Erlang C Calculator & Formula Explained

Erlang C Calculator

Erlang C Calculation Results

Erlang C Formula (Simplified Explanation):
The Erlang C formula calculates the probability that a call arriving at a contact center will have to wait in a queue. It considers the number of agents available, the average time it takes to handle a call, and the rate at which calls arrive.

Key Components:

• Traffic Intensity (A): The total call load in Erlangs (calls per hour * AHT in hours).
• Number of Agents (E): The number of available agents.
• Erlang C (P_delay): The probability a call experiences a delay.

The precise mathematical formula involves complex summations and factorials, often implemented via iterative algorithms or lookup tables for practical use. This calculator uses an approximation method suitable for typical call center volumes.

What is the Erlang C Formula?

The Erlang C formula is a cornerstone in contact center and telecommunications management, providing critical insights into queueing theory. It’s a mathematical model used to predict the probability that an incoming call will have to wait in a queue before being answered by an agent. This metric, often referred to as the Erlang C probability of delay, is fundamental for staffing calculations, ensuring that a contact center has the right number of agents to meet customer service level objectives.

Who Should Use the Erlang C Formula?

Anyone responsible for operational efficiency and customer experience in a high-volume inbound contact center environment should leverage the Erlang C formula. This includes:

• Call Center Managers: For workforce management, scheduling, and forecasting staffing needs.
• Operations Directors: To optimize resource allocation and control costs while maintaining service quality.
• Customer Experience (CX) Professionals: To understand how staffing levels impact customer wait times and satisfaction.
• Telecommunications Engineers: Designing and managing call routing and queuing systems.
• Business Analysts: Evaluating the impact of operational changes on key performance indicators (KPIs).

Common Misconceptions about Erlang C

Several misconceptions surround the Erlang C formula:

• It accounts for agent unavailability: Erlang C assumes all agents are available. It does not inherently factor in breaks, training, or sickness. Separate calculations are needed for these scenarios.
• It predicts exact wait times: Erlang C provides probabilities and averages. Actual wait times can fluctuate significantly due to call arrival patterns and call duration variability.
• It handles all call types: The standard Erlang C model is for a single queue with homogeneous agents. Complex multi-skill routing or blended agent environments require more advanced models.
• Lower Erlang C is always better: While a low probability of delay is desirable, striving for zero delay might lead to overstaffing and unnecessary costs. The goal is to find an optimal balance.

Understanding these nuances is crucial for accurate application of the Erlang C calculator.

Erlang C Formula and Mathematical Explanation

The Erlang C formula is derived from queueing theory and helps model the behavior of a telecommunications system with multiple servers (agents) and a single queue. The core idea is to balance the number of incoming calls with the capacity of the available agents.

The Derivation and Variables

The formula calculates the probability that all agents are busy when a call arrives, forcing it into a queue. While the full derivation involves concepts like Poisson processes for arrivals and exponential distributions for service times, leading to complex summations of factorials, we can understand the key components and their impact.

The formula for Erlang C (P_delay) is typically expressed as:

Erlang C = P(Delay) = _{(A^E / E!) * (E / (E – A))} / _{Σ^∞n=0 (Aⁿ / n!) + (A^E / E!) * (E / (E – A))}

Where:

• A = Traffic Intensity in Erlangs (average arrival rate × average handle time).
• E = Number of Agents (servers).
• n = Number of calls in the system.
• Aⁿ / n! = Term in the summation representing the probability of n calls in the system assuming infinite agents.
• A^E / E! = The term for exactly E calls in the system.
• E / (E – A) = The multiplier for the queueing probability when all agents are busy.

Note: The summation usually converges quickly, and for practical purposes, it’s often truncated or calculated iteratively.

Variables Table

Erlang C Formula Variables

Variable	Meaning	Unit	Typical Range
A (Traffic Intensity)	Total call load placed on the agents.	Erlangs	0.1 – 50+ (highly variable)
E (Number of Agents)	The number of agents available to handle calls.	Count	1 – 100+
Pdelay (Erlang C)	Probability a call experiences a delay (waits in queue).	Probability (0 to 1)	0 – 1
AHT (Average Handle Time)	Average time spent on a call (talk + hold + wrap-up).	Seconds / Minutes	30 – 1200 (seconds)
Arrival Rate (λ)	Average number of calls per unit time.	Calls per Hour / Minute	1 – 1000+ (per hour)
ASA (Average Speed of Answer)	Average time a caller waits in queue before being answered.	Seconds	1 – 600 (seconds)
Service Level Target	Desired % of calls answered within X time.	% & Minutes	e.g., 80% in 30 seconds

Practical Examples (Real-World Use Cases)

Let’s explore how the Erlang C calculator can be applied in different scenarios:

Example 1: Optimizing Staffing for a Customer Support Center

Scenario: A small e-commerce company’s customer support center receives an average of 150 calls per hour. The Average Handle Time (AHT) is 4 minutes (240 seconds). They aim to achieve an 80% Service Level, meaning 80% of calls should be answered within 30 seconds. Currently, they have 10 agents.

