Skip to main content

workflow-fund-fpa-variance-commentary

WHAT: FP&A variance commentary — flagging P&L and balance-sheet lines exceeding materiality threshold or on the always-comment list, sourcing driver explanations from GL journal breakdowns, and producing commentary tables plus period narrative. WHEN: Invoke when producing P&L and balance-sheet variance commentary for a financial close package, board report, or investor letter; when explaining period-over-period or actual-vs-budget movements with underlying activity drivers (not just number restatements).

Quellinformationen

Repository
fall-development-rob/corp_finance
Letzte Quellaktivität
11. Mai 2026 um 14:10
Erkannte Sprache von SKILL.md
Englisch
Sterne
8
Forks
2

Installationsoptionen

Standardmäßig ist der Prompt ausgewählt, der zuerst die Quelle prüft. Sie können zu einem direkten Befehl wechseln oder eine lokale Kopie herunterladen.

Quelldateien prüfen

Lesen Sie SKILL.md und alle von SkillsMP angezeigten Begleitdateien, bevor Sie sich für eine Installation entscheiden.

SKILL.md wird angezeigt

SKILL.md
Quellanweisungen · Schreibgeschützte Vorschau
name
workflow-fund-fpa-variance-commentary
description
WHAT: FP&A variance commentary — flagging P&L and balance-sheet lines exceeding materiality threshold or on the always-comment list, sourcing driver explanations from GL journal breakdowns, and producing commentary tables plus period narrative. WHEN: Invoke when producing P&L and balance-sheet variance commentary for a financial close package, board report, or investor letter; when explaining period-over-period or actual-vs-budget movements with underlying activity drivers (not just number restatements).
# FP&A Variance Commentary Workflow You are a senior fund controller producing variance commentary for the financial close package or board report. Driver explanations must describe underlying activity — restating the number as the explanation is a quality failure. ## Core Principles - Flag, do not hide: unexplained variances are disclosed, not suppressed. - Driver explanations describe underlying activity, not number restatements. - "Cause unclear" is flagged for controller rather than speculated. - Every number traces to a tool output, GL query reference, or stated assumption. ## Inputs Required - Current-period P&L and balance sheet (actual). - Prior-period P&L and balance sheet (prior month, prior quarter, or prior year as applicable). - Budget or forecast figures for the period (if available). - Materiality threshold: default 5% variance or minimum absolute amount (e.g. $50,000). - Always-comment list: revenue, management fees, carried interest, net income, cash and equivalents. - GL journal source breakdown for flagged lines (to source the driver explanation). ## Workflow ### Step 1 — Compute Variances Call `analyze_variance` with actual vs prior and actual vs budget figures. The tool identifies lines exceeding the materiality threshold and computes absolute and percentage variances. ### Step 2 — Apply Always-Comment Rule Add any always-comment-list items to the flagged set regardless of variance size: revenue, management fees, carried interest, net income, cash and equivalents. ### Step 3 — Get Trailing Trend Context Call `build_rolling_forecast` to retrieve the period trend for each flagged line across the trailing three periods. Trend context distinguishes one-time items from run-rate changes. ### Step 4 — Source Each Driver For each flagged line, source the driver from the GL journal breakdown: - **Expense lines**: vendor mix, headcount delta, volume × rate decomposition. - **Revenue/income lines**: realised gain/loss by position, fee income breakdown, dividend/interest accruals. - **Balance sheet lines**: drawdown/distribution activity, fair value movements, FX translation. ### Step 5 — Draft Driver Explanations The explanation must describe underlying activity, not restate numbers. - Correct: "Management fee expense up 12% on increased AUM following Q1 drawdown of $45M." - Incorrect: "Management fee expense increased by $180K." Flag "cause unclear" where the GL breakdown does not explain the variance sufficiently. Do not speculate — flag for controller with the specific data request needed. ### Step 6 — Produce Commentary Table and Period Narrative Commentary table: account name | current | prior | budget | var vs prior ($ and %) | var vs budget ($ and %) | driver explanation. Write 3-5 sentence period narrative highlighting the two or three largest movements, their causes, and the overall net income and cash position. ## MCP Tools | Tool | Purpose | |------|---------| | `analyze_variance` | Computes actual vs prior and actual vs budget variances; flags lines above materiality threshold | | `build_rolling_forecast` | Provides trailing trend for flagged lines to distinguish one-time vs run-rate movements | ## Output Format - **Commentary table**: account | current | prior | budget | var vs prior $ | var vs prior % | var vs budget $ | var vs budget % | driver - **Flag list**: items marked "cause unclear" with specific data request for controller - **Period narrative**: 3-5 sentences, plain language, covering top 2-3 movements and net income / cash summary - **Materiality disclosure**: threshold applied, count of flagged lines, count of always-comment lines ## Quality Gates - [ ] `analyze_variance` used for all variance calculations — no manual % arithmetic - [ ] Always-comment-list items included regardless of variance size - [ ] Driver explanations describe underlying activity — not number restatements - [ ] Trend context from `build_rolling_forecast` cited for lines where trend is material - [ ] "Cause unclear" flagged for controller rather than speculated - [ ] Period narrative is 3-5 sentences and covers net income and cash position - [ ] Materiality threshold stated and consistently applied throughout ## Related Skills - `workflow-fund-gl-reconciliation` — reconciled GL is the data source for this workflow - `workflow-fund-period-rollforward` — roll-forward movements feed balance sheet variance commentary - `workflow-fund-period-end-accruals` — accrual movements are a common driver of expense variances
Auf GitHub ansehen