This is a conversation we have often—and almost always with our most advanced clients. “Things are getting better on the ground. The comments are more positive. Teams can feel the difference. But the score itself isn’t moving anymore.”

The instinct is to conclude that the program is losing momentum. In reality, the opposite is often true: you’ve moved beyond the point where your metrics are still able to tell you what’s happening. Three mechanisms explain most cases.

1. Your customers are improving, but within their category

The Net Promoter Score (NPS) is structured around three categories: detractors (0–6), passives (7–8), and promoters (9–10). A customer who moves from 2 to 5 remains a detractor. A passive who moves from 7 to 8 remains a passive. A promoter who moves from 9 to 10 remains a promoter.

You can therefore improve the experience for thousands of customers without moving the score by a single point. A customer who moves from 0 to 6 gains six points but has no impact on the index; a customer who moves from 6 to 7 gains only one point but moves the index. The measure is insensitive to real gains and highly sensitive to category thresholds.

The full distribution of all eleven scores reveals the shift. You just have to look at it.

2. The same uniformity gap can describe two radically different networks

The uniformity index typically measures the distance between your best- and worst-performing averages. It is useful, but a range only sees its two endpoints and remains blind to everything in between.

Imagine a network where 10% of locations perform at 80% and 90% perform at 20%. The gap is 60 points. Two years later, 90% of the network performs at 80% and only 10% at 20%. The gap is still 60 points.

The same number. A network turned completely upside down. What changed was the distribution of performance, but the index doesn’t measure it.

3. Your measurement tool has reached its ceiling

When a customer experience index remains at 95% or higher, a significant proportion of individual ratings may already be at 100%. The instrument has no room left at the top. In psychometrics, this is known as a ceiling effect: when almost everyone selects “very satisfied,” the tool can no longer distinguish genuine delight from polite approval.

Imagine a bathroom scale calibrated up to 100 pounds. Everyone weighs 100 pounds, and everyone assumes they’re in peak condition. The problem isn’t the population, it’s the scale.

This isn’t a design flaw. A questionnaire calibrated for an organization at 78% may no longer be appropriate at 96%. The sensitivity of the measurement tool needs to evolve with the maturity of the organization.

What stagnation really means

When an index reaches a plateau, it doesn’t necessarily mean the program is losing momentum. Often, it means the program has succeeded, and its measurement tools haven’t been recalibrated since.

The answer isn’t to measure more. It’s to measure more precisely: look at distributions rather than averages, track where the mass is shifting rather than simply the range, and recalibrate scales so they continue to tell you something meaningful about a high-performing organization.

Your numbers aren’t necessarily stagnating. Your numbers aren’t necessarily stagnating. They’ve simply stopped teaching you something new.

If this conversation sounds familiar, it’s familiar to us too. It’s often where the most interesting work begins.