Inputs:

• Target Service Level: 30 seconds
• Number of Agents (E): 10
• Average Handle Time (AHT): 240 seconds
• Arrival Rate: 150 calls/hour
• Service Level Target: 80%

Calculation using the calculator:

• Traffic Intensity (A) = (150 calls/hour * 4 minutes/call) / 60 minutes/hour = 10 Erlangs
• Erlang C (Probability of Delay) ≈ 0.25 (or 25%)
• Calls Delayed ≈ 150 * 0.25 = 37.5 calls per hour
• Average Speed of Answer (ASA) ≈ 110 seconds
• Service Level Achieved ≈ 45%

Interpretation: With 10 agents, 25% of calls are expected to wait, resulting in an average wait time of over 1.5 minutes. Crucially, the Service Level target of 80% in 30 seconds is significantly missed (only 45% achieved). This indicates understaffing.

Action: The manager would increase the number of agents using the calculator until the Service Level Achieved meets or exceeds the 80% target. For instance, increasing to 13 agents might yield a much higher Service Level.

Example 2: Evaluating Impact of Increased Call Volume

Scenario: A bank’s call center handles 400 calls per hour with an AHT of 3 minutes (180 seconds). They have 30 agents and a target of answering 90% of calls within 20 seconds.

Inputs:

• Target Service Level: 20 seconds
• Number of Agents (E): 30
• Average Handle Time (AHT): 180 seconds
• Arrival Rate: 400 calls/hour
• Service Level Target: 90%

Calculation using the calculator:

• Traffic Intensity (A) = (400 calls/hour * 3 minutes/call) / 60 minutes/hour = 20 Erlangs
• Erlang C (Probability of Delay) ≈ 0.05 (or 5%)
• Calls Delayed ≈ 400 * 0.05 = 20 calls per hour
• Average Speed of Answer (ASA) ≈ 15 seconds
• Service Level Achieved ≈ 92%

Interpretation: With 30 agents, the traffic intensity is 20 Erlangs. The system is adequately staffed, with only 5% of calls delayed, an average wait of 15 seconds, and the Service Level target of 90% in 20 seconds is met.

What If? If a marketing campaign is expected to increase call volume by 25% (to 500 calls/hour), the manager would re-run the calculation with E=30 and Arrival Rate=500. The new Traffic Intensity becomes 25 Erlangs. The calculator would show a significantly higher Erlang C, longer ASA, and a missed Service Level, prompting the need for more agents.

How to Use This Erlang C Calculator

Our Erlang C calculator is designed for ease of use, providing immediate insights into your contact center’s performance. Follow these steps:

1. Input Key Metrics:
   • Target Service Level (Minutes): Enter the duration (e.g., 30 seconds, which is 0.5 minutes) within which you aim to answer a specific percentage of calls.
   • Number of Agents (E): Input the current or planned number of agents handling the queue.
   • Average Handle Time (AHT) (Seconds): Enter the average duration of a customer interaction, including talk time, hold time, and post-call work.
   • Arrival Rate (calls per hour): Input the average number of incoming calls received per hour.
   • Service Level Target (%): Enter the desired percentage of calls to be answered within the specified Target Service Level (e.g., 80 for 80%).
2. Click “Calculate Erlang C”: The calculator will process your inputs using the Erlang C formula and related calculations.
3. Review the Results:
   • Probability of Delay (Erlang C): The primary output, showing the likelihood a caller will wait. Lower is generally better, but balance cost.
   • Calls Delayed (per hour): An estimate of how many calls per hour will experience a wait.
   • Probability of Abandonment (Approx.): A rough estimate, as longer waits increase the chance callers hang up.
   • Average Speed of Answer (ASA) (seconds): The predicted average wait time for callers who *do* wait.
   • Service Level Achieved (%): The calculated percentage of calls answered within your defined target time, based on your inputs. Compare this to your target.
4. Interpret and Adjust: If the “Service Level Achieved” is below your target, you likely need more agents. Use the “Erlang C Simulation Table” and “Erlang C & Service Level Chart” to see how adding agents impacts performance. Use the “Reset” button to start fresh. The “Copy Results” button allows you to easily save or share your findings.

