Agent Utilization vs Occupancy Definitions and Tradeoffs

From WFM Labs
Utilization vs occupancy: the shrinkage gap between them.

Agent Utilization and Agent Occupancy are two distinct but frequently conflated metrics used in contact center workforce management to describe how productively agents are deployed relative to available time. Despite their surface similarity — both express a ratio of productive time to total time — they differ in what is included in the denominator and in what managerial levers affect each.[1] Misidentifying one for the other leads to flawed staffing targets, occupancy goals that inadvertently burn out agents, or utilization benchmarks that misrepresent the operational state of a contact center. Understanding the precise definitions, the relationship between them, and the appropriate target ranges for each is foundational to capacity planning and WFM operations at all maturity levels.[2]

Definitions

Occupancy

Occupancy is the proportion of time an agent spends handling contacts (talk time plus after-call work) relative to the time the agent is logged in and available to receive contacts. Formally:

Occupancy = (Handle Time) / (Handle Time + Idle Time)

or equivalently:

Occupancy = (Handle Time) / (Available Time)

where "Available Time" includes both handle time and idle time waiting for the next contact, but excludes time the agent has placed themselves in an auxiliary state (e.g., break, training, meeting). Occupancy is determined primarily by the ratio of offered load to staffing — it is a function of the queue's traffic intensity, not of individual agent behavior. In a well-managed queue, all agents experience approximately the same occupancy because work is distributed uniformly by the automatic call distributor (ACD).

The Erlang-C formula provides the theoretical occupancy for a given offered load and staffing level:

Occupancy = Offered Load (Erlangs) / Number of Agents

This relationship holds in the M/M/c queue and is approximately valid in most real-world inbound contact center settings.

Utilization

Utilization measures the proportion of total paid or scheduled time that an agent spends in productive activity — including not just contact-handling time but all work-related activities: handle time, idle time while available, after-call work, training, meetings, team huddles, and other auxiliary activities that represent legitimate productive use of the agent's time.

Utilization = (All Productive Time) / (Total Paid Time)

The complement of utilization is Shrinkage — the proportion of paid time during which agents are not available to handle contacts and are not engaged in scheduled productive activities. Shrinkage includes unplanned absence, unscheduled breaks, and other unproductive time.

A simplified relationship:

Utilization = 1 − Shrinkage

In this formulation, an agent who is scheduled and at their workstation handling contacts or in auxiliary activities has 100% utilization of their on-duty time; an agent who is absent has 0% utilization of their scheduled time.

The Key Difference

The critical distinction is the denominator:

  • Occupancy denominator = logged-in available time (excludes all auxiliary states and shrinkage)
  • Utilization denominator = total paid or scheduled time (includes all time, including shrinkage)

An agent can have high occupancy (85%) and moderate utilization (70%) simultaneously: they spend 85% of their logged-in available time handling contacts, but they are only logged in and available for 82% of their scheduled hours, with the remainder absorbed by breaks, training, and unplanned absence.[3]

Numerical Illustration

Consider an agent scheduled for 8 hours (480 minutes):

Activity Minutes
Handle time (talk + ACW) 290
Idle time (available, waiting for call) 70
Scheduled breaks and lunch 60
Team meeting 30
Unplanned absence 30
Occupancy = 290 / (290 + 70) = 290 / 360 = 80.6%
Utilization = (290 + 70 + 60 + 30) / 480 = 450 / 480 = 93.8%
Shrinkage = (60 + 30 + 30) / 480 = 120 / 480 = 25%

Note that utilization here includes the scheduled breaks and meetings as productive use of scheduled time; only the unplanned absence contributes to shrinkage.

Occupancy Targets and Agent Wellbeing

Sustained high occupancy is associated with agent stress, quality degradation, and increased attrition. When occupancy consistently exceeds 85–90%, agents have insufficient idle time between contacts to recover, complete after-call work accurately, or attend to administrative tasks. The result is a reinforcing deterioration: quality scores decline, error rates increase, attrition accelerates, and the resulting understaffing drives occupancy even higher.

Industry guidance from ICMI and Cleveland suggests occupancy targets in the range of 80–85% for most inbound contact center environments, with lower targets (75–80%) appropriate for technically complex or emotionally demanding contact types.[4] Higher targets (85–90%) may be sustainable in shorter interaction types (e.g., chat, simple IVR-assisted calls) where recovery time between contacts is less critical.

