Popular Call Centre Metrics and KPIs
With contact centres being among the most measured workplaces on earth, there’s no shortage of popular call centre metrics to choose from. The trick isn’t finding metrics — it’s choosing the right ones, using them well, and understanding how they interact.
Because 74% of contact centre leaders start their careers on the phones, many inherit a legacy scorecard and targets without the why behind them. This article cuts through the noise so you can make confident, context-aware decisions about which metrics truly matter for your operation.
Below we’ll cover the most widely used KPIs, where they come from, and the common traps to avoid — so your scorecard drives performance, not behaviour you didn’t intend.
Contact Centre Industry Standards and Benchmarks
There’s no single global “standard” scorecard you can copy-paste. But there are two high-value sources in Australia that provide evidence and context for what “good” looks like — and how to use it intelligently:
- Annual benchmarking (self-reported performance): The Contact Centre Best Practice Report aggregates KPI and practice data from hundreds of centres. It’s useful for understanding trends, ranges and peer comparisons (e.g., GOS, AHT, FCR, occupancy, shrinkage).
- Independent service assessment (outside-in quality): The Australian Call Centre Rankings evaluate real customer experiences against the Australian Call Centre Quality Standards across 80+ metrics via mystery shopping. This shows how service is actually delivered, not just how it’s reported.
How to use both: benchmark reports tell you where you sit; service standards tell you what customers feel. Together, they help you pick KPIs that balance operational efficiency with customer outcomes.
Reality check: anyone promising a one-size-fits-all set of call centre metrics is selling simplicity, not performance. Your mix should reflect your channels, demand profile, technology, staffing model and customer expectations.
The Modern Contact Centre: More Than Calls
Today’s leaders juggle people, tech, budgets, analytics, compliance and stakeholder expectations — while delivering outcomes across multiple channels. It’s not just phone anymore. Typical portfolios include:
- Emails
- Live chat
- SMS and proactive messaging
- Written correspondence
- Social media engagement
- Chatbots/virtual agents and assistants
- Video or co-browse support (for some industries)
Each channel has different demand patterns, handle times and quality expectations — which means the “right” KPIs vary by mix. Keep that in mind as we step through the most popular metrics below.
Sincere Apologies…
Before we dive in, a confession: narrowing this list down to just ten was a tough call. There are dozens of important call centre metrics that play a big role in managing a contact centre effectively — some laser-focused on the contact centre itself, others more about the overall customer experience and sentiment.
Metrics that didn’t quite make this “top ten” cut (but are still well worth tracking) include:
- Adherence
- Absenteeism
- Hold Time
- Net Promoter Score (NPS)
- Customer Effort Score (CES)
- CSAT (Customer Satisfaction)
- Forecast Accuracy
- Employee Satisfaction
- Calls Blocked
We cover the majority of these in other resources — you’ll find definitions, pros/cons and usage tips in the ACXPA CX & Call Centre Glossary.
The KPIs Australian Contact Centre Managers Value Most
In the 2024 Australian Contact Centre Industry Benchmarking Report, managers were asked which metrics mattered most to them. Their top responses — in order — were:
- Abandonment Rate
- Customer Feedback
- Average Speed of Answer
- First Call Resolution
- Grade of Service / Service Levels
- Average Handling Time (AHT)
- After Call Work
- Shrinkage
- Sales / Conversion
It’s no surprise there’s overlap between this list and our top ten below — but remember, no one KPI tells the whole story. The skill lies in combining the right metrics for your mix of channels, demand, and business goals.
The Top Ten Call Centre KPIs and Metrics
Here’s what I believe are the most popular and widely used call centre metrics today. I can’t stress this enough: never judge performance by a single metric in isolation. The best leaders look for patterns, relationships and cause/effect.
Most KPIs on this list are linked — a shift in one can ripple through several others. Understanding those connections is where experienced managers make their mark.
