Erlang Calculator
Calculate the required number of agents, traffic intensity, and service level for your call center using the Erlang C formula. This tool helps optimize staffing based on call volume, average handling time, and target service levels.
Erlang Calculator: Complete Guide to Call Center Staffing Optimization
The Erlang calculator is an essential tool for call center managers and workforce planners. Based on the Erlang C formula developed by Danish mathematician Agner Krarup Erlang, this calculator helps determine the optimal number of agents required to handle incoming calls while maintaining desired service levels.
What is the Erlang Formula?
The Erlang formula is a mathematical model used to predict call center performance. There are two main variations:
- Erlang B: Used for systems with blocked calls cleared (callers get a busy signal)
- Erlang C: Used for systems with delayed calls (callers enter a queue)
For call centers where calls are queued when all agents are busy, the Erlang C formula is most appropriate. The formula calculates:
- The probability that a call will have to wait
- The average waiting time
- The number of agents required to achieve a specific service level
Key Components of the Erlang C Formula
The formula incorporates several critical variables:
- Call arrival rate (λ): Number of calls per time unit
- Average handling time (AHT): Time to complete a call (including talk time and after-call work)
- Number of agents (N): Staff available to handle calls
- Service level target: Percentage of calls answered within a specific time
| Variable | Description | Example Value |
|---|---|---|
| λ (Lambda) | Call arrival rate (calls per hour) | 120 calls/hour |
| AHT | Average handling time (seconds) | 180 seconds (3 minutes) |
| N | Number of agents | 15 agents |
| Service Level | Percentage of calls answered within target time | 80% in 20 seconds |
How the Erlang Calculator Works
The calculator performs several key calculations:
1. Traffic Intensity Calculation
Traffic intensity (A) is calculated by:
A = (λ × AHT) / 3600
Where:
- λ = call arrival rate (calls per hour)
- AHT = average handling time (in seconds)
- 3600 = number of seconds in an hour
Example: For 120 calls per hour with 180-second AHT:
A = (120 × 180) / 3600 = 6 Erlangs
2. Agent Occupancy
Agent occupancy shows what percentage of time agents are busy:
Occupancy = (A / N) × 100
Where:
- A = traffic intensity
- N = number of agents
Ideal occupancy is typically between 70-85%. Higher occupancy leads to longer wait times.
3. Service Level Calculation
The service level is the percentage of calls answered within a target time. The Erlang C formula calculates the probability that a call will wait longer than the acceptable time.
The formula is complex and typically requires computational methods:
P(W > t) = (C(N,A) × e-(N-A)t/AHT) / (1 + (1-A/N) × (C(N,A)-1))
Where C(N,A) is the Erlang C function.
Practical Applications of Erlang Calculations
Staffing Optimization
Determine the exact number of agents needed for different times of day, ensuring cost efficiency while maintaining service quality.
Budget Planning
Accurately forecast staffing requirements for budgeting purposes, balancing service quality with operational costs.
Performance Benchmarking
Compare actual performance against Erlang predictions to identify operational inefficiencies or training needs.
Real-World Example
A mid-sized call center receives 500 calls per hour with an average handling time of 240 seconds. They want to answer 80% of calls within 20 seconds. Using the Erlang calculator:
- Traffic intensity = (500 × 240) / 3600 = 33.33 Erlangs
- Required agents = 42 (to achieve 80/20 service level)
- With 30% shrinkage, total staff needed = 55
- Agent occupancy = 79.4%
Common Mistakes in Erlang Calculations
- Ignoring shrinkage: Forgetting to account for breaks, training, and absenteeism (typically 20-40%)
- Using incorrect AHT: Not including after-call work time in average handling time calculations
- Overlooking call patterns: Applying the same staffing levels throughout the day instead of adjusting for peak times
- Misinterpreting service levels: Confusing “80% of calls answered in 20 seconds” with “average speed of answer is 20 seconds”
- Neglecting multi-channel contacts: Focusing only on calls while ignoring emails, chats, and other contact channels
Advanced Erlang Concepts
Erlang B vs Erlang C
| Feature | Erlang B | Erlang C |
|---|---|---|
| Call Handling | Blocked calls cleared | Calls enter queue |
| Typical Use Case | Systems where callers get busy signal | Call centers with wait queues |
| Key Metric | Blocking probability | Waiting probability and time |
| Staffing Impact | Requires fewer agents | Requires more agents for same service level |
| Customer Experience | Poor (callers can’t get through) | Better (callers can wait in queue) |
Multi-Skill Erlang Calculations
Modern call centers often use skills-based routing where agents have different skill sets. Advanced Erlang models can:
- Handle multiple call types with different AHTs
- Account for agents with multiple skills
- Optimize for blended environments (inbound + outbound)
- Incorporate priority routing
These models require more complex calculations and often use simulation software for accurate predictions.
