Unit Economics of Workforce Operations
Unit Economics of Workforce Operations provides the cost denominators that every WFM leader needs when talking to finance. Where Workforce Cost Modeling builds the per-FTE cost stack, unit economics inverts the lens: what does each unit of output actually cost to produce? This is the language CFOs speak — not headcount, not utilization, but cost per unit of value delivered.
Three unit metrics anchor the framework: cost-per-contact (CPC), cost-per-resolution (CPR), and cost-per-productive-hour (CPPH). Each tells a different story. CPC measures efficiency. CPR measures effectiveness. CPPH measures labor utilization. Together they give a complete economic picture of workforce operations.
Overview
Unit economics answers the question executives actually ask: how much does it cost us to do one thing? Headcount is an input. FTE cost is a supply metric. Unit economics connects supply cost to demand output — and that connection is where WFM earns its seat at the strategy table.
The shift matters because:
- Headcount-based budgeting hides inefficiency. Adding 50 agents to handle a volume spike looks like 50 × salary. Unit economics reveals whether each contact costs $7.20 or $11.40.
- Channel migration decisions require comparable denominators. Moving 30% of volume from voice to chat only saves money if the per-contact cost actually drops — and it might not, if chat concurrency assumptions are wrong.
- Automation business cases live or die on unit economics. An AI agent handling contacts at $0.25 each is compelling against $8.50 voice contacts but irrelevant if the AI can only handle 15% of contact types.
- Vendor benchmarking becomes meaningful. Comparing BPO pricing across vendors requires a common unit — and that unit is cost-per-contact or cost-per-resolution, not FTE rate.
Core Metrics
Cost-Per-Contact (CPC)
The foundational unit metric. CPC captures total labor cost divided by total contacts handled.
CPC = Total Fully Loaded Labor Cost ÷ Total Contacts Handled
"Fully loaded" means: base salary + benefits + employer taxes + facilities allocation + technology cost + supervision overhead + QA cost + WFM overhead. See Workforce Cost Modeling for the complete cost stack.
"Contacts handled" means: all contacts resolved, transferred, or dispositioned by the agent pool in the measurement period — not contacts offered, not contacts answered. The denominator must match what the workforce actually processed.
Benchmark ranges by industry (2025):
| Industry | CPC Range (USD) | Median |
|---|---|---|
| Financial services | $7.50–$14.00 | $10.20 |
| Healthcare | $8.00–$16.00 | $11.50 |
| Telecommunications | $5.50–$10.00 | $7.80 |
| Retail/e-commerce | $4.50–$9.00 | $6.40 |
| Technology/SaaS | $6.00–$12.00 | $8.70 |
Cost-Per-Resolution (CPR)
CPC adjusted for first-contact resolution. CPR penalizes operations that handle contacts cheaply but generate repeat calls.
CPR = CPC ÷ FCR
Where FCR = First Contact Resolution rate expressed as a decimal.
An operation with $8.00 CPC and 72% FCR has a CPR of $11.11. An operation with $9.50 CPC and 88% FCR has a CPR of $10.80. The "more expensive" operation is actually cheaper per resolved issue. This is why CPR is the metric that matters for total cost optimization — it captures the rework loop.
The relationship is non-linear. Improving FCR from 70% to 80% reduces CPR by 12.5%. Improving from 80% to 90% reduces it by 11.1%. The marginal return diminishes, but the absolute dollar savings remain significant at scale.
Cost-Per-Productive-Hour (CPPH)
The labor utilization metric. CPPH captures what each hour of actual productive work costs after accounting for shrinkage.
CPPH = Fully Loaded Annual Cost per FTE ÷ Annual Productive Hours
Annual productive hours = (Available hours per year) × (1 − Shrinkage Rate)
For a standard 2,080-hour work year at 35% shrinkage:
Productive hours = 2,080 × 0.65 = 1,352 hours
At $55,000 fully loaded annual cost:
CPPH = $55,000 ÷ 1,352 = $40.68
This is the number that makes managers uncomfortable: the shrinkage multiplier. An agent paid $26.44/hour ($55,000 ÷ 2,080) actually costs $40.68 per productive hour — a 54% markup driven entirely by shrinkage. At 38% shrinkage (common in operations with generous PTO and heavy ongoing training), the markup reaches 61%.
Channel Economics
Unit economics vary dramatically by channel. The variation is not just about wage rates — it is driven by concurrency, handle time, automation potential, and resolution complexity.
| Channel | CPC Range | Key Driver | Concurrency | Typical AHT |
|---|---|---|---|---|
| Voice | $6.00–$12.00 | 1:1 interaction, longest handle time | 1.0 | 6–10 min |
| Chat | $3.00–$6.00 | 2–3 concurrent sessions | 2.0–3.0 | 8–14 min per session |
| $4.00–$8.00 | Asynchronous, batch processing | N/A | 5–12 min handling | |
| SMS/Messaging | $2.50–$5.00 | Asynchronous, short interactions | 3.0–5.0 | 4–8 min total |
| AI Agent (automated) | $0.10–$0.50 | Compute cost, no labor | Unlimited | 1–3 min |
| AI-Assisted Agent | $4.00–$8.00 | Agent + AI copilot | 1.0–2.0 | 4–7 min |
Critical nuance: Chat CPC appears 50% cheaper than voice, but that depends on achieving target concurrency. An operation planning for 3.0 chat concurrency that actually runs at 1.8 (due to complex issues or agent skill gaps) sees chat CPC rise to $5.00–$10.00 — barely cheaper than voice. Always validate concurrency assumptions against actual data before building channel migration business cases.
