is fourth on the list your team looks at every morning and first on the list your finance director would care about. A quarter of the customers raise it, and they carry more than twice the revenue.
That reordering is what putting NPS to work means. Keep the score, connect it to the experiences behind it, rank those by financial consequence, and the number you present arrives with the actions it justifies.
Three places this usually breaks, and finding out which one is yours is half the value of running it.
Customer identifiers are missing from the verbatims. Plenty of survey deployments collect open text anonymously, or attach a response ID that does not join to anything commercial. If you cannot connect a comment to a customer, you cannot connect it to revenue, and this is the most common blocker by a distance.
The revenue sits in a system CX cannot query. The data exists, in finance or in the CRM, and getting a monthly extract requires a request, a justification and somebody else’s sprint. Worth knowing early, because it is a relationship problem rather than a technical one.
Themes are defined too broadly to price. “Customer service” is not a theme, it is a department. If your top theme covers 40% of your comments, it will also cover most of your revenue, and the ranking will tell you nothing. Themes have to be specific enough that fixing one is a project somebody could actually be given.
Why the rough version is not the real version
The spreadsheet proves the gap exists. It does not survive contact with a CFO, and here is why.
Sampling. A manual pass reads a fraction of the feedback, and Gartner found that 83.6% of standard survey verbatims provide low insight value, so the comments carrying the signal are a minority you have to read everything to find. Read 200 of 20,000 and your ranking has an unknown error bar.
Severity. Two customers raising the same theme are not carrying the same risk. One is mildly irritated and will renew. One has quietly decided this is the last contract. A straight count treats them identically, and that gap is most of the money.
Repeatability. A manual pass done once is a slide. Done monthly it is a routine, and only a routine changes what a team does.
None of that makes the exercise not worth doing. It makes it a diagnosis, not a system.
What changes when the second ranking exists
The priority list and the business case become the same document.
Roadmap arguments stop being opinion. “We should fix the app” is settled by whoever is most senior. “The app login failure carries the third largest financial exposure in our base” is a funding request, and funding requests compete on evidence. The board conversation then opens on exposure rather than on a score trend.
It is also good for the team that does it. Forrester has found that CX leaders who link their metrics to revenue are 29% more likely to secure budget increases, and Gartner has separately found the same 29% for organisations that calculate the link between satisfaction and business outcomes. Two research houses, two differently-worded studies, one figure.
Value Retail, which creates and operates The Bicester Collection, made exactly that connection. Their Director of Guest Experience Development, Mazhar Butt, described it this way:
“By merging spend data with the Adoreboard predictive insights, we could move beyond correlation and actually measure opportunity loss, how much money we have potentially lost when Trust falls below target. It makes the cost of poor customer experience visible in real financial terms.”
The phrase worth borrowing is “beyond correlation”. Plenty of teams can show that happier customers spend more. Far fewer can say what a specific drop is costing.
You already hold both halves
You already hold both halves of the second ranking. They sit in two systems that do not speak to each other, and the gap between the two lists is the argument you have been missing.
Doing that join every month, across all of your feedback rather than a sample, with emotional severity weighting the count, is our job. Running the rough version this week will tell you whether the gap is there. It usually is.
Curious what this looks like on your own data? Book a 15-minute walkthrough: adoreboard.com/demo