Jump to a Metric
Grade of Service (aka Service Levels)
The first of our most popular call centre metrics is the Grade of Service (GoS) — also known as Service Levels. It’s usually defined as the percentage of calls answered within a predetermined number of seconds.
In Australia, we love to shorten everything, so within the contact centre industry you’ll often hear “GOS” paired with a KPI like 80/30 — meaning 80% of calls answered within 30 seconds.
It’s popular because it’s easy to understand and immediately tells you something about the customer experience you aim to provide. But here’s the truth: there’s no universal industry-standard KPI for Service Levels. While some sectors (e.g., emergency services, certain regulated industries, or government contracts) have mandated targets, in most cases your Service Level should be aligned to your specific business requirements.
Why your Service Level target should be unique
- Emergency services might aim for 100/5 — 100% of calls answered within 5 seconds — because lives depend on it.
- Sales-driven environments often aim high to minimise lost revenue from callers who hang up when kept waiting.
- Service or complaint lines may tolerate longer waits if resolution quality is the higher priority.
The critical point: the higher your Service Level target, the more resources you’ll need — and that gets expensive fast. For example, moving from an 80/30 to a 90/30 target can require 20–30% more staff, depending on call volumes and handle times. Model before you commit.
Industry Benchmarks for Service Levels
From the 2024 Australian Contact Centre Industry Best Practice Report, the five most common targets in Australia were:
- 80/20 – most popular in 2024
- 80/30
- 70/30
- 80/60
- 70/40
While 80/30 had historically been the most common, 2024 saw 80/20 move into the top spot — reflecting shifts in budgets, tech, and expectations.
How to find the right target for your centre
Use the interactive Erlang C Calculator to model scenarios like:
- If call volumes rise 10%, how many extra agents to maintain SLs?
- If AHT drops 30 seconds, how many fewer agents could you run?
- If budget is cut 15%, what happens to SL performance?
That calculator also previews ACXPA’s Workforce Management (WFM) Hub with advanced modelling tools used by leading centres.
Level up your capability
Average Speed of Answer (ASA)
Average Speed of Answer (ASA) is the average time it takes for answered calls to be answered by an agent. If calls are answered in 25 seconds on average, your ASA is 25 seconds.
Why this matters: relying only on Service Levels (e.g., 80/30) hides what happened to the other 20%. ASA fills in some of that picture — but it still has traps.
Key things many leaders miss
- ASA excludes abandons. Calls that hang up before an agent answers are not in the average. Pair ASA with Abandonment Rate or you’ll get a flattering half-truth.
- Averages hide pain. You can post a 25-second ASA and still have a long tail of callers waiting 5+ minutes. Use distribution views or percentiles (e.g., 80th/90th) to surface outliers.
- Define your clock start. Some centres start ASA when a caller joins queue; others include IVR/menu time. Pick a definition and use it consistently across reports.
Industry context & benchmarks
There’s no universal “right” ASA — the best target depends on your goals, demand pattern and budget. In the 2024 Australian Contact Centre Industry Best Practice Report , self-reported averages varied by function, for example:
- Customer Service Calls — ~115 seconds
- Inbound Sales — ~97 seconds
- Tech Support / Helpdesk — ~130 seconds
Note: those are self-reported. For an outside-in view, the Australian Call Centre Rankings capture actual average wait time from live mystery shops. In 2025 YTD, the industry average wait time is 1:53. That isn’t ASA (because it includes everyone, including abandons), but it’s a strong indicator of the experience customers feel.
How to use ASA well
- Track ASA alongside Service Levels, Abandonment Rate, and Occupancy to see cause-and-effect.
- Add a longest wait or 90th percentile wait to expose outliers hidden by the average.
- If ASA is creeping up, check staffing vs forecast, AHT trends, schedule adherence, IVR routing, and whether callbacks/virtual hold would reduce perceived wait.
Go deeper
- Learn the nuances in our glossary explainer: Average Speed of Answer (ASA).