AI agent economics: The $0.10–$0.50 range for fully automated AI interactions reflects compute and API costs only. It excludes: AI development and maintenance cost (amortized), escalation handling by human agents, quality monitoring, and error remediation. The net AI CPC — including the human safety net — typically runs $0.80–$2.00 for mature deployments. Still dramatically cheaper than human-only, but not the 95% savings that raw numbers suggest.
Operation Type Economics
Unit economics differ by operation type because the definition of "one unit of output" changes.
| Operation Type | Unit of Output | Typical Unit Cost | Key Variables |
|---|---|---|---|
| Contact center | Contact or resolution | $5–$14 | Channel mix, FCR, AHT, concurrency |
| Back office | Transaction or case | $3–$25 | Complexity, automation rate, error/rework rate |
| Field service | Work order or visit | $150–$500 | Travel time, first-visit resolution, parts cost |
| Knowledge work | Deliverable or hour | $50–$200 | Expertise level, utilization rate, tool leverage |
For back-office operations, the unit cost range is wide because "transaction" spans from a simple data entry ($3) to a complex insurance claim adjudication ($25). Operations must define homogeneous work units before benchmarking makes sense.
Marginal Cost vs. Average Cost
This is where unit economics gets strategically useful — and where most WFM teams stop too soon.
Average cost divides total cost by total output. It is what it is.
Marginal cost asks: what does the next contact cost? In a workforce operation, marginal cost is almost always lower than average cost — until it suddenly is not.
The economics work like this:
Below the occupancy wall (occupancy < 85%):
- Fixed costs (facilities, technology, supervision, WFM) are already sunk
- The incremental contact is handled by existing staff absorbing one more interaction
- Marginal CPC ≈ variable labor cost only ≈ 40–60% of average CPC
At the occupancy wall (occupancy 85–92%):
- Existing staff are at capacity
- Service levels start degrading (longer wait times, higher abandons)
- Marginal cost includes service-level penalty and customer satisfaction damage
- Marginal CPC ≈ average CPC
Above the wall (occupancy > 92%):
- New capacity must be added: overtime, temp staff, or new hires
- Overtime runs 1.5× base cost; temp/contract labor runs 1.2–1.8× fully loaded FTE cost
- Marginal CPC = 1.5–2.5× average CPC
Strategic implication: WFM teams that report only average CPC miss the capacity story. A 500-agent center running at 78% occupancy has room to absorb 15–20% more volume at marginal cost well below average. That is a different conversation with finance than "we need 75 more agents."
Revenue per Contact
For sales and retention operations, the economic lens flips. Each contact is not just a cost — it generates revenue or saves revenue.
Net Unit Value = Revenue per Contact − CPC
A retention operation saving $45 per retained customer at $8.50 CPC generates $36.50 net value per contact. This reframes the WFM conversation entirely: understaffing a retention queue does not save money — it destroys value at $36.50 per abandoned call.
Revenue-per-contact metrics belong in every WFM business case for sales, retention, collections, and upsell operations. Without them, WFM optimization looks like pure cost reduction when it should look like revenue optimization.
Worked Example: Building a CPC Model for a 500-Agent Center
Scenario: A mid-size financial services contact center — 500 agents, blended voice and chat, goal of presenting unit economics to the CFO.
Step 1: Establish the fully loaded cost base
| Cost Component | Per Agent | Total (500 agents) |
|---|---|---|
| Base salary | $42,000 | $21,000,000 |
| Benefits (32% load) | $13,440 | $6,720,000 |
| Employer taxes (7.65%) | $3,213 | $1,606,500 |
| Facilities allocation | $4,800 | $2,400,000 |
| Technology (desktop, telephony, tools) | $6,000 | $3,000,000 |
| Supervision (1:15 ratio, loaded) | $5,333 | $2,666,500 |
| QA (1:25 ratio, loaded) | $3,200 | $1,600,000 |
| WFM overhead (1:75 ratio, loaded) | $1,067 | $533,500 |
| Total fully loaded cost | $79,053 | $39,526,500 |
Step 2: Calculate productive hours
Annual paid hours per agent: 2,080 Shrinkage rate: 38% (PTO 8%, training 4%, meetings 3%, coaching 2%, breaks 7%, absenteeism 4%, off-phone work 6%, other 4%) Productive hours per agent: 2,080 × 0.62 = 1,290 hours Total productive hours: 500 × 1,290 = 644,800 hours
The 38% shrinkage multiplier: Each agent's effective hourly cost is not $38.01 ($79,053 ÷ 2,080) — it is $61.28 ($79,053 ÷ 1,290). That 61% markup is invisible in headcount-based budgeting and is the single most common source of workforce budget variance.
