Variance analysis is the comparison of actual financial results against budget or forecast, broken down by driver so each gap points at a person and a response. The written explanation of why the gap occurred, and what it changes, is the deliverable.
How variance analysis works
The calculation is actual minus budget, or actual minus latest forecast, expressed in dollars and as a percentage. A variance is favourable when it helps profit and unfavourable when it hurts it, so revenue above plan and costs below plan are both favourable. The actuals side comes out of the ledger and has to be traceable back to source documents, which is the point of the record-keeping standards the ATO sets for business records.
The label is where people stop thinking too early. A favourable marketing variance is usually a campaign that did not launch, which is a delay, not a saving. An unfavourable cost of sales variance driven by volume above plan is a good month. Treating the sign as the answer is the most common way variance packs mislead a board.
Break the gap into drivers, not line items
A single number against a general ledger code tells you almost nothing. The useful split is by driver: volume against price against mix on the revenue side, and rate against usage against timing on the cost side. That is what turns a variance into a decision, because each driver points at a different person and a different response.
| What the variance shows | The question it should trigger | |
|---|---|---|
| Revenue variance | Actual revenue against plan | Is the gap volume, price or mix, and is it timing that reverses next month or demand that changes the forecast. |
| Cost of sales variance | Direct costs against plan | Did unit cost move or did volume move, because only one of those is a margin problem. |
| Labour variance | People cost against budget | Rate or headcount, and whether an underspend is a hiring plan running late rather than a saving. |
| Operating expense variance | Overheads against budget | Is this a deferral, a permanent reduction, or a cost that was accrued in the wrong period. |
Set a materiality threshold before you start, in dollars as well as percentage, and only write commentary above it. Explaining every line is how a variance pack gets long enough that nobody finishes it.
The commentary is the deliverable
A variance column with no explanation attached invites the reader to invent their own reason for the movement. What I push for is finance reporting that links directly to the business, moving beyond just focusing on the actuals to understanding what the numbers mean for stakeholders from a commercial perspective, so you can make recommendations and inferences about future actions.[1] That is the test. Not whether the variance is correct, but whether a non-finance reader finishes the page knowing what changed, why, and what happens next.
Variances only mean something against a statement people can already read, which is why what a profit and loss statement is actually telling you comes first.
Who owns it
Variance analysis normally sits with management accounting or FP&A rather than with the person who closed the ledger, and the reason is independence as much as skill. A senior site finance role typically covers management accounting, FP&A, budgets, working capital and business partnering, reporting directly into local leadership with a dotted line to the group financial controller in head office.[2] That dual line is deliberate: the commentary has to be credible to the operators and to the group at the same time.
In a company small enough that one person does the close and the commentary, expect the commentary to be the half that slips. It has no deadline enforced by anyone outside finance, so it quietly becomes optional.
The explaining half of this job is a distinct career track. Here is what a finance business partner actually does.
What AI changes, and what it does not
This is one of the finance tasks automation genuinely reaches. AI can handle 80 to 90 per cent of basic finance work such as checking, reconciliation, analysis, forecasting or modelling, but the last 10 per cent needs a qualified human who understands what the challenge is and what good looks like for the outcome.[3] Drafting the variance narrative is squarely in that last slice, because the explanation depends on knowing that the campaign slipped and the contract signed late.
The hiring consequence is the one worth planning for. Based on the conversations I have with CFOs and senior finance leaders, finance teams will get leaner while expectations grow, and if AI tools can generate reports, flag variances or draft commentary, leaders will opt for smaller teams with stronger commercial influence.[4] If you are adding a person to a reporting function now, add the one who can hold the conversation, not the one who can build the file.
Common questions
What is variance analysis in finance?
Variance analysis compares actual financial results against budget or latest forecast and explains the difference, in dollars and as a percentage. A variance is favourable when it helps profit and unfavourable when it hurts it. The comparison itself is arithmetic; the analysis is the written explanation of what drove the gap and what it changes.
What does a favourable variance actually mean?
Less than people assume. A favourable marketing variance is often a campaign that did not launch, which is a delay rather than a saving, and it usually returns next quarter. An unfavourable cost of sales variance driven by volume above plan can accompany a strong month. Reading the sign as the answer is the most common way a variance pack misleads a board.
How do you break down a variance properly?
By driver rather than by ledger code. On revenue, split volume, price and mix. Costs split into rate, usage and timing. Each driver points at a different owner and a different response, which is what turns a variance into a decision. Set a materiality threshold in dollars as well as percentage first, and write commentary only above it.
Who should do variance analysis?
Normally management accounting or FP&A rather than the person who closed the ledger, partly for independence and partly because the job needs someone talking to budget holders. A senior finance role of this type usually covers management accounting, FP&A, budgets, working capital and business partnering, reporting into local leadership with a dotted line to the group financial controller. Where one person does both the close and the commentary, the commentary is the half that slips.
References
- What I push clients towards: finance reporting should link directly to the business, moving beyond just focusing on actuals to understand what the numbers mean for stakeholders from a commercial perspective, enabling recommendations and inferences about future actions.
- From a role brief I wrote: a senior site finance role typically focuses on management accounting, FP&A, budgets, working capital and business partnering, reporting directly to local leadership while having a dotted line to the group FC in head office for financial support.
- My read on where automation stops: AI can handle 80-90% of basic finance tasks like checking, reconciliation, analysis, forecasting or modelling, but the last 10% requires a qualified human who understands what the challenge is and what good looks like for the outcome.
- Based on conversations with CFOs and senior finance leaders, many of them recorded on The CFO Track: finance teams will get leaner, but expectations will grow. If AI tools can generate reports, flag variances or draft commentary, leaders will opt for smaller teams with stronger commercial influence.