Every support team tracks at least one of these. Most track both. Fewer than half use them correctly - meaning they understand what each metric actually measures, what behaviors it incentivizes, and when each one is the right lens for a given decision.
This article gives you clean definitions, explains the perverse incentives that emerge when you optimize for the wrong metric, makes the case for a composite approach, provides industry benchmarks, and shows you how Velaro automates the collection of both.
Clean Definitions
- Measures satisfaction with a specific interaction
- Collected immediately after a conversation ends
- Typically: "How satisfied were you with this interaction?" 1–5 or 1–10
- Score = % who rated 4–5 (or 9–10)
- Transactional - reflects a single moment in time
- Best for: agent-level coaching, queue performance
- Measures likelihood to recommend the company
- Collected periodically (quarterly, post-milestone)
- Question: "How likely are you to recommend us?" 0–10
- Score = % Promoters (9–10) minus % Detractors (0–6)
- Relational - reflects the overall relationship
- Best for: retention risk, product feedback, executive reporting
The key distinction: CSAT measures the transaction; NPS measures the relationship. A customer can give you a 5/5 CSAT on a chat conversation and still be a detractor on your NPS survey - because CSAT captures the quality of that one interaction, not their cumulative experience with your company.
The Perverse Incentive Problem
Every metric creates incentives. When those incentives aren't aligned with the actual goal (customer retention, revenue, satisfaction), you get perverse outcomes.
⚠️ Optimizing for CSAT alone creates these problems:
Agents learn to close conversations quickly on a positive note - even when the underlying problem isn't solved. A customer who was told "our team will look into this and email you" rates the interaction 5/5 because the agent was friendly and responsive. The underlying issue wasn't resolved. CSAT looks great; the customer churns three weeks later when the problem recurs.
Another pattern: agents avoid transferring or escalating difficult conversations because a transferred chat often yields lower CSAT. The agent who hands off gets a poor score even when the escalation was correct. This discourages good behavior.
⚠️ Optimizing for NPS alone creates these problems:
NPS is collected quarterly and aggregated at the company level. Individual agents have almost no direct connection to their NPS impact - the signal is too diluted and too delayed to drive behavior change. Support teams that only report NPS miss the week-to-week, agent-by-agent data that's actually actionable for support operations. NPS tells you there's a problem. CSAT tells you where and with whom.
"CSAT without NPS optimizes for pleasant interactions. NPS without CSAT optimizes for feelings about the brand while ignoring daily execution. You need both - for different decisions."
The Composite Metric Approach
The right framework uses CSAT and NPS for different purposes at different levels of the organization:
Agent Level: CSAT
The metric agents can directly influence conversation-by-conversation. Weekly CSAT trends by agent are your coaching data. A 0.4-point drop over two weeks is an early intervention signal.
Team Level: CSAT + First Contact Resolution
Team CSAT tells you whether routing and staffing produce good experiences at scale. FCR tells you whether issues are actually resolved - not just closed politely.
Executive Level: NPS + Customer Effort Score
NPS captures relationship health. CES (how easy was it to get help?) is a powerful predictor of churn often missed by teams focused only on CSAT. Low effort = high loyalty.
Velaro collects CSAT automatically after every conversation and sends NPS surveys on your schedule.
See CSAT & NPS setupIndustry Benchmarks: What "Good" Looks Like
| Industry | CSAT Benchmark | NPS Benchmark | FCR Benchmark |
|---|---|---|---|
| E-commerce / Retail | 4.4+ / 5.0 | +35 to +55 | 78%+ |
| SaaS / B2B Software | 4.3+ / 5.0 | +40 to +60 | 72%+ |
| Healthcare | 4.0+ / 5.0 | +30 to +50 | 70%+ |
| Financial Services | 3.9+ / 5.0 | +20 to +45 | 80%+ |
| Telecommunications | 3.6+ / 5.0 | +10 to +30 | 65%+ |
| Travel / Hospitality | 4.2+ / 5.0 | +30 to +55 | 73%+ |
If your scores are below the benchmark for your industry, there's almost always a specific driver - response time, first-contact resolution rate, agent knowledge gaps, or channel coverage. The benchmark comparison tells you that you have a problem; the CSAT driver analysis tells you what the problem is.
