Customer service metrics: 12 KPIs to track in 2026 (with formulas)

Customer service metrics are quantitative indicators that measure how effectively your support team resolves issues, satisfies customers, and operates at scale. The 12 most important ones are CSAT, NPS, CES, FCR, First Response Time, Average Handle Time, Average Resolution Time, Churn Rate, Retention Rate, Customer Lifetime Value, Ticket Volume, and Cost Per Ticket.

This guide covers each metric with a formula, industry benchmark, and specific steps to improve it, so your team knows not just what to measure but what to do when a number looks wrong.

Quick reference: if you are tracking only 3 metrics today, start with CSAT (customer satisfaction), FCR (first contact resolution), and FRT (first response time). Together they cover quality, efficiency, and speed — the three dimensions that predict churn fastest.

What are customer service metrics?

Customer service metrics are quantitative measurements that show how well your support team is performing across three dimensions: quality of resolution, speed of response, and business impact. They turn subjective impressions (“customers seem happy”) into data you can act on.

The difference between a metric and a KPI matters in practice. A metric is any measurement your platform captures, such as average response time in minutes. A KPI is a metric tied to a strategic target, for example “first response time under 1 hour.” Every KPI is a metric, but not every metric becomes a KPI. Most support teams track 15-20 metrics but designate 3-5 as KPIs tied to quarterly goals.

For B2B SaaS teams, customer service metrics carry extra weight because support quality directly drives renewal decisions. A 2023 Zendesk benchmark report found that 73% of customers will switch to a competitor after multiple bad service experiences. Tracking the right metrics lets you catch warning signs before they become churn events.

3 categories of customer service metrics

Customer service metrics fall into three functional categories. Understanding which category a metric belongs to helps you avoid optimizing one dimension at the expense of another. Teams that focus only on speed metrics, for example, often see CSAT drop as agents rush through tickets.

Quality metrics

Quality metrics measure how well issues get resolved from the customer’s perspective. High scores here mean customers leave interactions satisfied and without needing to follow up. Key metrics: CSAT, NPS, CES, FCR.

Efficiency metrics

Efficiency metrics measure how your team uses time and resources. They matter for capacity planning and cost control. Key metrics: FRT, AHT, ART, Ticket Volume, Cost Per Ticket.

Business impact metrics

Business impact metrics connect support performance to revenue outcomes. They help leadership understand the ROI of investing in support. Key metrics: Churn Rate, Retention Rate, CLV.

Category	What it measures	Key metrics
Quality	How well issues are resolved from the customer’s view	CSAT, NPS, CES, FCR
Efficiency	Speed and resource usage	FRT, AHT, ART, Ticket Volume, Cost Per Ticket
Business impact	Connection to revenue and retention	Churn Rate, Retention Rate, CLV

Why you should track customer service metrics

Customer satisfaction is difficult to improve without measuring it first. Support teams that rely on gut feel tend to over-invest in the wrong areas and miss the signals that matter for retention. Tracking metrics gives you three practical advantages.

You can identify problems before they escalate. A spike in Average Resolution Time or a drop in FCR rate often precedes a CSAT decline by 2-3 weeks. Acting on the leading indicator is cheaper than recovering lost customer trust. For more on how to act on these signals, see our guide to customer service automation.

You can allocate resources based on data, not instinct. Ticket volume trends tell you when to hire, when to build self-service, and where to focus agent training. Without volume data, teams either over-staff (expensive) or burn out their agents (higher turnover).

You can connect support quality to business outcomes. Showing leadership that a 5% improvement in CSAT corresponds to a 3% reduction in churn turns support from a cost center into a revenue protection function. This framing is essential for getting budget for AI customer service tools or additional headcount.

Customer Satisfaction Score (CSAT)

CSAT measures how satisfied a customer was with a specific support interaction. It is the most immediate quality signal you have, collected right after a ticket closes or a chat session ends. Unlike NPS, which captures long-term loyalty, CSAT captures in-the-moment satisfaction with a single interaction.

How to calculate CSAT

CSAT = (Number of satisfied responses / Total responses) × 100

Example: 430 customers rated the interaction 4 or 5 out of 5. Total responses: 540. CSAT = (430 / 540) × 100 = 79.6%.

