CO&S Capacity Planning

Executive workforce model · 7,032 FTE · $900M AOP · Client Onboarding & Service
Monthly Business Review
Planning Horizon · Jan – Jun 2026
Portfolio Work Sample
J. Brownstein · VP Quant Analytics
Training data loaded — run models to fit estimatesRun models to fit estimates on this section
← Dashboard

Demand Forecast

How much work is coming? ARIMA case volume · rolling forecast vs actual · confidence intervals

How much work is coming?

Case volume drives every staffing decision. This module forecasts monthly onboarding & service demand and tracks forecast accuracy against realized volumes.

  • Jan 2026 outlook: 139K cases/mo (135–144K at 95% CI) — 24-mo In-line with plan @ +3% CAGR.
  • Forecast reliability: 2.2% backtest MAPE over 120 months — accurate enough to staff against.
  • Downstream link: Demand feeds the OLS capacity model → 7,325 required FTE under base.
ARIMA Demand Forecast

139K cases in the first forecast month (135–144K at 95% CI) — 2.2% backtest MAPE, reliable enough to staff against.

AIC 820.66-mo horizon24-mo In-line with plan @ +3% CAGR
MAPE 2.2%
History Forecast + CI Realized vs planDashed line = forecast origin
First month139KPlanning anchor for OLS FTE
95% CI135–144KJan ’26 band width
Backtest MAPE2.2%Within staffing tolerance
ARIMA → OLS FTE139K cases/mo drives 7,325 required FTE

Case volume — forecast vs realized actual

Forecast issued Dec 25 · compare as months elapse · 4 of 6 realized months included

Rolling MAPE0.9%
MAE1K
Bias0 K cases
CI hit rate100%
4 mo · through Apr 26

Supporting charts

Each chart includes a one-line takeaway. Click any visual for full methodology and coefficients.

Forecast vs actual (rolling)

Rolling holdout — tracks whether realized volumes stay inside the forecast band.

Executive

ARIMA backtest

Holdout MAPE validates the auto-selected ARIMA order before staffing decisions.

Technical

Seasonal pattern

Seasonal indices explain Q1 filing peaks and year-end volume surges.

Technical

Error distribution

Symmetric errors around zero — no systematic over- or under-forecast bias.

Technical

Scenario demand paths

Stress paths for surge and downturn scenarios feed capacity sensitivity.

Technical

Model Confidence

Demand forecast87%
MAPE 2.2%
FTE regression95%
R² 0.95
Supply forecast92%
MAPE 1.6%
Causal (DiD)86%
Directional p=0.253