Erlang C Isn't Broken — But Your Workforce Planning Might Be
Erlang C remains one of the most recognised and widely used workforce planning models in contact centres — and for good reason. It's simple, fast and still useful for estimating staffing requirements in traditional voice environments.
But while Erlang C isn't broken, many contact centres now apply it to environments it was never designed for. Voice is no longer the only workload. Channels behave differently, concurrency matters, skill-based routing matters, and real-time volatility matters.
This guide explains what Erlang C still does well, where it starts to break down, what leading contact centres use instead, and the practical next steps to strengthen your workforce planning capability.
Still useful — not the whole story
Erlang C is fast and reliable for stable, voice-led queues. It was simply never built for multi-channel, concurrent, skill-based work.
The real risk is over-reliance
Problems start when a single queueing formula becomes the only lens for staffing a modern, highly variable operation.
Widen the planning lens
Leading centres pair Erlang with scenario thinking, simulation, better forecasting and real-time optimisation.
The Real Issue Isn't Erlang. It's Over-Reliance.
There's a lot of industry noise suggesting that "nobody uses Erlang anymore" or that it's obsolete. That's too simplistic. Many contact centres still use Erlang C — directly in spreadsheets and calculators, or indirectly through Workforce Management tools that incorporate queueing logic in a broader planning engine.
The problem isn't that Erlang exists. The problem is when it becomes the only lens through which staffing decisions are made in environments that are now multi-channel, skill-based, highly variable and increasingly dynamic.
The point that matters
Erlang C is not wrong. It's simply too narrow to act as a complete workforce planning strategy in a modern contact centre.
What Erlang C Still Does Well
Don't dismiss Erlang C just because the operating environment has changed. At its core, it helps you understand the relationship between call volume, average handle time, service level targets and the number of people required to answer demand — genuinely useful, especially in voice-led environments where assumptions are relatively stable.
Fast baseline modelling
A quick way to estimate staffing needs without the complexity of a full simulation model.
Great for education
It clearly shows why service level, occupancy, delays and resourcing are tightly connected, not independent decisions.
Useful in simpler environments
Where voice dominates, routing is straightforward and variability is moderate, it can still be directionally sound.
A solid reference point
Even advanced WFM environments use Erlang-style logic as a baseline before layering in more complex assumptions.
Where Erlang C Starts to Break Down
Erlang C was built for a much narrower operating context than most contact centres now deal with. Once you move beyond a relatively simple voice queue, the model starts to lose explanatory power.
Multi-channel operations
Chat, messaging, email and social all behave differently — different expectations, response logic and pacing. A voice-queue model cannot represent them all.
Concurrency
Agents now handle more than one contact at a time on chat and messaging. Capacity is no longer simply one person handling one interaction.
Customer patience & abandonment
Erlang assumes structured waiting behaviour. In reality patience varies by channel, intent, time of day and urgency.
Skill-based routing & blended teams
Overlapping skills, specialist queues, priority rules and overflow logic mean staffing is about capability and utilisation across a system, not one queue.
⚡ Volatility and real-world messiness
Erlang relies on averages — average handle time, average arrival rates, average conditions. But contact centres run on spikes, peaks, campaigns, outages, roster gaps and sudden shifts in demand.
Erlang gives you a number. Modern operations need a range, a risk profile and a response plan.
What Leading Contact Centres Use Instead
There's no single "new Erlang" that has replaced the old model. The shift is broader: leading centres are moving from formula-led planning to system-led planning — using methods that reflect how the contact centre actually behaves rather than forcing everything into one queueing formula.
Simulation-based planning
Discrete-event simulation models arrivals, routing, channel mix, skills, patience and variability — giving a range of outcomes, not one fixed answer.
AI-driven forecasting
Beyond static historical averages: seasonality, external events, campaigns, billing cycles, anomaly detection and more frequent reforecasting.
Multi-channel workload modelling
Channels modelled differently by service expectation, pacing, concurrency and work type — vital as chat and messaging grow.
Real-time optimisation
Reforecasting intra-day, monitoring adherence and queue performance, and adjusting staffing or routing as conditions emerge.