Industry benchmark

A good CSAT score for B2B SaaS is 75-85%. Above 90% indicates excellent performance. Below 60% signals serious issues that need immediate attention. Most helpdesk platforms (Zendesk, Freshdesk, Dashly) calculate CSAT automatically from post-conversation surveys.

How to set the KPI for CSAT

Set a baseline from your last 90 days of data, then target a 5% improvement per quarter. Avoid setting a single company-wide CSAT target without accounting for channel differences: live chat typically scores 10-15 points higher than email because of the real-time nature of the interaction.

How to improve CSAT

The fastest CSAT levers are response speed and resolution quality. Slow responses are the single most common CSAT complaint across B2B support teams. Reduce response time by routing complex tickets to specialized agents and automating tier-1 queries. Resolution quality improves when agents have instant access to a well-maintained knowledge base during conversations. Survey timing also matters: sending the CSAT survey within 1 hour of ticket close increases response rates by 30-40%.

Net Promoter Score (NPS)

NPS measures long-term customer loyalty by asking one question: “How likely are you to recommend us to a colleague?” on a 0-10 scale. Unlike CSAT, which is transactional, NPS captures the cumulative effect of all support interactions over time. It is the metric most closely correlated with churn and expansion revenue in B2B SaaS.

How to calculate NPS

NPS = % of Promoters (score 9-10) minus % of Detractors (score 0-6)

Example: 200 responses. 100 Promoters (50%), 40 Detractors (20%), 60 Passives (30%). NPS = 50% — 20% = 30.

Industry benchmark

For B2B SaaS, an NPS above 30 is good, above 50 is excellent. The average NPS across software companies is around 31 (Retently 2024 benchmarks). A negative NPS (more detractors than promoters) is a strong churn warning signal.

How to set the KPI for NPS

Run NPS surveys quarterly and after major product releases or pricing changes. Segment results by account tier: enterprise accounts with an NPS below 20 should trigger an immediate customer success outreach, not just a survey follow-up email.

How to improve NPS

NPS improves slowly because it reflects accumulated experience. The highest-leverage tactics are: closing the loop with detractors within 48 hours of a survey response, resolving their issue, and documenting what caused it. Detractors who receive a personal follow-up convert to passives or promoters at a rate of 30-40% according to Bain data. Systematically fix the root causes behind detractor feedback rather than treating each response individually.

Customer Effort Score (CES)

CES measures how much effort a customer had to put in to get an issue resolved, typically on a 1-7 scale where 1 = very easy and 7 = very difficult. CES is the best predictor of repeat contact: customers who rate an interaction as high-effort are 4x more likely to contact support again within 30 days (CEB research). Use it to diagnose friction in specific processes like onboarding, billing, or account changes.

How to calculate CES

CES = Sum of all CES scores / Number of responses

On a 1-7 scale (1 = strongly agree the company made it easy), lower is better. A CES below 3 is excellent. Above 5 means customers find your support process genuinely painful.

Industry benchmark

Industry average CES: around 3.5-4.0 on a 7-point scale. Best-in-class teams score below 3. For context: the top quartile of B2B support teams have a CES under 2.5.

How to set the KPI for CES

Collect CES after specific high-friction interactions: complex troubleshooting, billing disputes, account migrations. Use it alongside CSAT, not as a replacement. CSAT tells you whether the outcome was satisfying, CES tells you whether the process was painful.

How to improve CES

Reduce the number of steps customers need to take to reach resolution. Common improvements: proactive status updates so customers do not need to chase for updates, self-service options for the most common request types (password resets, invoice copies, usage reports), and eliminating unnecessary verification steps that slow down straightforward requests.

NPS vs CSAT vs CES: which one to use when

All three metrics measure customer satisfaction but from different angles and timeframes. Using all three gives you a complete picture; relying on just one creates blind spots.

	NPS	CSAT	CES
What it measures	Long-term loyalty and advocacy	Satisfaction with a specific interaction	Ease of getting an issue resolved
Survey timing	Quarterly or after key milestones	After each support ticket closes	After a specific task or process
Scale	-100 to +100	1-5 or 1-10 (% satisfied)	1-7 (lower = better)
Industry benchmark	>30 B2B SaaS	>75%	<3
Best for	Predicting churn, expansion signals	Per-agent quality monitoring	Self-service and portal design

Use NPS for strategic account health. Use CSAT for day-to-day quality control. Use CES when you suspect a specific process (onboarding, billing, cancellation) is generating friction.

