variance-commentary-writer

Variance Commentary Writer

This skill takes a budget-vs-actual (and optionally prior-year) data set and produces the variance analysis and written commentary that goes into a monthly management reporting pack. It computes favourable/adverse variances, builds a contribution bridge from budget to actual, screens movements against a materiality threshold so the narrative stays focused, and drafts commentary in a disciplined what / why / so-what structure. It is for FP&A analysts, financial controllers, and finance business partners who own the month-end commentary and want a consistent, defensible first draft.

When to use this

• Month-end or quarter-end close: actuals are in and you need P&L commentary for the reporting pack.
• Budget vs actual, forecast vs actual, or actual vs prior-period analysis.
• Preparing board or management committee slides that need a tight narrative, not just a table.
• You have many line items and need to filter to the few that matter rather than commenting on everything.
• Building a repeatable commentary template across business units or cost centres.

Inputs

Provide a tabular file (.xlsx or .csv). The bundled sample samples/budget_vs_actual_sample.csv shows the expected shape. Useful columns:

• entity / cost centre (optional, for multi-unit packs)
• period (e.g. 2026-05)
• line_item (e.g. Revenue, Marketing, Salaries)
• category (Revenue, COGS, Gross Profit, Opex, Finance, Tax, Other) for grouping and bridge logic
• actual, budget, and optionally prior_year (or forecast) as numeric values

Sign convention: state it explicitly. The sample uses revenue/income as positive and costs as negative (natural sign), so a single arithmetic rule applies throughout. If costs are stored as positives, confirm before computing favourability. Also tell the skill the reporting currency, the units (actuals vs thousands), and the materiality basis you want (see below). If not provided, ask once, then proceed with stated defaults.

How to do it

1. Validate and normalise the data

• Confirm every commentary line has actual and budget. Flag blanks, text in numeric columns, and obvious unit mismatches (e.g. one row in thousands, the rest in units).
• Establish the sign convention up front. Convert everything to natural sign (income +, cost -) internally so favourability is a single rule.
• Reconcile that subtotals tie: Gross Profit = Revenue + COGS; Operating Profit = Gross Profit + sum(Opex) + Other. If a provided subtotal does not equal the sum of its components, stop and report the break before writing any commentary. Commentary on numbers that do not foot is worse than no commentary.

2. Compute the variances

For each line, compute against budget (and against prior year if supplied):

• Absolute variance: Variance = Actual - Budget
• Percentage variance: Variance % = (Actual - Budget) / |Budget| (guard against Budget = 0; if zero, report the absolute movement and mark "n/m" for the percentage).
• Favourability on a natural-sign basis: a positive variance on an income/profit line is favourable (F); a positive variance on a cost line (cost becoming less negative, i.e. lower spend) is also favourable. The single rule: Variance > 0 on natural-sign data is favourable to profit; Variance < 0 is adverse (A). Always label F or A; never leave the reader to infer direction from a sign.

3. Build the budget-to-actual bridge

Decompose the movement in operating profit (or the headline metric) into additive contributions so the parts sum to the total:

• Start from Budget profit, walk through each driver (Revenue, COGS / margin, each material Opex bucket), and arrive at Actual profit.
• Contribution of each line to the profit movement equals its natural-sign variance. Verify the bridge ties: sum(line variances) = Actual profit - Budget profit exactly. If it does not tie, you have a sign or scope error.
• Where useful, split Revenue variance into price/rate vs volume if the data supports it (Volume effect = (Actual qty - Budget qty) x Budget price; Price effect = (Actual price - Budget price) x Actual qty); otherwise keep it as a single revenue line and say so.

4. Apply a materiality threshold

Discuss only movements that matter. Use a dual test and flag a line if it breaches either limb:

• Absolute: variance >= a stated currency amount (default suggestion: 5% of the relevant subtotal, or a round figure agreed with the pack owner).
• Relative: |Variance %| >= a stated percentage (common default 10%). Smaller items are aggregated into a single "other movements, net" line so the bridge still foots. State the chosen thresholds explicitly in the output so the reader knows what was filtered. Never silently drop a large absolute variance just because its percentage is small (a big base hides this), and never over-index on a large percentage on a tiny base (mark it "n/m" if immaterial in cash terms).

