workload-analyzer

name: workload-analyzer description: "Analyzes federal contract workload attachments (Section J spreadsheets, PWS data tables, site lists, demand history) to identify pricing risks, demand trends, geographic concentration, and hidden modeling assumptions — then produces a structured handoff to price-to-win. USE WHEN the user uploads or references a workload spreadsheet, sites list, demand history table, staffing table, CLIN volume estimate, or any Section J attachment with quantitative data. Also triggers on 'analyze this workload', 'what does this data tell us about staffing', 'help me understand site scope', 'workload trend', 'CM volume', 'site count', or 'what should I price for option years'. Builds on data-analyzer's statistical foundation — do not use data-analyzer directly for govcon workload data; use this skill." license: MIT metadata: personas_primary: cost_estimator personas_secondary: [capture_manager] shipley_phases: [capture, strategy] capability: analyze skill_role: standalone skill_family: pricing skill_family_label: Pricing & Workload category: govcon-platform version: 1.0.0 status: active runtime: tools

Workload Analyzer

Govcon-specific workload analysis skill that classifies any Section J quantitative attachment into a known data archetype, applies the matching analytical framework (borrowing from data-analyzer's EDA and comparative workflows), and emits a structured pricing-risk summary for price-to-win consumption.

Foundation: This skill composes data-analyzer's statistical methods — see data-analyzer/references/statistical-methods.md for test selection and data-analyzer/references/pitfalls.md for analytical guardrails. Do not re-implement those; reference them when depth is needed.

References (load when needed):

• references/archetypes.md — 4 workload data archetypes + recognition signals
• references/analysis-frameworks.md — per-archetype analytical playbook
• references/ptw-handoff.md — structured output envelope format for price-to-win
• assets/report_template.md — report structure for workload analysis output

Workflow 0: Schema Recognition (always run first)

Objective: Classify what data you have before doing any analysis. Let the data structure and content determine the approach — never assume column names or schema.

1. Inventory the data — count sheets/tabs, rows, columns; note column headers verbatim; identify any totals rows, notes, or footnotes
2. Classify each sheet into one or more archetypes (load references/archetypes.md):
   • Geographic Scope
   • Demand/Volume History
   • Manning/Staffing
   • CLIN/Task Structure
3. Identify the join key — the field that links sheets together (site code, CLIN ID, location name, etc.)
4. Extract hidden assumptions — footnotes, adjustment factors, phone-resolution percentages, exclusion notes, any multipliers applied to raw data
5. State what you have — brief plain-language summary of data shape, date range, coverage, and any gaps before proceeding to analysis

If multiple archetypes are present, run the relevant workflows and synthesize in Workflow 3.

Workflow 1: Geographic Scope Analysis

Applies to: Geographic Scope archetype — site lists, location tables, AOR assignments.

1. Count and classify sites — total count, CONUS vs. OCONUS split, breakdown by country/state/region; flag any unusual locations (remote islands, combat zones, areas with restricted access)
2. Identify geographic concentration — what percentage of sites are in which regions; any single country or state that dominates scope
3. OCONUS risk tier each location — use travel burden, SOFA/access constraints, per-diem differential, and labor availability as risk dimensions; load references/analysis-frameworks.md for the tiering rubric
4. Flag scope anomalies — duplicate entries, missing location fields, ambiguous site names, sites that appear in one sheet but not another
5. Produce scope summary — total sites, CONUS/OCONUS counts, high-risk OCONUS count, concentration index

Workflow 2: Demand / Volume Analysis

Applies to: Demand/Volume History archetype — case counts, ticket volumes, service requests, PM/CM events, or any unit-of-work × time × location table.

1. Establish the time series — identify years/periods covered; compute totals per period; calculate year-over-year growth rate
2. Apply adjustment factors — locate any phone-resolution ratios, exclusion percentages, or multipliers noted in footnotes; apply them to get field-touch volume; document each factor explicitly
3. Run Pareto on site-level volume — which sites drive 20/50/80% of total volume; identify the high-volume outliers by name
4. Trend analysis — is volume growing, flat, or declining? Apply data-analyzer Workflow 2 (Pattern Detection) to the time series; project volume into option years using observed growth rate
5. Volatility analysis — coefficient of variation per site; flag sites with year-over-year swings >30% as pricing risk
6. Cross-tab with geography (if sites data available) — join on the shared key; compute volume per OCONUS site vs. CONUS site; surface the OCONUS sites with highest volume as highest-exposure pricing items

Workflow 3: Synthesis & PTW Handoff

Objective: Combine workflow outputs into a ranked pricing-risk list and structured handoff for price-to-win.

1. Build the risk register — each site or CLIN gets a risk score combining:
   • Volume tier (high/medium/low based on Pareto)
   • Geographic tier (OCONUS remote / OCONUS standard / CONUS)
   • Trend direction (growing / flat / declining)
   • Volatility flag (high variance = pricing risk)
2. Identify the top pricing traps — items where PWS projections appear flat but historical data shows growth; adjustment factors that inflate or deflate apparent demand; scope with high OCONUS concentration
3. Validate PWS vs. actuals (if both are present) — compare government estimates against historical actuals; calculate the gap; assess whether the government is underestimating, overestimating, or has a plausible rationale
4. Produce option-year projections — apply observed growth rate to base-year estimate; show the range (conservative/base/optimistic) for each option period
5. Format PTW handoff — load references/ptw-handoff.md; emit the structured envelope; call out any assumptions the estimator must validate

Deliverable: Use assets/report_template.md for the output structure.

Quick Reference

Guardrails

• Never assume schema — run Workflow 0 first, every time; column names vary by contract
• Surface every adjustment factor — hidden multipliers are pricing traps for the estimator who only reads the totals row
• Distinguish statistical significance from pricing significance — a site averaging 0.5 CMs/year is statistically unremarkable but pricing-significant if it is in Diego Garcia
• Do not extrapolate beyond observed range without flagging it — option-year projections are estimates; label confidence explicitly
• Cross-reference data-analyzer/references/pitfalls.md before drawing trend conclusions from fewer than 3 data points