First Contact Resolution rate (FCR)

FCR measures the percentage of customer issues resolved in a single interaction, without the customer needing to follow up. It is one of the highest-ROI metrics to improve: SQM Group research shows that each 1% improvement in FCR reduces support operating costs by approximately 1% and increases CSAT by 1-3 points. For B2B SaaS, high FCR also reduces churn risk by eliminating the frustration of repeat contacts.

How to calculate FCR

FCR = (Issues resolved on first contact / Total issues) × 100

Example: 850 tickets resolved on first contact out of 1,200 total tickets in a month. FCR = (850 / 1200) × 100 = 70.8%.

Define “first contact” clearly before measuring. Most teams use: no follow-up contact on the same issue within 7 days after ticket close. Without a consistent definition, FCR numbers are not comparable across periods.

Industry benchmark

Industry average FCR: 70-75%. Best-in-class teams reach 80%+. Below 65% means a significant share of customers are contacting you multiple times for the same issue, compounding volume and cost. Phone support typically achieves higher FCR (75-80%) than email (65-70%) because of real-time back-and-forth.

How to improve FCR

The two main FCR killers are agents lacking the right information and agents lacking the authority to resolve issues without escalating. Fix both:

• Build and maintain a knowledge base agents can search in real time during a conversation. Agents who find answers in under 30 seconds resolve issues on first contact far more consistently.
• Give frontline agents authority to issue refunds, apply credits, and extend trials without manager approval up to a defined threshold (for example, $100 or 1 month of plan value). Most “need to escalate” situations cost more in repeat contacts than the concession itself.
• Use AI to classify incoming tickets and route complex issues directly to senior agents on first contact, rather than bouncing from tier-1 to tier-2. This is what AI support agents handle automatically for Dashly customers.

Churn Rate

Churn Rate measures the percentage of customers who stop using your product in a given period. For B2B SaaS, it is the single most important business impact metric because it determines whether growth compounds or bleeds out. Even a small improvement in churn has outsized revenue impact: a team with 5% monthly churn loses 46% of its customer base per year.

How to calculate Churn Rate

Churn Rate = (Customers lost in period / Customers at start of period) × 100

Example: 500 customers at the start of the month, 25 cancelled. Churn = (25 / 500) × 100 = 5%.

Industry benchmark

Healthy B2B SaaS churn: below 5% annually (0.4% monthly). Above 10% annually requires immediate attention. Enterprise segments typically achieve under 2-3% annual churn because of longer contracts and higher switching costs. SMB segments of 7-10% annually are common but signal room for improvement.

How to set the KPI for Churn Rate

Track churn monthly by cohort (by signup month, plan tier, and industry segment) rather than as a single global number. Global churn hides which customer segments are actually healthy. A company with 5% total churn might have 2% enterprise churn and 12% SMB churn — completely different problems requiring different solutions.

How to improve Churn Rate

Support-side churn reduction focuses on identifying at-risk accounts before they cancel. Warning signals include: multiple repeat contacts on the same issue, declining product login frequency, and NPS scores below 6. When these signals appear, proactive outreach from a customer success manager outperforms a reactive support ticket. Set up automated alerts to flag these patterns before the renewal date. See how to detect churn signals early with behavioral triggers.

Customer Retention Rate

Retention Rate is the inverse of Churn Rate. It measures the percentage of customers who stay with you over a given period. While churn focuses attention on what you are losing, retention frames the same data as what you are protecting. Both views are useful. Retention is typically reported annually for investor and board reporting; churn is reported monthly for operational management.

How to calculate Retention Rate

Retention Rate = ((Customers at end of period — New customers acquired) / Customers at start of period) × 100

Or simply: Retention Rate = 100% minus Churn Rate (for the same period).

Industry benchmark

B2B SaaS target: above 90% annual retention. Best-in-class enterprise SaaS: 95-97% annual retention. Below 85% annually signals a systemic problem with product-market fit or onboarding, not just support quality.

How to improve Retention Rate

Retention improvements compound. Raising retention from 85% to 90% does not just add 5% more customers per year — it changes the trajectory of your entire customer base over 3-5 years. The highest-leverage retention investments for support teams are faster issue resolution (reducing frustration-driven churn), proactive education during the first 90 days (reducing “didn’t understand the product” churn), and executive business reviews for enterprise accounts at 90-day intervals.

