NPS vs CSAT — When to Use Which

NPS and CSAT measure different things. Plain-language guide to when each metric is appropriate, common mistakes, and how Tabular Pro implements both.

Net Promoter Score (NPS) and Customer Satisfaction (CSAT) are the two most common customer feedback metrics. They measure different things. Picking the wrong one for your use case produces noisy data that nobody trusts.

Here's the honest breakdown.

NPS — Net Promoter Score

"On a scale of 0–10, how likely are you to recommend [Company/Product] to a friend or colleague?"

Score = % Promoters (9–10) minus % Detractors (0–6). Range: −100 to +100.

What it measures: broad customer sentiment and word-of-mouth potential.

When NPS is the right metric:

  • Brand-level relationship tracking
  • Quarterly board metric
  • Comparing across companies (industry benchmarks exist)
  • Detecting trend changes over time

When NPS is the wrong metric:

  • Measuring satisfaction with a specific transaction (use CSAT)
  • Measuring effort of a specific task (use CES)
  • B2B with small customer counts — the score is noisy at low N
  • After every interaction (creates survey fatigue, dilutes signal)

CSAT — Customer Satisfaction

"How satisfied were you with [the experience / product / interaction]?" (1–5 or 1–7 scale)

Score = % rating 4 or 5 (top-2-box). Range: 0–100%.

What it measures: satisfaction with a specific touchpoint or interaction.

When CSAT is the right metric:

  • Post-purchase, post-support, post-onboarding
  • Measuring a specific channel or feature
  • High-frequency, transactional measurement
  • Benchmarking against an internal baseline

When CSAT is the wrong metric:

  • Brand-level relationship measurement (use NPS)
  • Cross-company comparison (CSAT scales aren't standardized — a 1–5 vs 1–7 scale produces different numbers)
  • Predicting future behavior (NPS correlates better with retention)

CES — Customer Effort Score (the third option)

"[Company] made it easy for me to handle my issue." (Strongly Agree / Disagree on 1–7 scale)

When CES is the right metric:

  • Post-support interactions
  • Friction reduction projects
  • Self-service flow optimization

CES is the best predictor of customer churn after a service interaction. NPS is the best predictor of long-term retention and growth. CSAT is the best for "how did this specific moment go."

Common mistakes

  1. Using NPS for everything. NPS asked after every support ticket dilutes the signal and annoys customers. Use CSAT/CES for transactions, save NPS for relationship tracking.
  2. Comparing CSAT scores across companies. Different scales and different framing produce different scores. CSAT comparisons are only meaningful within your own data.
  3. Ignoring the verbatim. The score is summary; the open-end response is the explanation. Don't ship NPS without open-end coding.
  4. Acting on small samples. NPS at N=30 has wide confidence intervals. Wait for a stable read.
  5. Treating the score as the goal. The goal is improving the underlying experience. The score is a thermometer, not the patient.

How Tabular Pro implements all three

  • Native NPS, CSAT, and CES question types with the standard scales
  • Auto-calculated scores at the dashboard level
  • Verbatim collection with AI open-end coding so the "why" feeds into the same dashboard
  • Tracker workflow for repeat measurement waves
  • Weighting and significance testing for cross-segment comparisons

Related: Quantitative Research Platform · Cross-tab analysis · vs Delighted