6 min read · 1,334 words
A Net Promoter Score of 62 looks like a trophy on a slide. It told nobody in the room whether the account it came from was about to renew, expand, or quietly let the contract lapse at the next procurement cycle. NPS was never built to answer that question, and in B2B it mostly doesn’t.
I sat through a QBR a few years ago where the CX team opened with a regional NPS of 58 — “world-class,” someone said, and a few people nodded. Two months later we lost a mid-sized account that had scored a 9 on the same survey. The champion who filled it in had moved teams; the new procurement lead had never been asked a single question. The score was real. It was also measuring a person who no longer had a vote.
What NPS was actually built to measure
Fred Reichheld introduced the Net Promoter Score in a 2003 Harvard Business Review article, built on consumer research: one person, one relationship, one purchase decision, asked whether they’d recommend a company to a friend. That’s a coherent question for a bank customer or a streaming subscriber. It is a much stranger question to ask a procurement manager who didn’t choose your company, doesn’t personally use your product, and whose job is explicitly not to be swayed by likability.
Why B2B breaks the model
Gartner’s research on B2B buying groups puts the average number of stakeholders involved in a purchase decision at six to ten people, spanning procurement, end users, finance, and an executive sponsor who may never touch the product. NPS surveys typically reach one of those people — usually whoever’s email is in the CRM, usually the friendliest one. You are not measuring the account. You are measuring your best relationship inside the account, and calling it representative.
The academic critique of NPS’s predictive power isn’t new. A 2007 paper by Timothy Keiningham and colleagues, published in the Journal of Marketing, tested the claim that NPS is the single best predictor of growth against a broader set of satisfaction metrics and found no evidence it outperformed them. That paper is nearly two decades old and the debate still hasn’t settled — which itself tells you the “one number” claim was always a marketing line, not a measurement finding.
What the score misses in a renewal-driven business
In a B2B relationship, the commercial outcome you actually care about is renewal, expansion, or churn — and none of those live on a single detractor-to-promoter scale filled in by one contact. Forrester’s CX Index research has consistently found that customer experience quality correlates with revenue growth at the account level, tracked over time, not at the level of a single quarterly survey score. Qualtrics’ XM Institute has published similar findings: the organisations that outperform on customer-led growth are the ones tracking outcome metrics tied to the account, not sentiment metrics tied to an individual.
Zendesk’s CX Trends research adds a B2B-specific data point worth citing directly: a majority of business buyers now say experience is as important as the product itself in their purchase and renewal decisions — which makes it even more important that whatever metric represents “experience” on a dashboard actually reflects the account, not one respondent’s mood on a Tuesday.
Here’s the honest problem with NPS in B2B: it is cheap to collect, easy to put on a slide, and comparable across quarters, so it survives long after people stop trusting it. I’ve watched three separate organisations keep reporting NPS at the board level specifically because the alternative — building a proper multi-stakeholder health score — takes real work, and nobody wanted to own that project.
| Metric | What it captures | What it misses in B2B |
|---|---|---|
| NPS | Sentiment of one respondent, one moment | Whether the whole buying group is aligned; whether the account is commercially healthy |
| CSAT | Satisfaction with a specific interaction | Long-term account trajectory; only measures a moment, not a relationship |
| Net Revenue Retention | Actual account-level commercial outcome | Lagging — tells you what already happened, not why |
| Executive sponsor engagement | Whether the account has a champion above the day-to-day contact | Harder to quantify; requires CRM discipline most teams don’t have |
What to track instead
None of this means throw out satisfaction data — it means stop treating one score as if it were a commercial forecast. A more honest account-health model combines: net revenue retention (the actual number), expansion or contraction trend over the last two renewal cycles, breadth of stakeholder engagement (how many distinct people in the account have you had a real conversation with in 90 days), and a qualitative signal from the executive sponsor specifically, not just whoever answers the survey. This is close to the account-health scoring that customer success teams at PLG software companies have been building for years — B2B services and manufacturing businesses have mostly not caught up.
I wrote about a related measurement trap in why B2B brand awareness is the wrong metric to track — the pattern repeats: a metric built for consumer, single-decision-maker contexts gets imported into B2B wholesale, survives because it’s simple to report, and quietly stops meaning what everyone assumes it means.
How do you retire NPS without the room feeling like you’re hiding bad news?
This is the real objection, and it’s fair — NPS has boardroom credibility precisely because everyone already understands it. The honest approach isn’t to kill it overnight; it’s to keep reporting it for context while introducing the account-health metrics alongside it, and let the gap between the two tell the story. When NPS says 58 and net revenue retention says 91%, that gap is more informative than either number alone, and it’s a much harder story to wave away with “we’re doing fine.”
What this means for you
If NPS is the headline metric in your CX reporting, don’t remove it — recontextualise it. Add net revenue retention, stakeholder breadth, and executive sponsor engagement as companion metrics for every account, not just the ones flagged at risk. I’ve argued for a similar shift on the data side in data-driven marketing and unlocking insights — the organisations that win aren’t the ones with the fanciest single number, they’re the ones willing to track something less flattering and more true.
Frequently asked questions
Is NPS completely useless in B2B?
No — it’s a reasonable pulse-check on individual relationships and it’s cheap to run consistently. The problem is treating it as a proxy for account health or a predictor of renewal, which the underlying research doesn’t support in multi-stakeholder buying contexts.
What should replace NPS as the primary CX metric in B2B?
There isn’t a single replacement — that’s the point. Net revenue retention, stakeholder breadth, and executive sponsor sentiment together give a more honest picture than any single score, including NPS itself.
Why do B2B companies keep using NPS if it’s flawed?
It’s simple to collect, easy to benchmark against competitors, and boards already understand it. Building a proper multi-stakeholder account-health model takes real operational work that most CX teams aren’t resourced to do.
How many stakeholders are typically involved in a B2B buying decision?
Gartner’s research puts the average buying group at six to ten people spanning procurement, end users, finance and an executive sponsor — which is exactly why a single-respondent survey score struggles to represent the account.
Does a high NPS mean an account is safe from churn?
Not reliably. A high score from one satisfied day-to-day contact says nothing about whether the executive sponsor is engaged or whether procurement is quietly evaluating alternatives — which is where most enterprise churn actually starts.
If your CX dashboard leads with NPS, what’s the metric sitting quietly next to it that actually predicted your last churn? I’d like to know what other teams have found works.
Related reading: customer journey mapping that survives contact with reality.