First Response Time (FRT)

FRT measures the time between a customer submitting a support request and the first human response from your team. It is the metric customers feel most immediately and comment on most often in CSAT surveys. A slow first response creates the impression of disorganization even if the eventual resolution is perfect. FRT does not measure quality — only speed — so always track it alongside CSAT.

How to calculate FRT

FRT = Total first response time across all tickets / Number of tickets in the period

Example: 300 tickets in a week, total first response time of 900 hours. Average FRT = 3 hours.

Industry benchmark

Email support: under 4 hours (best-in-class under 1 hour). Live chat: under 1 minute. Social media: under 1 hour. Phone: immediate or under 2 minutes hold time. Benchmarks vary significantly by channel; report FRT separately for each channel you operate.

How to set the KPI for FRT

Set FRT targets per channel and per account tier. Enterprise accounts typically have contractual SLAs (for example, 1-hour FRT in business hours). SMB accounts might have a 4-hour target. Treat a breach of an enterprise SLA as a P1 incident, not just a statistic to review at month-end.

How to improve FRT

The fastest FRT improvements come from automation, not hiring. AI agents and chatbots can acknowledge a ticket instantly and resolve tier-1 queries (password resets, status updates, basic FAQ) before a human needs to engage. For complex tickets requiring a human, smart routing ensures the right agent sees it immediately rather than sitting in a general queue. Using saved replies for recurring issue types cuts average handle time and indirectly improves FRT by freeing agents to start new conversations sooner.

Average Handle Time (AHT)

AHT measures the average time an agent spends on a single interaction, including active conversation time, hold time, and post-interaction wrap-up (documenting notes, updating the CRM). It is the primary efficiency metric for voice and chat support. High AHT inflates cost-per-ticket and limits how many customers each agent can serve. But artificially low AHT — caused by rushing agents — correlates with lower FCR and CSAT. The goal is the right AHT, not the lowest one.

How to calculate AHT

AHT = (Total talk time + Total hold time + Total wrap-up time) / Total number of interactions

Example: 200 calls in a day. Total talk time: 1,400 minutes. Total hold: 200 minutes. Total wrap-up: 400 minutes. AHT = (1,400 + 200 + 400) / 200 = 10 minutes per call.

Industry benchmark

Phone support AHT average: 6 minutes (ICMI benchmark). Chat support: 8-10 minutes. Email: tracked as resolution time, not handle time. Benchmarks vary heavily by industry and product complexity. A SaaS company with a complex product should expect AHT 20-30% higher than a consumer e-commerce brand.

How to set the KPI for AHT

Segment AHT by issue type before setting targets. Billing disputes have different natural AHT than password resets. A single global AHT target punishes agents handling genuinely complex issues and rewards those who cherry-pick simple tickets. Use AHT as a coaching metric for individual agents, not as a competitive leaderboard.

How to improve AHT

Reduce wrap-up time with AI-assisted note-taking and CRM auto-fill. Reduce hold time with real-time access to account data during the conversation — agents who need to put customers on hold to look up information extend AHT significantly. Improve talk time efficiency with clear escalation paths so agents do not spend time attempting to resolve issues above their tier. Regular call recording reviews (twice per month per agent) identify specific habits that inflate handle time.

Average Resolution Time (ART)

ART is different from AHT. AHT measures time spent in a single interaction. ART measures total elapsed time from when a customer first contacts you to when the issue is fully closed, even if it takes multiple interactions over several days. A ticket that requires three exchanges over 48 hours has a low AHT per interaction but a high ART overall. ART is the metric customers experience — they care how long their problem takes to go away, not how fast each individual reply arrives.

How to calculate ART

ART = Total time to resolve all tickets in period / Number of tickets resolved in period

Count from the timestamp of the first customer message to the timestamp when the ticket is marked resolved (not just “waiting for customer”). Exclude tickets where resolution is blocked by the customer’s own slow responses — or track “business hours” resolution time separately from calendar time.

Industry benchmark

Email/ticket support: best-in-class under 24 hours. Good: under 48 hours. B2B enterprise (complex technical issues): under 8 business hours for standard issues; under 4 hours for priority. Phone/chat single-session: under 10 minutes per session.

