Call Center Workforce Management: How to Schedule by Call

It’s 2:15 on a Tuesday and your queue just exploded. Three agents are on lunch, two called out this morning, and the wait time on your dashboard is climbing past four minutes. You scheduled the same number of people you always do, because that’s what felt right. Now customers are abandoning calls and the ones who do get through are already annoyed.

Then Thursday afternoon arrives and the opposite happens. You’ve got eight agents staring at an empty queue, refreshing screens, while you quietly burn payroll on a stretch where the phones barely ring. You feel both problems in the same week, and neither one is your fault exactly — you just don’t have a reliable way to match the number of people on the floor to the number of calls coming in.

Call center workforce management is the discipline that fixes this. It replaces “schedule what feels right” with a repeatable process: predict how many contacts you’ll get, calculate how many agents that requires, and build a schedule that lines up with the curve. Done well, it’s the difference between a calm floor and a daily fire drill.

Call center workforce management (WFM) is the practice of forecasting contact volume, calculating the staff needed to hit a service-level target, and scheduling agents to match demand interval by interval. It combines call volume forecasting, Erlang-based staffing math, and shift scheduling so you have enough agents on at busy times and don’t overpay during quiet ones.

WFM Basics for Call Centers: The Four Moving Parts

Before you touch a schedule, it helps to see the whole loop. Workforce management isn’t one task — it’s four that feed each other.

Forecast, staff, schedule, manage the day

Forecast answers “how many contacts will arrive, and when?” Staffing answers “how many agents does that volume require to hit our target?” Scheduling turns that requirement into actual shifts, breaks, and lunches for named people. And intraday management is what you do when reality drifts from the plan — someone calls out, volume spikes, a system goes down.

Skip any one of these and the others wobble. A perfect forecast with a sloppy schedule still leaves you short at 2 p.m. A tight schedule built on a bad forecast just means you’re precisely wrong. The goal of wfm basics in a call center is to get all four working as a cycle you repeat every week.

Service level is the number everything serves

Most centers anchor on a service-level target — a common one is answering 80% of calls within 20 seconds, often written “80/20.” Everything in WFM exists to protect that number at a cost you can live with. When you decide how many agents to schedule, you’re really deciding how often you’ll hit your target. Write yours down before you start; it’s the yardstick for every staffing decision that follows.

Call Volume Forecasting: Predicting the Curve

You can’t staff what you can’t predict. Call volume forecasting is the foundation, and the good news is that contact patterns are far more regular than they feel in the moment.

Use your own history first

Your phone system or contact platform already logs how many calls arrived in each interval — usually 15- or 30-minute blocks. Pull at least several weeks of that data and you’ll see the shape: a morning ramp, a midday peak, an afternoon lull, maybe an early-evening bump. Mondays run heavier than Fridays in most centers. The first and last days around a weekend or holiday tend to spike.

Forecasting is mostly about respecting those patterns instead of averaging them away. A daily average tells you nothing useful, because nobody calls at the average rate — they call in waves.

Adjust for the things history can’t see

History gives you the baseline. Then you layer on what you know is coming: a billing cycle that drives questions on the 1st, a marketing email going out Wednesday, a product launch, a seasonal rush. Build the forecast at the interval level, because two days with identical call totals can need completely different staffing if one is flat and the other has a brutal noon spike.

Here’s an illustrative interval forecast for a single weekday (your real numbers will differ):

Interval	Forecast calls	Avg handle time	Pattern
8:00–9:00	40	5 min	Morning ramp
9:00–10:00	95	5 min	Climbing
11:00–12:00	140	5 min	Daily peak
1:00–2:00	75	5 min	Post-lunch dip
4:00–5:00	110	5 min	Late surge

The totals matter less than the shape. Staffing to that 11 a.m. peak is a different problem than staffing the 1 p.m. dip, and a single headcount can’t serve both.

Erlang Staffing for a Call Center: Turning Volume Into Headcount

Once you have a forecast, you need to convert calls into agents. This is where new managers usually guess — and guess wrong, because the math isn’t intuitive.

Why you can’t just divide

Say you expect 140 calls in an hour and each takes 5 minutes to handle. That’s 700 minutes of work, or about 11.7 agent-hours. So 12 agents, right? No. If you staff exactly to the workload, queues explode the moment calls cluster — and calls always cluster. You need a cushion for the randomness of arrivals.

The standard tool for sizing that cushion is the Erlang C formula, which is the heart of Erlang staffing in a call center. You feed it three things: forecast volume for the interval, average handle time, and your service-level target. It returns the number of agents needed to hit that target given how unpredictably calls actually arrive. You don’t have to compute it by hand — free Erlang calculators and most WFM platforms do it — but you should understand what it’s telling you.

