WFM Equilibrium — WrightForge Lab

Demand

Daily call volume 3,000

Avg handle time 270s

Service level target 80/20

Supply

Agent headcount (FTE) 55

Hours per shift 7.0h

Shrinkage & Occupancy

Shrinkage rate 40%

Sick/Absent8%

Vacation10%

Training5%

Other7%

After-call work 45s

Break time / shift 60m

Cost model

Agent hourly rate$28.00 CAD

Cost basisBare wage

Excludes loaded costs (benefits, pension ~25-35%), technology, facilities, supervision, and BSU overhead allocation. True cost/call is typically 1.5–2× this estimate.

Operating hours

Daily window08:00–17:00

Interval length15 min

Intervals per day36

Working days / month22

Weekend staffing30% of FTE

Call distribution

Intraday patternDouble peak

Dual gaussian: peaks at 10:30 (σ=1.2h) and 14:00 (σ=1.4h), base load 30% of peak. Typical for municipal/government service centres.

Weekly weightsMon→Sun

20% / 18% / 17% / 16% / 15% / 8% / 6% — Monday-heavy, weekend-light pattern.

Monthly weights4 equal weeks

26% / 25% / 25% / 24% — slight front-loading.

Staffing model

Schedule typeUniform

All agents follow the same shift. Includes ramp-up (07:30–09:00), lunch dip (12:00–13:00, −15%), and ramp-down (16:00–17:30). No staggered or split shifts.

Shrinkage breakdownProportional

Sick 27% / Vacation 33% / Training 17% / Other 23% of total shrinkage rate. Displayed in sidebar, not independently adjustable.

Behavioral model

Caller patience180s avg

Exponential patience distribution. Abandon probability: P = 1 − e^(−ASA / patience). Industry range is 60–180s depending on call type.

IVR containment12%

Percentage of calls self-served before reaching queue. Used in Sankey flow only — does not affect Erlang C calculations.

Disposition split78/15/7

Resolved 78% / Transferred 15% / Escalated 7%. Used in Sankey flow only.

Target answer time20 sec

The denominator in "80/20" service level. SL% slider adjusts the numerator only.

Mathematical model

Queuing modelErlang C

Arrival processPoisson

Service timesExponential

Erlang C assumes M/M/N queue: Poisson arrivals, exponential service, N servers, infinite queue, no abandonment in the base formula.

SL = 1 − P(w) × e^(−(N−A) × t/AHT)
ASA = P(w) × AHT / (N − A)

Where N = agents, A = traffic intensity (arrival rate × AHT), t = target wait, P(w) = Erlang C probability of waiting. Abandon rate is modeled separately via exponential patience overlay.

Service Level

—

% in target

Avg Speed Answer

—

seconds

Abandon Rate

—

% gave up

Occupancy

—

agent utilization

Cost / Call

—

CAD

Effective Staff

—

net of shrinkage

Flowing Waiting At risk Abandoned Agent

Speed 12×

Queue: 0 Serving: 0 Resolved: 0 Abandoned: 0 Total: 0

Service Level

—

target: 80/20

ASA

—

avg wait (sec)

Abandon %

—

caller patience

Occupancy

—

agent load

Cost / Call

—

CAD

Eff. Staff

—

net FTE