How to improve ART

ART is primarily reduced by improving FCR. When issues are resolved on the first contact, ART equals AHT. When tickets require multiple exchanges, ART can be 10-20x higher than AHT. Additional levers: clear ticket ownership (one agent owns a ticket from open to close, not a rotating queue), automated status updates to customers so they do not send follow-up emails asking “any update?” which reset the SLA clock, and automation for common multi-step resolution workflows.

Customer Lifetime Value (CLV)

CLV estimates the total revenue a customer generates over their entire relationship with your company. It connects support investment directly to business value: improving CSAT or reducing churn by 1% is worth more for a customer base with high CLV. Support teams use CLV to prioritize high-value accounts for proactive outreach and premium SLAs.

How to calculate CLV

CLV = Average Annual Revenue per Customer × Average Customer Lifespan (in years)

Or, for SaaS with variable plans: CLV = Monthly Recurring Revenue per customer / Monthly Churn Rate.

Example: $500 MRR per customer, 2% monthly churn. CLV = $500 / 0.02 = $25,000.

Industry benchmark

CLV varies by product and segment. A useful rule of thumb for B2B SaaS: a healthy CLV: CAC ratio is 3:1 or higher. If your average CLV is $10,000, your Customer Acquisition Cost should be under $3,333. Support teams that reduce churn extend CLV without the sales team needing to acquire new customers.

How to set the KPI for CLV

Segment CLV by customer tier (enterprise, mid-market, SMB) and use these segments to set support SLAs. High-CLV accounts justify premium response times and dedicated CSMs. Low-CLV accounts should be served efficiently through self-service and automated workflows to keep cost-per-ticket low.

How to improve CLV through support

Every reduction in churn extends CLV. Support’s most direct CLV lever is identifying expansion opportunities during service interactions. An agent who resolves a billing question can also notice the customer is near their usage limit and mention the next plan tier. This is not upselling — it is proactive advice that prevents the customer from hitting a frustrating wall and churning. Train agents to flag these opportunities and pass them to customer success, not to close them in the support channel.

Ticket volume and support backlog

Ticket volume is the total number of new support requests your team receives in a given period (day, week, or month). Backlog is the number of tickets open beyond your target resolution time. These two metrics are the foundation of capacity planning. Everything else — staffing decisions, automation investments, self-service strategy — follows from understanding your volume patterns.

How to use ticket volume for planning

Track volume by day of week and time of day to identify predictable peaks. Most B2B SaaS teams see Monday morning spikes (customers return from the weekend with backlogged issues) and drops on Friday afternoons. Use these patterns to stagger agent shifts rather than staffing uniformly across the week.

Rough staffing formula: if your weekly volume is 400 tickets and average AHT is 12 minutes, you need approximately 80 agent-hours per week (400 × 12 min / 60). At 6 productive hours per 8-hour agent day, that is around 14 agent-days per week. Add 20% buffer for complexity variance.

How to reduce backlog

A growing backlog is often a symptom of insufficient self-service, not insufficient headcount. Before hiring, audit your top 10 ticket types by volume. If 3-4 of them are answerable with a well-placed knowledge base article or an automated bot response, deploying self-service for those types can reduce inbound volume by 20-30% without additional staff. Real examples of automated customer service show the deflection rates achievable with this approach.

Cost Per Ticket (CPT)

Cost Per Ticket is total support costs divided by total tickets resolved in a period. It connects your operational metrics to financial performance and is essential for justifying automation or tooling investments. If deploying an AI agent costs $500/month and reduces your ticket volume by 200 tickets, the breakeven point is CPT above $2.50.

How to calculate CPT

CPT = Total support costs in period / Total tickets resolved in period

Include all costs: agent salaries, benefits, software licenses, management overhead. Exclude capital costs like hardware or one-time setup fees unless amortizing over a relevant period.

Industry benchmark

CPT varies widely: $2-$5 for simple e-commerce support, $15-$30 for complex B2B SaaS technical support, $40-$80 for enterprise with dedicated CSMs. The useful benchmark is your own trend over time: is CPT falling as you automate and scale, or rising as complexity increases? Automation investments are justified when they reduce CPT by more than their cost.

How to reduce CPT

CPT falls when volume grows faster than costs (scale) or when automation handles a portion of tickets without adding headcount. Focus automation on your highest-volume, lowest-complexity ticket types first. Automating 30% of tickets at CPT $0.50 each (bot cost) versus $20 each (human cost) produces significant savings quickly.

Frequently asked questions about customer service metrics