Staffing Model
Definition
Staffing Model
A staffing model is the plan a call center uses to work out how many agents to hire, schedule, and deploy against expected call volume. It maps demand to headcount, shifts, and skill mix. Right-sized staffing is the top lever on cost and service.
Get it wrong on the low side and queues balloon, abandonment spikes, and service levels miss. Get it wrong on the high side and payroll swallows the margin. Most contact centers spend 60–70% of their budget on people.
A good staffing model links three inputs — forecast, shrinkage, and occupancy target — to one output: agents needed per interval. Teams rebuild it monthly for the forecast and revalidate weekly against actuals.
The 2024 Deloitte Global Contact Center Survey found that 76% of operations leaders now list workforce planning as a top-three priority, up from 61% pre-pandemic.
That shift lands hardest on accuracy. A model built on stale call center forecasting inputs will miss by double digits in peak weeks, and no amount of clever scheduling recovers the gap.
Key takeaways
- A staffing model converts forecast call volume into agents required per 15- or 30-minute interval.
- The core formula uses Erlang C: agents are a function of volume, handle time, and target service level.
- Shrinkage of 25–35% is normal, so the plan has to gross up for it rather than ignore it.
- Fixed, variable, blended, and outsourced are the four common staffing structures.
- BPO staffing lets buyers flex headcount 20–40% up or down against seasonal peaks.
How it works
A staffing model runs in four steps: forecast interval-level demand, convert that demand into required agents with Erlang C, gross the result up for shrinkage, then schedule that headcount across shifts and check occupancy holds.
- Forecast interval-level call volume from 12–18 months of history plus known drivers such as marketing, product launches, and seasonality.
- Calculate required agents with Erlang C or a Monte Carlo simulation against the target in the service level agreement, typically 80/20.
- Gross up for shrinkage, the hours agents are paid for but are not on the phone: breaks, meetings, training, and sick leave.
- Schedule the grossed-up headcount across shifts, checking that adherence and occupancy rate both stay inside target.
Here is the simplified formula most workforce managers memorise: agents required = (forecast contacts × average handle time in seconds) ÷ (interval length in seconds × target occupancy).
Say a 30-minute interval expects 200 calls at a handle time of 300 seconds, with 85% target occupancy. Agents = (200 × 300) ÷ (1,800 × 0.85) = 39 on the floor. Gross that up 30% for shrinkage and you schedule 51.
| Model | How it works | Best for | Typical cost premium |
|---|---|---|---|
| Fixed | Full-time headcount on flat 8-hour shifts | Steady inbound, low variance | Baseline |
| Variable | Part-time staff, split shifts, on-call cover | Peaked retail and seasonal traffic | 5–10% lower |
| Blended | Full-time core plus a part-time surge tier | Most mid-size operations | 3–7% lower |
| Outsourced | A provider handles hiring, scheduling, flex | Rapid scale, cost arbitrage | 40–70% lower |
Erlang C — the queuing math named after Danish engineer Agner Krarup Erlang — assumes calls arrive at random and that callers wait rather than hang up. Planning suites bake it in alongside skill-based routing and multi-channel adjustments.
Erlang C also overstates the agents you need when abandonment is high. Operations above a 10% abandon rate blend in a simulation model to avoid overstaffing the shoulder intervals.
Examples
Large providers run staffing models at a scale in-house teams rarely see: multi-country coverage, seat pools that flex by the thousand, and blended shift structures rebuilt every quarter against retail and banking demand curves.
Concentrix, a listed customer experience provider, runs staffing models across 70+ countries. It blends Manila and Bogotá capacity for follow-the-sun coverage of US banking and technology accounts.
Its 2024 annual report cited a 22% year-on-year gain in flex-headcount utilisation, which buyers should ask about before signing any seat commitment.
Teleperformance, the France-based outsourcing group, uses a hub-and-spoke model. A fixed core team is topped up by a variable work-from-home tier, branded Cloud Campus, that scales 15,000 seats against retailer holiday peaks.
Alorica rebuilt its US retail staffing model in Q3 2024 around a 60/40 blend: 60% full-time and 40% part-time on four-hour minimum shifts. Black Friday coverage rose 31% with no extra fixed cost.
Foundever, formed by the 2023 merger of Sitel and Sykes, publishes an annual CX Index. Its 2025 edition reported that clients on its variable-staffing model cut cost-per-contact by 18% against the fixed-headcount baseline.
Filipino provider SPI CRM and Indian giant Infosys BPM both quote 20–40% headcount elasticity inside 30 days as standard. That flex band is why buyers route peak volume through business process outsourcing contracts.
Seasonality makes the choice obvious in retail. A mid-size retailer typically sees November volume run close to double its July baseline, so a fixed model pays a third of the floor to sit idle for months.
Related terms
A staffing model sits inside a wider planning stack. These terms cover the inputs it consumes, the targets it plans against, and the artefacts it produces once headcount is agreed for every interval of the week.
- Workforce Management: the parent discipline covering forecasting, scheduling, and real-time adherence.
- Forecasting: the demand input that feeds every staffing calculation.
- Shrinkage: the non-productive paid hours the plan has to gross up for.
- Occupancy Rate: the productivity target the plan aims at, usually 80–85%.
- Service Level Agreement: the contractual service target the plan has to hit.
- Business Process Outsourcing: the delivery route for outsourced staffing arrangements.
- Schedule: the operational artefact the model produces once headcount is set.
FAQ
Six questions cover what planners ask most: how the model differs from a roster, how often to rebuild it, which shrinkage figure to trust, what outsourcing changes, which service level to aim at, and which tools do the math.
What is the difference between a staffing model and a schedule?
The staffing model tells you how many agents you need per interval; the schedule tells you which named agents work which shift. The model is the math, and the schedule is the roster built from it.
How often should a staffing model be rebuilt?
Rebuild the full model, including forecast, headcount, and shrinkage assumptions, every month. Weekly variance reviews catch drift, and interval-level tuning happens daily inside the planning tool.
What shrinkage number should I use?
Industry benchmarks put shrinkage at 25–35% of paid hours, per the 2024 ContactBabel US Contact Center Decision-Makers Guide. Retail and healthcare usually sit higher, while outbound sales sit lower. Use your own trailing 90-day actuals whenever you have them.
Does outsourcing change the staffing model?
Yes. A provider absorbs the hiring, training, and scheduling risk, so the buyer’s model becomes a capacity purchase, locked seats plus a flex band, rather than a full headcount plan.
What service level does the staffing model aim for?
The most common target is 80/20, meaning 80% of calls answered in 20 seconds, followed by 90/30. Financial services and premium tech support often push to 90/15, and that target is set in the contract, not by the planning team.
Which software calculates it?
Erlang-based planning runs inside NICE CXone, Genesys Cloud, Verint, Calabrio, and Playvox, all of which layer shrinkage, multi-skill overlays, and AI-driven arrival forecasting on top of the core math.
Compare staffing quotes from 3,000+ vetted providers on the Outsource Accelerator hubs directory.
Independent