- Model the impact of AHT, volume and staffing on ASA using our online Erlang C Calculator and explore the tools inside the WFM Hub.
- Build leadership capability with: Introduction to Contact Centre Management and Contact Centre Management Fundamentals.
Abandonment Rate
Abandonment Rate is the percentage of callers who enter a queue to speak with an agent but hang up before they’re connected.
People abandon for many reasons — long waits, frustration, or even the wrong kind of hold music. That article dives into the science behind why hold music affects customer patience, plus practical tips for improving it. Once you understand the psychology, you can make intentional choices that keep callers engaged longer.
If you decide to upgrade, you can find suppliers of professional call centre audio solutions — from high-quality hold music to voiceovers and IVR prompts — so your queue time sounds as professional as the service you aim to deliver.
In the 2024 Australian Contact Centre Industry Best Practice Report, Abandonment Rate was ranked the #1 most important metric by managers — because it’s often the clearest signal that your service levels, staffing, or routing need attention.
How to calculate it
Formula: Abandoned Calls ÷ Total Incoming Calls = Abandonment Rate
Example: If 100 calls join your queue in an hour and 10 hang up before speaking to an agent, your abandonment rate is 10%.
What influences abandonment rate
- Service Levels & Staffing: The longer people wait, the more likely they’ll hang up.
- Callback options: Offering virtual hold can reduce frustration and improve completion rates.
- Call purpose: Urgent or high-value calls tend to have lower tolerance for waiting.
- Queue experience: What customers hear while waiting — music, announcements, tone — can extend or shorten their patience.
When a higher abandonment rate isn’t bad
Not all abandons are negative. For example, if you play an IVR or hold message that resolves the customer’s need — like a recorded outage update — hanging up might mean they got what they needed without using an agent’s time.
If I call my power company and hear a message saying “We’re aware of the outage, power will be restored in 30 minutes,” I’ll hang up happy, not annoyed.
Industry benchmarks & 2025 rankings data
From the 2024 Best Practice Report: average abandonment rate was 8% (down from 9% in 2023). Self-reported average time before abandoning: 109 seconds.
From the Australian Call Centre Rankings (actual performance measurement) in 2025 YTD:
- 9.8% of calls industry-wide were still unanswered after 15 minutes.
- The worst sector: Banks — only 27.5% of calls answered within 15 minutes, including new business opportunities like home loan enquiries.
Reducing abandonment
- Model the impact of extra staff, shorter handle times, or improved routing using the Erlang C Calculator.
- Leverage WFM Hub tools to optimise schedules and forecast more accurately.
- Train leaders with courses like: Introduction to Contact Centre Management or Contact Centre Management Fundamentals.
Average Handle Time (AHT)
Average Handle Time (AHT) measures the average duration of a customer interaction from start to finish, including:
- Average Talk Time (ATT) — the time an agent spends speaking with the customer.
- After Call Work (ACW) — the time an agent spends after the call on wrap-up tasks, such as entering notes into the system.
Formula: Average Talk Time + Average After Call Work = Average Handle Time.
Example
If an agent spends 180 seconds talking to a customer and 60 seconds on ACW, the AHT is 240 seconds.
Why AHT matters — and when it doesn’t
While shorter AHT often means lower cost, it can also backfire. Rushing calls may hurt resolution, increase repeat contacts, and damage customer satisfaction. That’s why AHT should never be managed in isolation — instead, balance it with metrics like First Call Resolution and Call Quality.
With self-service and automation handling simple queries, the calls reaching agents today are typically more complex — which explains why AHT is increasing industry-wide.
What really drives AHT
It’s a mistake to treat AHT purely as an agent performance metric. Many factors that influence handle time are outside the agent’s control, such as:
- System speed and reliability
- Number of screens or applications required to resolve a query
- Mandatory scripts or compliance checks
- Process complexity or policy restrictions
- Availability of information from other departments
Improving AHT often means fixing processes and technology, not just coaching agents to “go faster.”