Why This Matters Operationally
This isn't just a technical WFM debate. The quality of your planning model has real consequences right across the operation.
Customer experience
Underestimating demand or misreading channel behaviour drives longer waits, missed SLAs, higher abandonment and more customer effort.
Employee Experience
Weak planning creates avoidable pressure, higher occupancy, more volatility and greater burnout risk on the frontline.
Financial performance
Poor planning drives both over- and understaffing at different times — you spend too much while still delivering poor outcomes.
Leadership confidence
When planning outputs repeatedly miss operational reality, trust in the discipline erodes and decisions become reactive.
Digital channel growth
Self-service, chat, bots and asynchronous support need models that reflect how those channels behave — not voice with different labels.
The most capable contact centres don't ask "Should we stop using Erlang?" They ask "What is Erlang still useful for, and what else do we need around it?"
Practical Next Steps for Contact Centres
For most organisations the answer isn't to rip out existing planning logic overnight. The practical path is to raise planning maturity step by step.
Stress-test your current assumptions
Where do planning outputs regularly diverge from reality? Are volumes more volatile than you assume? Are digital channels distorting workload? Is customer patience changing?
Think in scenarios, not just targets
Instead of a single staffing number, ask what happens if conditions shift — volume spikes, AHT climbs, or a digital campaign lands harder than expected.
Model channels differently
Voice, chat, messaging and asynchronous work should not be treated as if they behave the same way. The more you blend channels, the more this matters.
Match your tools to your complexity
If your WFM environment cannot model concurrency, scenario risk or real-time optimisation, it may be time to review alternative approaches and technology.
Erlang Tools & WFM Resources
If this article highlighted gaps in your current approach, ACXPA has resources to help you take the next step.
Reviewing your WFM tooling?
Explore workforce management technology and consultants in the ACXPA Supplier Directory.
Frequently Asked Questions
Is Erlang C still relevant in modern contact centres?
Yes — for stable, voice-led queues it's still a fast, useful way to estimate staffing. The issue isn't Erlang itself; it's relying on it alone in multi-channel, concurrent, skill-based operations it was never designed for.
What is Erlang C used for?
Estimating how many agents you need to meet a service level target given call volume and average handle time in a voice queue. It links volume, handle time, service level, occupancy and staffing into one model.
Where does Erlang C break down?
Multi-channel work, concurrency (chat and messaging), highly variable customer patience, skill-based routing, and the real-world volatility that simple averages cannot capture.
What are the alternatives to Erlang C?
There's no single replacement formula. Leading centres use simulation, AI-driven forecasting, multi-channel workload modelling and real-time optimisation — often keeping Erlang-style logic as a baseline.
Does Erlang C work for chat and Omnichannel?
Not well on its own. Chat and messaging involve concurrency and different pacing that a single-queue voice model doesn't reflect, so they need to be modelled differently.
Is there a free Erlang C calculator?
Yes — ACXPA provides a free online Erlang C Calculator for quick staffing estimates in voice environments, with advanced interval and simulation tools available to members.
Where to Next
Summary
Erlang C still has a place in contact centre workforce planning. It's useful, practical and relevant in the right context. But treating it as a complete answer in a modern contact centre is where the real problem begins.
The most capable operations aren't simply replacing Erlang with another formula. They're widening the planning lens — using scenario thinking, simulation, improved forecasting, multi-channel modelling and real-time optimisation to better reflect operational reality.
So the question isn't whether Erlang is dead. It's whether it's telling you enough.
Related Articles
- About the Author
- Latest Articles
After more than 30 years in contact centres and CX, I’ve built, run, outsourced and transformed just about every kind of customer operation there is — and I’m still as fascinated by getting it right as I was on day one. That’s what I love about this profession: there’s always more to learn.
As founder and CEO of ACXPA, I’m on a mission to raise the standard of customer experience as a profession — Australian-founded, and increasingly on the global stage. Through ACXPA and our specialist training arm, CX Skills, we help businesses and their frontline teams deliver customer experiences that actually work — on the phone, digital and in person — with the skills, standards and insight to back it up.