Critically, occupancy is not a target that can be directly managed by individual agents or supervisors in real time. It is the emergent result of staffing decisions. The only way to reduce occupancy sustainably is to add agents (reducing traffic intensity) or to reduce offered load (e.g., through self-service deflection or Average Handle Time reduction). Supervisors who pressure agents to "stay available" in an understaffed environment are not reducing occupancy — agents are already maximally available; they have no idle time to give.

The Power of One Connection

The Power of One principle describes how marginal staffing changes produce disproportionate service-level impacts in contact center queueing systems. The same principle applies to occupancy: adding one agent to an understaffed queue reduces occupancy for all agents in that queue, not just the marginal agent. This is because arrival rates are distributed across a larger server pool, reducing each agent's share of offered load. The occupancy reduction per added agent is nonlinear — larger at high occupancy levels (near the knee of the Erlang-C curve) and smaller at lower occupancy levels. This nonlinearity reinforces the importance of avoiding sustained high-occupancy operation.

Utilization Targets and Shrinkage Management

While occupancy is managed through staffing levels, utilization is managed through Shrinkage control. An organization that allows excessive unplanned absence, overruns breaks, or accumulates unauthorized auxiliary time will see low utilization even at appropriate staffing levels — resulting in waste without service-level benefit.

Target utilization levels depend on the shrinkage budget built into the staffing model. A contact center that plans for 30% shrinkage will target approximately 70% utilization of paid time as the baseline expectation. The question of whether actual utilization meets this target is a WFM operations question, addressed through real-time adherence monitoring and schedule compliance reporting.

Implications for Multi-Channel and Blended Operations

In blended environments where agents handle multiple contact types — inbound calls, outbound calls, email, chat — occupancy and utilization require more careful definition. An agent who is simultaneously handling a chat session and an email task is not idle, but may appear to have low occupancy on either channel individually. Multi-channel occupancy requires aggregation across channels, weighted by the proportion of attention each channel commands. See Multi-Channel and Blended Operations.

Common Misconceptions

  • Misconception: High utilization means agents are too busy and should not be given more work. Correction: Utilization includes all productive activities including idle time. An agent at 70% utilization may have substantial idle time while available, which is a function of staffing level, not agent behavior.
  • Misconception: Low occupancy means agents are shirking. Correction: Low occupancy is a consequence of appropriate staffing relative to offered load. Idle time between contacts is both expected and desirable in a properly staffed queue.
  • Misconception: Occupancy and utilization can be set as targets for individual agents. Correction: Occupancy is a queue-level outcome; individual agents experience approximately the same occupancy as their peers when work is equitably distributed. Utilization is an organizational average, not an individual target that can be meaningfully set at the agent level.

Maturity Model Considerations

At L1–L2 maturity, occupancy and utilization are often used interchangeably or not tracked at all. Staffing decisions are made without reference to occupancy targets, and burnout from sustained high occupancy is attributed to individual agent characteristics rather than systemic understaffing.

At L3, organizations compute both metrics separately, maintain documented target ranges, and incorporate shrinkage budgets into headcount planning. High occupancy is recognized as a systemic indicator requiring staffing action, not a supervisory coaching issue.

At L4–L5, occupancy and utilization are tracked at interval granularity, integrated with real-time workforce management systems, and used to trigger intraday staffing adjustments. See WFM Labs Maturity Model.

Related Concepts

References

  1. ↑ Cleveland, B. (2012). Call Center Management on Fast Forward, 4th ed. ICMI Press.
  2. ↑ Koole, G., & Mandelbaum, A. (2002). Queueing models of call centers: An introduction. Annals of Operations Research, 113(1–4), 41–59.
  3. ↑ ICMI. (2020). The Right Occupancy Rate for Your Call Center. International Customer Management Institute.
  4. ↑ ICMI. (2020). The Right Occupancy Rate for Your Call Center. International Customer Management Institute.
Retrieved from "https://wiki.wfmlabs.org/w/index.php?title=Agent_Utilization_vs_Occupancy_Definitions_and_Tradeoffs&oldid=1503"