Industry trend
According to the 2024 Australian Contact Centre Industry Best Practice Report, average AHT across Australia rose to 543 seconds in 2024, up from 507 seconds in 2023.
Spotting outliers with the WFM Hub
One of the fastest ways to improve AHT is to identify individual or process outliers that skew your average. The ACXPA Workforce Management (WFM) Hub includes tools like the Outlier Detection Wizard, which visually flags agents or teams operating outside normal ranges so you can coach effectively and fairly.
Go deeper
- Learn more in our glossary explainer: Average Handle Time.
- Model the impact of staffing, call mix, and AHT changes using the Erlang C Calculator.
- Improve leadership capability with: Introduction to Contact Centre Management or Contact Centre Management Fundamentals.
Shrinkage
Shrinkage is the percentage of paid time that agents are not available to handle customer contacts because they’re in other activities — like training, meetings, or on leave. It’s a cornerstone metric for workforce planning because it directly affects how many staff you need to meet your Service Level targets.
Put simply: shrinkage tells you how much of your rostered time is “lost” to activities other than answering calls, emails, or chats.
Why shrinkage matters
- If you underestimate it, you’ll be under-staffed and queues will blow out.
- If you overestimate it, you’ll be over-staffed and waste budget on idle time.
- It’s a non-negotiable input for accurate Erlang calculations and roster modelling.
What’s included in shrinkage
- Training
- Coaching
- Meetings
- Annual leave
- Public holidays
- Paid breaks
- Sick leave / unplanned leave
- System downtime / operational blockers
Shrinkage % = (Total Paid Time Lost ÷ Total Paid Time) × 100
Example calculation
Annual terms for a full-time agent (261 paid days after weekends):
| Activity | Days lost |
|---|---|
| Annual leave | 20 |
| Sick leave | 10 |
| Public holidays | 8 |
| Training | 5 |
| Coaching | 5 |
| Team meetings | 2 |
| Total days lost | 50 |
Shrinkage = 50 ÷ 261 × 100 = 19%. Plan staffing for the available time, not 100% of paid time.
- Only counting leave and forgetting activities like training or coaching.
- Using a generic “industry average” shrinkage instead of calculating your own.
- Not updating assumptions when your business changes (new channels, more training, different rosters).
Related tools and learning
First Call Resolution (FCR)
First Call Resolution (FCR) is the percentage of customer enquiries that are fully resolved in a single interaction — with no need for the customer to contact you again about the same issue within an agreed time window.
- “Resolved” is subjective unless you define it precisely (by reason, channel and time window).
- Poorly designed targets can discourage helpful transfers or callbacks and actually increase repeat contacts.
- Many FCR drivers are outside the agent’s control (policies, processes, systems, authorisations).
Define FCR before you measure it
How to calculate FCR
FCR % = (Interactions with no repeat within X days ÷ Total eligible interactions) × 100
Example: 7,200 eligible calls this month. 5,220 had no repeat within 7 days. FCR = 72.5%.
Ask on the post-call survey: “Was your issue fully resolved today?” The % answering “Yes” is Survey FCR.
Best practice: use both. Operational shows behaviour; Survey shows perception. Any gap is diagnostic gold.
What really drives FCR
- Process & policy: Can agents actually fix it first time, or do rules force handoffs?
- System access & tools: Read/write access, knowledge, entitlement visibility.
- Authority & skills: Refund/credit limits, exception handling, diagnostic skill.
- Upstream quality: Poor forms, broken IVR routing, unclear comms create preventable repeats.
Coach agents, yes — but fix the system first. High FCR is a whole-of-business outcome.
Design to avoid gaming
- Report FCR alongside Call Quality and Customer Feedback so speed-only behaviours don’t win.
- Count warm transfers that resolve as FCR (or you’ll teach agents to bounce callers).
