Guide

How to write variance commentary that survives the board meeting

A template, worked examples and the phrases to stop using, so your commentary answers the follow-up question before anyone asks it.

Eddie Ningombam Updated Nov 2026

The short answer. Good variance commentary says, in two or three sentences, what moved and by how much, what drove it (traced to a source), whether it will repeat, what it means for the forecast, and who owns the next step. Lead with the number, name the driver rather than the account, and never write "due to timing" without saying when it reverses.

What commentary is for

Variance commentary has one job: let a reader who was not in the analysis make a decision without asking a follow-up question. Every follow-up a CFO or board member has to ask is a sign the commentary left something out. The usual gaps are the same four: the driver, the evidence, whether it repeats, and what happens next.

This guide goes deeper on Step 6 of How to explain a budget vs actual variance.

The template

Five parts, two or three sentences

1. [Line] was [amount] ([%]) [over/under] [budget/forecast].
2. Driven by [driver], from [source].
3. This is [timing / one-time / structural].
4. [Forecast impact, with amount and period].
5. [Owner] will [next step] by [date].

Parts 1 to 3 are facts. Parts 4 and 5 are judgment and action. Keeping them in that order makes it obvious to the reader which is which.

Worked examples

All figures below are illustrative.

LineCommentary
New ARR (under)New ARR was $180K (15%) under plan. Two enterprise deals worth $210K slipped from September to October per the CRM close dates; both are in contract review. This is timing. Q4 new ARR forecast rises by $210K if both close; the CRO confirms at the Oct 10 pipeline review.
Payroll (over)Payroll was $42K (3%) over budget. Three engineers started in August instead of the planned October (HRIS start dates), adding $36K, and FX on overseas payroll added $6K. The early starts raise full-year payroll by the same $36K; FX is now forecast at the current rate.
Software (over)Software was $8.9K (21%) over budget because an annual software renewal was billed in Q3 instead of Q4 (invoice dated Sep 2). This is timing and reverses in Q4. No change to the full-year forecast.
Cloud hosting (over)Cloud hosting was $31K (12%) over budget. Usage ran 12% above plan after the new reporting feature launched; the rate per unit was flat. This is structural. Q4 hosting is forecast $33K higher; engineering is reviewing query costs and reports back by Oct 20.
Marketing (under)Marketing was $55K (22%) under budget because the October conference sponsorship was paid in October, not September as phased. This is timing. No change to the full-year forecast.

Phrases that weaken commentary

Instead ofWrite
"Due to timing."What moved, from when to when, and the period it reverses in.
"Higher than expected costs."The driver: price, volume, mix, timing, one-time or FX.
"Driven by various factors."The two or three factors that make up most of the amount, with numbers.
"We are monitoring."Who is doing what, and when you will know.
"One-off."What the event was, and why it will not repeat.
"In line with expectations."Against what: budget, forecast or last year, and by how much.

One variance, three audiences

The same variance needs different levels of detail for different readers. Write the detailed version first; the others are summaries of it.

AudienceExample (illustrative)
Budget ownerCC001 was $27K over: contractor extension INV-4471 ($14.2K), annual software renewal billed early ($8.9K), September offsite ($3.9K). Please confirm by Oct 15 whether the contractor continues into Q4.
CFOOpex was $87K (1.5%) over. About $47K is timing or one-time and will not affect the year. The open item is a $14K a quarter contractor extension in CC001; decision due Oct 15.
BoardOpex finished 1.5% over plan, mostly timing. Full-year opex guidance is unchanged unless we extend one contractor engagement, which would add about $14K a quarter.

Favorable variances need commentary too

A line that comes in under budget is often spend that is late, not spend that is saved. Say which it is. "Marketing is $55K under because the sponsorship moved to October" tells the reader not to bank the saving. "Marketing is $55K under because we cancelled the sponsorship" is a different decision.

When you cannot fully explain it yet

Write what you know, what you do not, and when you will. For example: "Of the $27K, $23K is traced to two invoices and the offsite. The remaining $4K is under review with AP and will be confirmed by Friday." Partial and honest beats complete and guessed. A guessed explanation that turns out wrong costs more trust than the variance itself.

Review checklist before you send

  • Every material line leads with the amount and percentage
  • Each driver has a source someone could check
  • The drivers add up to the total, or the remainder is named
  • Each item is labeled timing, one-time or structural
  • Forecast impact is stated with an amount and a period, or "no change"
  • Open items have an owner and a date
  • None of the weak phrases above

Keep a commentary log

Store each month's commentary with the evidence behind it. When the same line moves next quarter, the first draft is already written and the reasoning is on record. This is one of the cheapest ways to reduce decision latency.

Frequently asked questions

How long should variance commentary be?

Two or three sentences per material line, and one sentence for everything grouped under "other". If a line needs more, the extra detail belongs in an appendix with the evidence.

Should commentary compare to budget, forecast or last year?

Pick one primary comparison and state it. Most teams use budget for the board and the latest forecast for internal reviews. Mixing comparisons without labels is a common source of confusion.

Who should write variance commentary?

FP&A drafts it, the budget owner confirms the driver, and the finance lead signs off on the forecast impact. The owner's confirmation is what makes the driver credible.

E
Eddie Ningombam
Founder, PhrasIQ

Building InSightOS, the decision infrastructure layer for enterprise FP&A teams. Previously in finance operations and data infrastructure. Writing about decision latency, financial reasoning, and what it takes for FP&A to own the strategy conversation.

Get new articles in your inbox
FP&A strategy, variance analysis, and decision infrastructure. No noise.
✓ You're subscribed. First issue incoming.