Erlang in Omnichannel Environments
With the rise of digital channels, Erlang principles are being adapted for:
- Live chat: Using similar queuing theory with different handling time distributions
- Email: Modeling response time SLAs instead of immediate answer times
- Social media: Incorporating variable response time expectations
- Blended operations: Balancing real-time and asynchronous channels
Implementing Erlang in Your Call Center
Step-by-Step Implementation Guide
-
Data Collection:
- Gather historical call volume data (by 15-30 minute intervals)
- Measure accurate average handling times (including ACW)
- Determine current service levels and customer expectations
-
Tool Selection:
- Choose between spreadsheet-based calculators or dedicated WFM software
- Consider integration with your ACD system for real-time data
- Evaluate cloud-based vs on-premise solutions
-
Pilot Testing:
- Run calculations for a specific team or time period
- Compare predicted vs actual performance
- Adjust shrinkage factors based on real data
-
Full Implementation:
- Roll out to all teams with proper training
- Establish regular review cycles (weekly/monthly)
- Integrate with forecasting processes
-
Continuous Improvement:
- Monitor accuracy of predictions
- Adjust models as call patterns change
- Incorporate new channels as they’re added
Integration with Workforce Management
Erlang calculations form the foundation of modern Workforce Management (WFM) systems. Advanced WFM tools:
- Automate Erlang calculations using real-time data
- Incorporate machine learning to improve forecasts
- Provide intra-day management capabilities
- Offer “what-if” scenario planning
- Generate optimized schedules that balance efficiency and agent preferences
Industry Benchmarks and Standards
While optimal metrics vary by industry, here are some general benchmarks:
| Metric | Retail | Telecom | Financial Services | Healthcare |
|---|---|---|---|---|
| Service Level (X/Y) | 80/20 | 85/20 | 90/20 | 95/15 |
| Average Speed of Answer (seconds) | 28 | 22 | 18 | 12 |
| Agent Occupancy (%) | 80-85 | 75-80 | 70-75 | 65-70 |
| Shrinkage (%) | 25-30 | 30-35 | 35-40 | 40-45 |
| Average Handling Time (seconds) | 240 | 300 | 360 | 420 |
Note: These benchmarks are illustrative. Actual targets should be based on your specific customer expectations and business requirements.
Limitations of Erlang Calculations
While powerful, Erlang models have some limitations:
- Assumes random call arrivals: Doesn’t account for call spikes or patterns
- Fixed handling times: Assumes all calls take the same amount of time
- No agent skill differences: Basic models treat all agents as identical
- No call abandonments: Doesn’t account for customers hanging up
- Steady-state assumption: Assumes the system has been running long enough to reach equilibrium
For these reasons, many call centers use Erlang as a starting point and then adjust based on real-world performance data.
Future Trends in Call Center Staffing
AI and Machine Learning
Emerging technologies are enhancing traditional Erlang models:
- Predictive staffing: AI analyzes multiple data sources to predict call volumes more accurately
- Real-time optimization: Machine learning adjusts staffing levels intra-day based on actual patterns
- Agent performance prediction: AI identifies which agents are likely to perform best during peak times
- Automated scheduling: AI generates optimized schedules considering hundreds of constraints
Cloud-Based Workforce Management
Cloud solutions are making advanced WFM accessible to businesses of all sizes:
- Real-time dashboards with Erlang calculations
- Mobile apps for agent self-service
- Integration with CRM and contact center platforms
- Automated forecasting with machine learning
- Global workforce optimization for distributed teams
The Rise of Gig Agents
The gig economy is changing call center staffing models:
- On-demand agents can be brought online during peak times
- Erlang calculations help determine when to activate gig workers
- Blended models combine full-time agents with part-time/gig workers
- New metrics needed to account for variable agent availability
Authoritative Resources on Erlang Calculations
For those seeking to deepen their understanding of Erlang theory and its applications in call centers, these authoritative resources provide valuable insights:
- National Institute of Standards and Technology (NIST) – Offers research on queuing theory and its applications in service systems. Their publications on telephone traffic engineering are particularly relevant.
- NIST/SEMATECH e-Handbook of Statistical Methods – Includes sections on queuing theory that underpin Erlang calculations.
- Columbia University Industrial Engineering and Operations Research – Their research on stochastic models and queuing systems provides advanced insights into the mathematical foundations of Erlang calculations.
These resources offer both theoretical foundations and practical applications of Erlang calculations in modern contact center environments.
Conclusion: Mastering Erlang for Call Center Excellence
The Erlang calculator remains one of the most powerful tools for call center optimization, nearly a century after its development. By understanding and properly applying Erlang principles, call center managers can:
- Significantly improve service levels and customer satisfaction
- Optimize staffing levels to balance cost and performance
- Make data-driven decisions about capacity planning
- Identify operational inefficiencies
- Prepare for seasonal fluctuations and growth
While the mathematical foundations are complex, modern tools make Erlang calculations accessible to any call center. The key to success lies in:
- Collecting accurate input data
- Regularly validating predictions against actual performance
- Adjusting models as business conditions change
- Combining Erlang with other workforce management techniques
- Continuously improving based on real-world results
As call centers evolve to handle multiple digital channels, the core principles of Erlang remain relevant. The future will likely see these models enhanced with AI and machine learning, but the fundamental relationship between call volume, handling time, and staffing requirements will continue to be governed by the queuing theory that Erlang pioneered.
By mastering Erlang calculations and their practical application, call center leaders can achieve the delicate balance between operational efficiency and exceptional customer service that defines world-class contact center performance.