- Use reason-level FCR (e.g., “Password reset” vs “Complex billing discrepancy”) for fair comparisons.
- Publish a clear playbook for what counts/doesn’t — then calibrate monthly with QA and team leaders.
Practical implementation steps
- Create/clean your contact reason taxonomy (10–20 parent reasons max).
- Build repeat logic in reporting (Customer ID + reason + X-day window across all channels).
- Stand up a failure reasons tag set (why didn’t it resolve first time?).
- Target top failure modes with cross-functional fixes; track impact weekly.
Learn and improve
- Sharpen your approach in: Introduction to Contact Centre Management and Contact Centre Management Fundamentals.
- Benchmark your FCR and resolution performance with our Contact Centre Benchmarking services — outside-in assessments that surface repeat-contact drivers and improvement opportunities.
Occupancy
Occupancy is the percentage of time agents spend actively handling customer contacts (talk + hold + After Call Work) compared with the total time they’re logged in and available for work. It’s a core measure of efficiency.
Understanding the balance
- Low occupancy: Too many staff rostered — you’re paying for idle time.
- High occupancy: Not enough staff — long waits, stressed agents, quality slips.
Most centres aim for 80–85% as the sustainable sweet spot.
How to calculate occupancy
Occupancy % = (Contact Handling Time ÷ (Contact Handling Time + Available Time)) × 100
Example: Agents spend 80 hours this week on talk/hold/ACW and have 20 hours of available time.
Occupancy = 80 ÷ (80 + 20) × 100 = 80%
Trade-offs & pitfalls
- Higher Service Levels usually mean lower occupancy (more agents ready).
- Sustained 90%+ occupancy drives burnout and repeat contacts — expect quality drop.
- Measure by interval (e.g., 15/30 mins) — daily averages hide peak pain.
- Multi-skill routing and long ACW can distort occupancy if definitions aren’t consistent.
- Fix root causes (forecast error, adherence, AHT, IVR leakage) before pushing targets.
Cost per Call (or Cost per Contact)
Cost per Call is the average cost of handling each interaction. It’s great for internal decision-making (targets, business cases, channel mix) — but it’s rarely apples-to-apples across organisations.
How to calculate
Cost per Call = Total Operating Cost ÷ Total Calls (or Contacts)
Example: Total operating cost = $1,000,000. Calls handled = 40,000.
$1,000,000 ÷ 40,000 = $25 per call
What to include (fully loaded)
- Salaries & on-costs (agents, leaders, WFM, QA)
- Technology/licences (telephony, CRM, WEM, QA)
- Premises (rent, utilities) or WFH stipends
- Training, coaching, recruitment
- Vendor fees / BPO charges
- Shared overheads (allocate sensibly)
Use it well
- Track trends over time; the direction is more useful than the absolute number.
- Pair with FCR and Quality — cheaper isn’t better if issues repeat.
- Look at Cost per Resolved Contact for a truer efficiency view.
- Split fixed vs variable costs; know your marginal cost when volumes move.
- Report by channel (call, chat, email) — they have very different cost profiles.
Common pitfalls
- Benchmarking blindness: “Industry average” is mostly meaningless without identical scope, mix and cost allocation rules.
- Allocation games: Dumping enterprise overhead into the centre can make your cost look unjustifiably high.
- Per-agent targets: Don’t weaponise cost per call at the agent level — it’s a planning metric, not a performance KPI.
Employee Attrition
Employee Attrition (aka turnover) measures the percentage of your workforce that leaves during a given period. In contact centres, it’s often seen as an indicator of engagement, leadership effectiveness, and organisational health — but not all attrition is bad.
Types of attrition
- Voluntary: Employees choose to leave.
- Involuntary: You ask employees to leave.
How to calculate
Attrition % = (Number who leave ÷ Average headcount) × 100
Example: 100 FTE, 12 leavers in 12 months = 12% annual attrition (≈ 1 person per month).
