← ccPlanning Academy · Metrics track

CSAT, NPS & quality

Free visual lesson · about 5 minutes · short quiz at the end

ccPlanning academy · metrics

CSAT, NPS & quality

The experience metrics — what they capture, and where they mislead.

Why they matter

The operational metrics don’t measure the point.

Service level and AHT measure the machine. CSAT, NPS and quality try to measure the outcome — did we actually help, and how did it feel? They’re the closest thing to what the contact centre exists for.

CSAT

“How satisfied were you with this contact?”

Usually a post-contact rating. It’s immediate and contact-specific — great for spotting which interactions went well. But it captures the people who choose to respond, who skew to the delighted and the furious. The quiet middle goes unheard.

NPS

“Would you recommend us?”

A relationship metric, not a contact one. It reflects the whole brand — product, price, last week’s outage — not just this call. Holding a contact-centre agent’s bonus to NPS blames them for things they don’t control. Useful at company level; misleading at agent level.

Quality scores

The internal view of “good.”

An evaluator scores contacts against a rubric — did the agent verify identity, show empathy, resolve the issue? Valuable for coaching, but only as good as the rubric and the consistency of scorers. A checklist that rewards saying the script can punish actually helping.

The sampling trap

Small, biased samples masquerade as truth.

A 4% survey response rate or ten QA evaluations a month is a tiny, skewed slice. Treating a 3-point CSAT move on that base as real — and pinning it to one agent — is reading noise. Know your sample size before you act.

Using them well

Triangulate; coach, don’t rank.

No single experience metric is the truth — read CSAT, quality and repeat-contact rate together. Use them to coach and to find systemic problems, not to build league tables off noisy samples. The point is a better experience, not a tidier number.

Mind the base

A “3-point CSAT drop” that isn’t

An agent’s CSAT falls from 92 to 89 and a manager schedules a difficult conversation. But it’s built on eleven surveys this month — two grumpy responses move the number three points. That’s not a trend; it’s noise.

Pin a tiny, self-selected sample to one person and you’re managing randomness. Always ask “out of how many?” before you act.

The takeaway

Closest to the point — handle with care.

CSAT is contact-level (and self-selected), NPS is brand-level (not the agent’s fault), quality is your own rubric (only as fair as it’s built). Triangulate them, mind the sample, and use them to improve rather than to rank.

Now test yourself ↓

1 / 8

Slides done? Here’s the same idea in a bit more depth — the part worth keeping.

In depth: the metrics closest to the point — and the easiest to misread

Service level and AHT measure the machine; CSAT, NPS and quality try to measure the outcome — did we actually help, and how did it feel? They’re the closest thing to what the contact centre exists for, which is exactly why misusing them does so much damage. Each captures something real and each misleads in a specific, predictable way.

Three metrics, three blind spots

CSAT is a post-contact rating — immediate and contact-specific, great for spotting which interactions went well, but it captures only the people who choose to respond, who skew to the delighted and the furious while the quiet middle goes unheard. NPS is a relationship metric, not a contact one: it reflects the whole brand — product, price, last week’s outage — so it’s useful at company level but misleading at agent level, and pinning an agent’s bonus to it blames them for things they don’t control. Quality scores are your own internal view of “good,” valuable for coaching but only as good as the rubric and the consistency of scorers — a checklist that rewards saying the script can punish actually helping.

The sampling trap, and how to use them

The biggest practical error is treating small, biased samples as truth: a 4% survey response rate or ten QA evaluations a month is a tiny, skewed slice, so a three-point CSAT move on that base — especially pinned to one agent — is usually noise. Know your sample size before you act. Used well, no single experience metric is the truth: read CSAT, quality and repeat-contact rate together, use them to coach and to find systemic problems, and resist building league tables off noisy samples. The point is a better experience, not a tidier number.

The principle to remember: these are the metrics closest to the point, so handle them with care. CSAT is contact-level and self-selected, NPS is brand-level and not the agent’s fault, quality is only as fair as the rubric — triangulate them, mind the sample, and use them to improve rather than to rank.

Quick quiz

Five questions. Pick an answer to each, then check your score.

1. What do CSAT, NPS and quality try to measure that operational metrics don’t?

SL and AHT measure the machine; experience metrics measure the point.

2. Why is NPS misleading at the individual agent level?

NPS is a relationship metric; pinning it to one agent blames them for things they don’t control.

3. What’s a key bias in CSAT?

The quiet middle rarely responds, so CSAT over-represents the extremes.

4. What’s the risk of a quality rubric?

Quality is only as good as the rubric and the consistency of scorers.

5. How should you use experience metrics?

No single experience metric is the truth — read them together and coach, don’t rank.