Two ideas Erlang makes obvious

First, the curve isn’t linear. Going from 80/20 to 90/10 costs more agents than you’d expect, and chasing a near-zero wait time gets wildly expensive. Pick a target that balances customer experience against payroll.

Second, bigger pools are more efficient. A center handling 1,000 calls an hour needs proportionally fewer “extra” agents than one handling 50, because randomness smooths out at scale. That’s why splitting one team into tiny isolated queues quietly raises your staffing cost.

Erlang C assumes voice calls where people wait in a queue. For chat or email, where agents handle several contacts at once or replies aren’t immediate, you’ll need different math — but for phone-first centers, it’s the workhorse.

Call Center Scheduling: Building Shifts That Match the Curve

Now you have an agent requirement for every interval. Call center scheduling is the act of covering that requirement with real shifts — and it’s harder than it looks, because people work in continuous blocks while demand moves up and down all day.

Match shape, not just totals

The classic mistake is scheduling everyone 9-to-5 and then drowning at the 11 a.m. peak and the 4 p.m. surge while overstaffing the lull in between. Instead, stagger start times, vary shift lengths, and place lunches and breaks deliberately in the slower intervals. Move breaks out of your peak and into your dip and you’ve effectively added coverage where you needed it without adding a single labor hour.

Plan for shrinkage

“Shrinkage” is every paid hour an agent isn’t available to take contacts: breaks, lunches, training, meetings, coaching, time off, and the small gaps between calls. It’s commonly a meaningful chunk of paid time, and if you ignore it you’ll schedule the Erlang number and still come up short, because a third of your “staff” is in a training room or on break.

The fix is simple to state: if the math says you need a certain number of agents on the phones, you have to schedule more people than that to absorb shrinkage. Track your own shrinkage rate from real data rather than guessing.

A simple weekly rhythm

Day	Task
Monday	Pull last week’s actuals; compare forecast vs. real volume
Tuesday	Update next week’s interval forecast
Wednesday	Run staffing math; build the draft schedule
Thursday	Confirm availability and time-off; publish
Daily	Watch service level; flex breaks and adjust on the fly

Make it a routine, not a scramble, and each week’s forecast gets a little sharper.

Managing the Day: When Reality Drifts

No forecast survives contact with a Tuesday. Intraday management is watching your live service level and reacting fast: shifting breaks earlier, asking for voluntary overtime when a surge hits, or releasing people early when it’s dead. The teams that hold their targets aren’t the ones with perfect forecasts — they’re the ones who notice drift within minutes and adjust. Keep your published schedule current so everyone’s looking at the same plan when you make a change.

Call center work overlaps with other shift-scheduling challenges, too. If you run extended or overnight hours, the patterns in our clopening shifts guide apply, and managing sudden absences is its own discipline — see our take on a last-minute call-outs policy. More scheduling playbooks live in the /category/hospitality hub.

How ShiftSynch helps

ShiftSynch is built for busy service teams: organize staff into teams, build shifts around your peaks with rotation patterns, manage time-off and availability, and track labor in clear reports, on web and mobile.

Start free — no credit card required (1 team, up to 10 staff); paid plans start at $19/month with a 14-day trial.

Start free on ShiftSynch

Workforce management sounds like a big system, but it’s really one honest loop: predict the calls, do the staffing math, schedule to the shape of the day, and adjust when the day surprises you. Start with your own interval history this week and build from there. Each cycle you run makes the next forecast more accurate and your floor a little calmer.

Frequently Asked Questions

Q: What are the wfm basics every call center manager should know? The basics are four linked steps: call volume forecasting, staffing math to convert volume into agents, scheduling those agents into shifts and breaks, and intraday management to handle the unexpected. They run as a weekly cycle anchored to a service-level target, like answering 80% of calls within 20 seconds, so every decision protects that number.

Q: How does call volume forecasting actually work? Start with your phone system’s interval-level history — calls per 15 or 30 minutes — across several weeks. Patterns repeat: morning ramps, midday peaks, afternoon dips, heavier Mondays. Use those patterns as your baseline, then adjust for known events like billing cycles, campaigns, or seasonal rushes. Forecast at the interval level, never as a daily average.

Q: Why do I need Erlang staffing for a call center instead of simple division? Dividing total call minutes by shift length ignores that calls arrive in random clusters, so a workload-only count leaves you short the moment volume bunches up. The Erlang C formula sizes the extra cushion needed to hit your service-level target. Free calculators and WFM tools run it for you from volume, handle time, and target.

Q: How is call center scheduling different from normal shift scheduling? Demand changes every interval, so a flat 9-to-5 schedule overstaffs lulls and crumbles at peaks. Effective call center scheduling staggers start times, varies shift lengths, and places breaks in slower intervals. You also schedule extra people beyond the raw agent requirement to cover shrinkage — breaks, training, and time off that pull agents off the phones.