5. Draft commentary in what / why / so-what

For each material line, write a tight three-part note:

• What: the fact and the number. "Marketing was 186k vs budget 150k, 36k (24%) adverse."
• Why: the driver, in business terms, not accounting terms. Distinguish timing (phasing that reverses) from permanent (run-rate change). If the data does not tell you the cause, state the candidate drivers to investigate rather than inventing one. Do not fabricate explanations.
• So-what: the implication. Does it change the full-year outlook, signal a run-rate shift, require reforecast, or carry a cash or covenant consequence? If neutral (pure timing that reverses next month), say so.

Keep each note to two or three sentences. Lead with the largest adverse and largest favourable movements. Quantify everything; avoid vague adjectives ("significantly", "broadly in line") unless paired with a number.

6. Frame the headline

Open the commentary with a two-to-three sentence executive summary: headline metric vs budget (F/A and %), the one or two drivers that explain most of the gap, and the net read-through to the full-year position. This is what an executive skims.

Output

Produce a Markdown management commentary workpaper with these sections:

1. Header - entity, period, currency, units, sign convention, materiality thresholds applied.
2. Executive summary - 2-3 sentences (step 6).
3. Variance table - columns: Line Item | Category | Actual | Budget | Variance | Variance % | F/A | Material? (Y/N). Subtotals for Gross Profit and Operating Profit that foot to the line items.
4. Bridge - an ordered budget-to-actual walk (text or simple table) with each contribution and a check that the parts sum to the total movement.
5. Commentary - one what/why/so-what block per material line, grouped Revenue then Costs, ordered by magnitude.
6. Items flagged for follow-up - data breaks, unexplained variances, or lines needing input from the business.

If the user wants a spreadsheet, mirror the variance table as a CSV (one row per line item, plus subtotal rows) so it can paste into the pack. Keep currency formatting consistent and show the sign convention in a note cell.

Quality checks

• Subtotals foot, and the bridge ties exactly to Actual profit minus Budget profit.
• Every variance carries an explicit F/A label, not just a sign.
• Budget = 0 cases are handled with "n/m", not a divide-by-zero or a misleading large percentage.
• No fabricated explanations. Where the cause is unknown, the note says "to be confirmed with [owner]" and lists candidate drivers.
• Timing vs permanent is stated for each material item.
• The materiality thresholds are disclosed; aggregated immaterial items still appear in "other movements, net".
• Units and currency are consistent across every figure; thousands vs units is not mixed.
• Tone is neutral and factual; no overstatement, no speculation presented as fact.

Worked example

Using samples/budget_vs_actual_sample.csv (Aurora Components Ltd, 2026-05, natural sign, currency units, thresholds: absolute >= 25k or |%| >= 10%):

• Revenue 4,820k vs budget 4,500k: +320k (7.1%) favourable.
• Cost of Sales -2,980k vs -2,700k: -280k (10.4%) adverse, partly the cost of higher volume.
• Gross Profit 1,840k vs 1,800k: +40k favourable - revenue gain largely offset by margin compression worth investigating.
• Professional Fees -122k vs -60k: -62k (103%) adverse - the largest Opex miss; likely one-off advisory, classify as timing pending confirmation.
• Bad Debt Expense -45k vs -15k: -30k (200%) adverse - a specific provision; permanent unless recovered, flag credit-risk read-through.
• Marketing -186k vs -150k: -36k (24%) adverse; Travel -38k vs -30k below the absolute limb, aggregate into "other movements, net".

Executive summary draft: "Operating profit was broadly favourable to budget on a 7% revenue beat, but the gross margin compressed ~150bps and one-off professional fees plus a 30k bad-debt provision absorbed most of the upside. Excluding the likely-one-off advisory spend, the run-rate remains in line; the bad-debt provision warrants a credit-risk follow-up." Each material line then gets its what/why/so-what block, and the bridge confirms the line variances sum to the operating-profit movement.

Disclaimer

Disclaimer: This skill is a drafting and analysis aid, not professional advice. It does not provide accounting, audit, tax, investment, or legal advice. All output must be reviewed and approved by a qualified professional before use or reliance.

---

Part of LedgerSkills - security-vetted Claude Code skills for finance teams. New skills scanned and tested weekly; subscribe on the site.