NPS Burnout Pt. 2 — On classic CX metrics. Or: Never trust a statistic you …

The week after that steering committee was strangely quiet. No follow-up questions. No escalation. Just silence, as if the room had collectively decided to move past the topic. I sat at my desk, the CMO’s sentence still ringing in my ears, and thought about what to do. My first impulse: better slides. More context. An executive summary that brings out the benefit more clearly. Maybe a dashboard that puts the key numbers on a single page.

It took me about two hours to realize that this was the wrong way to go.

NPS, CSAT, and CES are well designed. They measure what they are meant to measure: sentiment, perception, and a customer’s perceived effort at a specific point in time. In the right hands, with the right context, they provide valid signals about the state of a customer relationship. The problem lies elsewhere. These metrics show where customers stand. They do not answer why they are there or what that position costs or delivers for the company. Anyone who regularly sits in steering committees knows the pattern: The numbers are presented. There is a brief nod. Sometimes a question. Then the meeting moves on. Disinterest? No, no one wants to make a decision based solely on sentiment.

Because at some point, after a few years of consistent measurement, most programs encounter the same phenomenon: the numbers barely move anymore. NPS (or CSAT) fluctuates within a narrow range. The scores remain stable, sometimes up a point, sometimes down one. Internally, this is interpreted as stagnation, and sometimes as proof that feedback programs are merely a nice-to-have. Often, it is neither. It is the instrument’s diminishing marginal value. Put differently: given this starting point, what exactly is there left to decide?

Recently, a “new” classification of metrics—Leading vs. Lagging Indicators—has been gaining attention. The idea is sound: examine what the past explains (Lagging) and identify signals for future developments (Leading). I, too, once believed this was the solution.

Yet a metrics taxonomy alone provides no additional insight as long as the fundamental question remains the same: which specific moments along the journey make the difference.

What actually gets heard in executive suites: operational signals. Handovers that stall. Response times that fluctuate. Rework that increases. Reliability that declines.

These patterns immediately resonate with management because they translate directly into costs, risk, and customer behavior. No statistical voodoo or room for interpretation is needed. No context needs to be explained first. “Customers are frustrated because …” + “Frustrated customers represent … EUR of the pipeline.”

CX managers who speak this language get a different seat at the table.

The question I asked myself during those weeks was simple: What if we operationalized measurement differently? What if we stopped capturing moods and started tracking impact?

The next episode explores what this means in concrete terms and what we found along the way.