Key Factors That Affect Erlang C Results

Several interconnected factors influence the outcomes predicted by the Erlang C formula and, consequently, your contact center’s performance. Understanding these is key to effective management:

1. Arrival Rate (Calls per Hour): This is a primary driver. Higher call volumes directly increase the traffic intensity (A), making it more likely that agents will be busy. Peaks and troughs in call arrival patterns (e.g., seasonal, daily, or campaign-driven) significantly impact staffing needs. Accurate forecasting is essential.
2. Average Handle Time (AHT): A longer AHT means each agent is occupied for a longer duration per call. This reduces agent availability for subsequent calls, increasing the traffic intensity and the probability of delay. Optimizing AHT (without sacrificing quality) by improving agent training, call scripting, or knowledge base resources can significantly reduce staffing requirements.
3. Number of Agents (E): This is the most direct control variable. Increasing the number of agents directly reduces the probability of delay (Erlang C) and improves the Average Speed of Answer (ASA) and Service Level. However, more agents mean higher personnel costs, so finding the optimal number is crucial.
4. Service Level Target: A more aggressive Service Level target (e.g., 90% of calls in 10 seconds) requires significantly more agents than a less stringent one (e.g., 70% in 60 seconds) for the same call volume and AHT. The target directly influences the required Erlang C value.
5. Shrinkage: This refers to the percentage of paid time that agents are not available for handling calls due to breaks, lunches, training, meetings, coaching, and unexpected absences (sick leave). Standard Erlang C calculations assume 100% availability. You must factor in shrinkage by calculating the *effective* number of agents available and potentially adjusting staffing targets upwards to compensate. For example, if you have 100 agents scheduled but 30% shrinkage, you only have 70 agents effectively available.
6. Call Abandonment Rate: While Erlang C primarily focuses on delay probability, high wait times increase the likelihood of callers abandoning their calls before reaching an agent. This isn’t directly part of the Erlang C calculation but is a critical consequence. Modeling abandonment often requires additional formulas or simulations based on the calculated ASA.
7. Occupancy Rate: This measures the percentage of time agents spend actively handling calls or in wrap-up activities versus their available time. While Erlang C doesn’t directly calculate occupancy, very high occupancy rates (e.g., > 85-90%) often indicate agent burnout and potential quality issues, even if Service Level targets are met. Managing occupancy is key to agent retention and sustained performance.

Frequently Asked Questions (FAQ)

• What is the difference between Erlang C and Service Level?
  Erlang C is the probability that a call will have to wait in the queue. Service Level is a performance target, typically expressed as a percentage of calls answered within a specific time threshold (e.g., 80% of calls answered in 30 seconds). Erlang C helps determine the staffing needed to *achieve* a target Service Level.
• Does Erlang C account for call abandonment?
  The basic Erlang C formula does not directly calculate abandonment. However, the predicted Average Speed of Answer (ASA) derived from Erlang C can be used to estimate abandonment rates, as longer waits significantly increase the probability callers will hang up. More advanced models exist for integrated abandonment calculation.
• How do I use Erlang C results for scheduling?
  Use Erlang C to calculate the base number of agents needed during specific periods. Then, overlay anticipated call volume variations (e.g., time of day, day of week) and adjust for shrinkage (breaks, training) to create detailed agent schedules that meet service level goals cost-effectively.
• What is considered a “good” Erlang C value?
  There’s no single “good” value, as it depends on business strategy and cost tolerance. Generally, an Erlang C below 0.1 (10% probability of delay) indicates ample staffing for that specific traffic load. However, achieving very low Erlang C might be cost-prohibitive. The goal is to find a balance that meets Service Level targets efficiently.
• Can Erlang C be used for outbound call centers?
  The standard Erlang C model is designed for inbound call centers handling unpredictable arrival patterns. While principles can be adapted, direct application is limited. Outbound dialing often uses different models focusing on dialer efficiency and right-party connection rates.
• What if my call arrival isn’t a Poisson process?
  The Erlang C formula assumes arrivals follow a Poisson distribution. If your arrival patterns are highly non-random (e.g., predictable batch calls from IVR callbacks), the Erlang C prediction might be less accurate. More sophisticated queueing models or simulations might be necessary for higher precision in such cases.
• How does AHT variability affect the results?
  Erlang C assumes a constant average AHT. Significant variability in AHT can impact actual performance. If AHT is highly variable, some calls might take much longer, tying up agents and increasing wait times beyond what the average suggests. This is another area where simulations can provide deeper insights than basic Erlang C.
• Is there a limit to the number of agents Erlang C can calculate for?
  Mathematically, the formula can handle large numbers of agents. Practically, computational precision might become a factor for extremely high agent counts or traffic intensities. Most modern calculators and software libraries handle typical ranges (up to hundreds of agents) effectively.
• Can I use this calculator for agent effectiveness, not just staffing?
  While the calculator primarily focuses on staffing based on arrival rates and AHT, you can infer agent effectiveness. If you consistently need more agents than anticipated based on industry benchmarks for a given call volume and AHT, it might suggest your AHT is too high, or your agents require more training or better tools. Conversely, needing fewer agents might indicate high efficiency.

Related Tools and Internal Resources