Compensation Scenario Modeling

Type: data Origin repo(s): anycomp (most full-featured), people-analyst, calculus, decision-wizard Extraction readiness: needs extraction (heavy business logic bleed) Depends on: segmentation dimensions (to scope scenarios), Monte Carlo simulation (for stochastic scenarios), statistical analysis (for equity checks) Last reviewed: 2026-04-24

What it is

Define a compensation scenario — budget envelope, merit matrix, pay structures, proposed actions — then sandbox it against the real employee population to produce: cost impact, compa-ratio distribution shift, equity checks by demographic slice, AI-generated narrative summary, and an exportable packet (CSV, PDF, or push to downstream systems).

Data shape

• comp_scenarios — scenario metadata (owner, name, cycle, status, assumptions JSONB).
• merit_matrices — grid of performance × position → percentage raise (per-scenario overrides).
• pay_structures — grade/band definitions, min/mid/max per grade.
• scenario_actions — proposed per-employee actions (raise, promotion, re-grade, no action).
• scenario_results — computed impact metrics (total_cost, compa_ratio_distribution JSONB, equity_report JSONB).
• Optional scenario_simulation_runs — Monte Carlo output when stochastic inputs are used.

UI / surface shape

• Scenario list with status chips, last-run timestamp, cost delta.
• Scenario builder: budget input → segment scope → merit matrix editor → action preview → run button.
• Results dashboard: cost card, compa-ratio histogram (before/after), equity report table, AI narrative summary.
• Export: CSV, PDF, or direct-push API to downstream HRIS.

Variants in the wild

• Anycomp — most full-featured. VOI integration, scenario kit/templates, comprehensive export.
• Calculus — embeds inside Tier 1 recipes; less standalone UI.
• People-analyst — emphasis on Monte Carlo + stochastic inputs.
• Decision-wizard — wraps a scenario in the Kepner-Tregoe analytical frame (criteria-weighted).

Primary files in origin

• server/scenarios/ (anycomp)
• server/merit-matrix/ — matrix CRUD + application
• server/equity-check/ — slice-by-demographic reporting
• app/scenarios/[id]/ — scenario builder UI

Next-version notes

• The merit matrix editor is re-invented per-app; extract as a shared component.
• Equity check logic should be one canonical implementation in the toolbox (currently drifts between repos — different statistical thresholds).
• AI narrative generation is a good candidate for the canonical Budgeted Agent pattern (already documented as a hub canonical).

Related patterns

• P09 — Runtime Provider Registry (for export adapters)
• P40 — Flexible JSONB State Column (scenario assumptions and results)
• P104 — Idempotent Batch Webhook Receiver (for downstream HRIS push)
• Canonical: Budgeted AI Agent with Pause/Resume (for narrative generation)