CustomerGauge NPS Intelligence · 2026
The NPSHall of Shame
Real scores. Questionable methods. Zero accountability.
We analysed thousands of publicly disclosed NPS scores. Some were rigorous,
credible, and genuinely impressive. Others were… not.
This is a celebration of the latter.
↓ scroll to be appalled ↓
A number without a methodology
is a claim, not a measurement.
Net Promoter Score (NPS) is one of the most powerful customer metrics ever invented. It is also one of the most
abused. In the right hands — measured consistently, across a representative sample,
reported transparently — it is a genuinely useful signal of customer loyalty and business health.
In the wrong hands, it becomes a press release accessory: a number
plucked from a convenience survey of seventeen delighted early adopters, dressed up
in a tuxedo and paraded in front of investors as evidence of world-class customer experience.
The companies below did not necessarily lie. They may have measured something,
at some point, with some respondents. But what they disclosed —
and how — tells you everything about whether their NPS means anything at all.
Crime: Touched the ceiling
"An NPS of 99 means that out of every 100 patients, 99 would enthusiastically recommend you. The one holdout was probably unconscious."
NPS 99 is not a score. It is a statistical miracle — or a sample of
about twelve people who all happened to be having the best week of their lives.
Uptiv disclosed this in a press release announcing a strategic investment, which is
roughly equivalent to asking your mum to write your Glassdoor review.
Infusion care patients are, by definition, grateful — they just received
life-improving treatment. Measuring NPS at peak gratitude is like asking
someone to review a restaurant immediately after you saved them from choking.
Methodology? Sample size? Auditor? Conspicuously absent.
🚩 PR context only
🚩 No sample size
🚩 Peak-moment capture
🚩 Zero prior disclosures
View the source disclosure →
Crime: Reported a range, not a number
"Somewhere between 97 and 99. We're not sure. We'd measure more precisely but we were too busy being excellent."
Genuine NPS programs report a specific number. That is rather the point of the metric.
Reporting a range of 97–99 suggests either that multiple surveys were run
(in which case, which one counts?) or that the number was estimated from memory
over lunch. Disclosed on Fast Company's Most Innovative Companies list —
a context specifically designed to make things sound impressive.
The healthcare industry average they cite as comparison is 38–58,
which is true and useful. Their own number appearing to be double that average
is less useful and more suspicious.
🚩 Range, not a score
🚩 Award context
🚩 No methodology
🚩 Captive patient base
View the source disclosure →
Crime: Hiding in an executive hire announcement
"We hired Tom. Also, our NPS is 95. These two facts are unrelated but please remember both."
The NPS was disclosed in a local Las Vegas newspaper covering the appointment
of a senior vice president. This is a classic camouflage technique:
bury a marketing claim inside a personnel announcement so it acquires
the veneer of corporate news without anyone asking inconvenient questions.
No prior disclosures. No sample. No auditor. No methodology.
Just: Tom is here, and we are, apparently, beloved.
Tom would presumably like the number to be true. We wish Tom well.
🚩 Hire announcement context
🚩 Local press source
🚩 No methodology
🚩 First and only disclosure
View the source disclosure →
Crime: Measuring the wrong people entirely
"Our volunteers love us. Shocking, we know. Next we'll survey people we gave free pizza to."
This one is not even subtle. The NPS of 94% was measured
among participants in their CSR volunteer programme —
people who chose to give up their free time to do good works under
Kaizen Gaming's banner. Volunteers self-select for enthusiasm.
They are, by construction, the most committed and positive subset of
any population. Measuring their NPS and presenting it alongside
business customer scores is a category error so fundamental
it deserves its own textbook chapter.
It is the equivalent of a restaurant chain measuring NPS
among its own chefs and calling it customer satisfaction.
🚩 Wrong respondent group
🚩 Self-selected sample
🚩 CSR ≠ customer NPS
🚩 Misleading comparison
View the source disclosure →
Crime: The suspicious decimal
"92.31. Not 92. Not 92.3. 92.31. The precision of someone measuring with a micrometer and the credibility of someone measuring with a feeling."
Two red flags in one number. First: the context —
this NPS was reported for a specific Innovator of the Year Award winner
within OSF HealthCare, not the institution as a whole. It is a sub-unit,
measuring a programme, not a company measuring its customers.
Second: the decimal. An NPS of 92.31 implies a very small
sample — the kind where individual respondents move the needle by fractions
of a point. Our rough maths suggests fewer than 200 respondents to produce
that level of decimal precision. If you need two decimal places,
you probably need more respondents.
🚩 Sub-unit, not company NPS
🚩 Implied tiny sample
🚩 Award context
🚩 Not comparable to enterprise NPS
View the source disclosure →
Crime: Measuring at peak euphoria
"We surveyed pilots immediately after telling them they got their dream job. They were surprisingly positive. Science."
Simtech's NPS of 93 was captured from APS MCC graduates —
trainee pilots who had just completed their course and were on their way
to airline placements. This is the single highest-emotion moment
in a pilot's career. They have just been handed the keys to the sky.
Asking them if they'd recommend Simtech at this precise instant is not
customer experience measurement. It is euphoria harvesting.
Survey the same cohort six months into their first job, dealing with
roster disputes and 4am standby calls, and the number might look different.
Methodology matters. Timing is methodology.
🚩 Peak-moment capture
🚩 Graduation context
🚩 Self-selected completers
🚩 No follow-up measure
View the source disclosure →
Crime: Measuring a captive market at its happiest moment
"We asked people who just bought their dream home if they liked us. They said yes. We are choosing to believe this is about us."
Title insurance is a product that most consumers do not choose.
Their mortgage lender or real estate agent selects the title company.
The customer simply signs the paperwork at closing — the same closing
at which they receive the keys to their new home,
typically the second-happiest day of their adult life.
WFG measures NPS at this precise moment, in this captive market,
and has arrived at 94 — both years they've disclosed it.
The consistency is, if anything, the most suspicious part:
real NPS programmes show variance. Identical scores year-over-year suggest
either a remarkably stable measurement or a remarkably stable methodology
problem.
🚩 Captive market
🚩 Peak life-moment
🚩 Suspiciously stable YoY
🚩 Direct ops only, not full business
View the source disclosure →
01
Peak-moment capture
Surveying customers at the highest emotional point of their experience — graduation, purchase completion, post-treatment — and presenting it as ongoing loyalty. You're measuring the high, not the relationship.
02
Sample opacity
Publishing a score without disclosing n=. If your NPS is 94 and your sample is 31 people, that is a different number than if your sample is 3,100. The scale is not the score.
03
Context laundering
Measuring a sub-programme, award winner, or volunteer cohort and presenting it alongside whole-company NPS benchmarks. Not all NPS is customer NPS.
04
PR-only disclosure
Releasing NPS inside a fundraising announcement, hire press release, or award entry — contexts where scrutiny is low and motivation to impress is high. Credible NPS appears in financial reports.
05
The solo disclosure
Publishing NPS once, with no prior score and no commitment to publish again. Real programmes report consistently. A single score with no history is a marketing claim, not a metric.