Step 3: Calculate contacts handled
Blended AHT: 7.2 minutes (voice 8.5 min at 60% mix, chat 5.5 min at 40% mix with 2.2 concurrency) Contacts per productive hour: 60 ÷ 7.2 = 8.33 Annual contacts: 644,800 × 8.33 = 5,371,184
Step 4: Calculate unit metrics
CPC = $39,526,500 ÷ 5,371,184 = $7.36
At 76% FCR: CPR = $7.36 ÷ 0.76 = $9.68
CPPH = $79,053 ÷ 1,290 = $61.28
Step 5: Sensitivity analysis
| Variable Changed | New CPC | Delta | Insight |
|---|---|---|---|
| Shrinkage 38% → 33% | $6.69 | −$0.67 (−9.1%) | 5-point shrinkage reduction = strongest single lever |
| AHT 7.2 → 6.5 min | $6.65 | −$0.71 (−9.6%) | 42-second AHT reduction matches shrinkage impact |
| FCR 76% → 82% | CPR $8.98 | −$0.70 (−7.2%) | FCR gain reduces resolution cost, not contact cost |
| Attrition 45% → 35% | $7.02 | −$0.34 (−4.6%) | Attrition impact flows through training/ramp cost |
| Chat mix 40% → 55% | $6.91 | −$0.45 (−6.1%) | Channel shift at verified concurrency |
Benchmarking Unit Economics
Comparing unit economics across operations requires normalization. Raw CPC comparisons across companies are meaningless without controlling for:
- Contact complexity: A Tier 1 password-reset operation and a Tier 3 technical support center cannot share a CPC benchmark. Segment by contact type before comparing.
- Channel mix: An all-voice operation will always have higher CPC than a 60% chat operation. Compare within channel or use channel-weighted benchmarks.
- Geography: A US-based center and an offshore center have fundamentally different cost structures. Compare quality-adjusted CPC (see Labor Arbitrage and Global Workforce Optimization).
- Scope of "fully loaded": If one operation includes facilities and technology in their CPC and another does not, the comparison is wrong. Define the cost boundary before benchmarking.
Practical benchmarking approach:
- Define homogeneous contact types
- Establish consistent cost boundaries (what is "fully loaded")
- Normalize for channel mix
- Adjust for geographic cost differences
- Compare CPR (not just CPC) to capture quality differences
- Track trends over time, not just point-in-time snapshots
Executive Communication
When presenting unit economics to C-suite:
Lead with CPR, not CPC. CPC invites the question "how do we make this cheaper?" CPR invites "how do we resolve issues more efficiently?" The second framing protects quality.
Show the shrinkage multiplier explicitly. Executives who see "$42,000 salary" and "$7.36 cost per contact" do not understand the connection. Walk them through: salary → loaded cost → productive hours → contacts → CPC. The 61% shrinkage markup is always a revelation.
Frame marginal cost for capacity decisions. When volume increases 10%, the answer is not "10% more agents." Show the marginal cost curve — existing capacity can absorb the increase at 40-60% of average CPC if occupancy is below 85%.
Use revenue-per-contact for revenue-generating queues. Reframe from "we need 20 more retention agents at $79K each" to "each unstaffed retention position leaves $450,000 in annual save opportunity on the table."
Maturity Model Position
Unit economics capability maps to the WFM Labs Maturity Model:
- Level 1 (Reactive): No unit economics tracked. Budgets based on headcount × salary.
- Level 2 (Managed): CPC calculated quarterly. Used for annual budgeting.
- Level 3 (Optimized): CPC, CPR, and CPPH tracked monthly by channel. Used for operational decisions.
- Level 4 (Strategic): Marginal cost modeling. Revenue-per-contact for sales/retention queues. Unit economics drive channel strategy and automation investment.
- Level 5 (Predictive): Real-time unit economics dashboards. Predictive CPC modeling integrated with demand forecasting. Unit economics embedded in all workforce investment decisions.
See Also
- Workforce Cost Modeling
- WFM Role in Labor Cost Management
- ROI Frameworks for WFM Technology
- Cost of Delay in Staffing Decisions
- Total Cost of Workforce Ownership
- Labor Arbitrage and Global Workforce Optimization
- Workforce Investment and Human Capital ROI
References
- COPC Inc., Customer Experience Standard, Release 7.0 — cost-per-contact benchmarking methodology.
- Gartner, Contact Center Cost Optimization Framework, 2024 — channel economics and unit cost benchmarks.
- McKinsey & Company, The Next Horizon for Contact Centers, 2024 — AI agent economics and channel migration analysis.
- Deloitte, Global Contact Center Survey, 2024 — cross-industry CPC benchmarks by geography and channel.
- MetricNet, Contact Center Benchmarking Database, 2025 — cost-per-contact and cost-per-resolution benchmarks across 4,000+ centers.