When attrition can be positive
- Underperformers exiting — helps lift overall capability.
- Internal moves to other roles in the organisation — good for the business, though it may hurt the contact centre in the short term.
When attrition is negative
- Loss of talent to competitors or other industries.
- Driven by poor leadership, culture, or working conditions.
- Linked to lack of development or career pathways.
Why high attrition is costly
- Recruitment and training costs rise.
- Lower productivity until new hires reach proficiency (often 6+ months).
- Customer experience inconsistency (varies by agent experience).
- Morale drop as remaining staff pick up the workload.
Why very low attrition can also hurt
- Stagnation — harder to drive change.
- Costs increase as wages rise with tenure.
- Risk of “system workarounds” becoming embedded.
Industry snapshot
Australian contact centre attrition has trended down from 32% in 2023 to 27% in 2024. Larger centres (500+ seats) average above 43%, while smaller centres tend to see lower rates but can feel the loss more personally.
Call Quality
Call Quality is how you define, measure and improve the standard of service your agents deliver. Everyone agrees it matters — where teams stumble is turning “quality” into something objective, calibrated and actionable.
Two types of quality metrics
- Compliance metrics — checks whether agents follow defined processes, scripts, legal/regulatory requirements, and internal policies (e.g. ID verification, mandatory disclosures).
- Performance metrics — assesses the skill and effectiveness of the interaction, such as rapport-building, problem-solving, and delivering a positive customer outcome.
Many call centre “quality” programs only measure compliance — which can miss the bigger picture of the customer’s experience.
Start with a clear standard
Anchor your program to an agreed framework so “quality” isn’t subjective. We use the Australian Call Centre Quality Standards — a practical outside-in view of what great looks like. See also our Quality Assurance glossary entry for tips on setting up a QA program.
Typical components (customise to suit)
- Greeting & identification
- Discovery & active listening
- Ownership & next steps
- Accuracy & compliance
- Resolution effectiveness
- Communication & tone
- Closing & confirmation
- Security/ID checks (where relevant)
Scoring & calibration
- Use behavioural rubrics (what “good” looks like) instead of vague 1–5 stars.
- Run regular calibration between QA, TLs and operations so scores mean the same thing everywhere.
- Track a Quality Score % (points achieved ÷ points available) and show rationale in coaching notes.
How many calls should you assess?
There’s no global standard. In practice, 5–15 calls per agent per month (random, stratified by reason/queue) gives useful coaching signals without drowning the team. Use more when you’re new, in remediation, or changing process.
The rise of AI in QA (what to know)
- Pros: huge coverage (100% of interactions), faster insights, objective pattern spotting, searchable transcripts.
- Requirements: a solid rubric mapped to your standards, human calibration loops, and privacy/consent settings nailed.
- Reality check: AI can surface issues at scale; coaches still turn those into behaviour change.
Independent assessments & benchmarking
- For an external perspective, use our Call Quality Assessment service — we assess your calls against the Australian Call Centre Quality Standards and provide targeted improvement actions.
- Explore broader measurement programs: Contact Centre Benchmarking (overview), Contact Centre CX Benchmarking, Monthly Snapshot Service, and learn how to get your Call Centre Ranking.
Training to build a QA framework
Want to create a quality program that goes beyond compliance? Our Quality Framework for Contact Centres course shows you how — ACXPA Members receive 25% off.
Tools & suppliers
Exploring QA and CX tech? Browse our directories: Call Centre Technology or CX Technology.
- About the Author
- Latest Articles
After spending over 30 years working in contact centres and CX, one thing I’ve learnt is there is always something more to learn!
I’m thrilled to be the inaugural CEO of ACXPA, and together with the rest of the team, we’re focused on helping Australian businesses deliver efficient and effective customer experiences via phone, digital and in-person by empowering their employees with the skills, industry insights and professional support networks they need to succeed.