How to Set Up Both in Velaro Automatically
CSAT Collection
In Velaro, CSAT surveys fire automatically when a chat conversation is closed. Configuration options:
- Survey timing: Immediately on close, or with a 2-hour delay (for async channels like email where the customer may not be at their screen)
- Scale: 5-star rating (default) or emoji-based for mobile-heavy audiences
- Optional follow-up question: Open-text "What could we have done better?" - this is where your qualitative coaching data lives
- Agent attribution: CSAT scores are automatically attributed to the closing agent and rolled up by team, channel, and time period in the Analytics dashboard
NPS Collection
NPS in Velaro can be sent:
- On a schedule (quarterly to your entire customer base, via email)
- Post-milestone (after 90 days of service, after a support ticket is resolved, after onboarding completes)
- Triggered by CSAT patterns - customers who give 3 or fewer stars on CSAT can be automatically flagged for an NPS survey and a proactive follow-up from a success manager
The CSAT/NPS Disconnect: When to Investigate
The most revealing signal is when CSAT and NPS diverge. Some patterns to watch for:
- High CSAT, Low NPS: Customers like your agents but don't like your product or company. Support is performing well; the product or sales experience is failing. This is a signal to escalate to product or customer success, not support operations.
- Low CSAT, High NPS: Customers love your brand but recently had a bad support experience. This is a temporary signal - perhaps a team member is struggling or a new product issue is generating contact. Intervene quickly before relationship goodwill is depleted.
- Both declining simultaneously: Systemic problem. Could be a product issue driving support volume, a staffing gap, or a process breakdown. This is the pattern that requires immediate cross-functional attention.
Velaro's dashboard surfaces both metrics side by side with trend lines, making the divergence pattern visible in real time rather than in your quarterly review.
The Bottom Line
CSAT and NPS answer different questions and belong in the same dashboard. CSAT tells you whether your support team is executing well today; NPS tells you whether customers plan to stay and refer others tomorrow. Teams that track only one are flying half-blind. Run both, watch for divergence between them, and use the divergence to find problems before they show up in churn.
Frequently Asked Questions
What is the difference between CSAT and NPS?
CSAT (Customer Satisfaction Score) measures satisfaction with a specific interaction - typically asked immediately after a support contact. NPS (Net Promoter Score) measures overall brand loyalty and likelihood to recommend - typically asked on a periodic basis. CSAT is transactional and operational; NPS is relational and strategic. Both are useful; neither replaces the other.
Which is better: CSAT or NPS?
Neither is universally better - they serve different purposes. CSAT is better for diagnosing support quality and agent performance because it ties directly to specific interactions. NPS is better for predicting retention, referrals, and long-term revenue because it captures overall brand sentiment. Most mature support teams use both and watch for divergence between them as an early warning signal.
How do you measure customer satisfaction in support?
The most common approach is a post-chat CSAT survey with a 1–5 or thumbs up/down rating, sent immediately after the conversation closes. Aim for at least a 20% response rate for statistical reliability. Supplement CSAT with first contact resolution rate, average handle time, and sentiment analysis on chat transcripts to build a complete picture of support quality.
What is a good CSAT score for live chat?
A good CSAT score for live chat is 85% or higher on a percentage-satisfied basis. Best-in-class live chat teams achieve 90–95%. Scores below 75% indicate systematic issues with resolution quality, wait times, or agent training. Live chat typically outperforms phone and email on CSAT because the format is faster, more convenient, and produces a complete transcript customers can reference.
When should I use CSAT vs NPS?
Use CSAT after every support interaction to measure execution quality and catch agent or process issues quickly. Use NPS on a quarterly or semi-annual basis to measure overall brand and product sentiment. If you can only run one: use CSAT for support teams (it's more actionable at the operational level) and NPS for executive and product reporting (it's more predictive of long-term business outcomes).