Variance analysis

Variance analysis includes a variance decomposed into price, volume, mix and timing components where the underlying data supports it, prior-period and budget comparisons shown side by side rather than a single number in isolation, root-cause notes for the largest variances each month, and a short list of items flagged as worth a follow-up conversation rather than buried in a report nobody reads past the summary line. The specific lines analyzed, whether it is cost of goods sold, a labor cost category, or a revenue line affected by pricing changes, is agreed with you and set out in writing before the first month's analysis runs.

Once the month closes and the budget-versus-actual comparison identifies which lines moved, we pull the underlying transaction detail for the largest variances and work through the decomposition: isolating what a rate or price change alone would have produced, what a volume change alone would have produced, and what is left over as mix or timing. For lines where a bill of materials, a rate card or a per-unit cost exists, the decomposition is more precise; for lines without that structure, we work from whatever detail the transaction history and any operational notes you provide can support, and note where the split is an estimate rather than an exact calculation.

Manufacturers with a bill of materials and a standard cost per unit get the cleanest decomposition, since price and volume effects on material cost are usually calculable directly from purchase and production records. Construction companies comparing actual job costs to a bid use this to see whether an overrun was a material price increase, more labor hours than estimated, or scope that crept beyond the original bid. Trucking and logistics companies use it to separate a fuel price spike from a volume increase in miles run. Any business, including a service business with variable labor costs, that sees a cost or revenue line swing enough to matter benefits from knowing the actual driver rather than guessing.

The most common problem is a variance reported as a single dollar figure with no breakdown, so a business owner assumes the whole swing is one thing, usually price, when it is actually a mix of volume growth and a real price increase that need different responses. We fix this by splitting the number wherever the data allows. Another common problem is forcing a clean price-volume-mix split onto data that genuinely does not support one, producing a number that looks precise but is not meaningful; we flag those cases explicitly instead of manufacturing false precision. A third is treating a timing difference, an invoice landing in the wrong period, as a real cost increase; timing effects get called out separately so they do not distort the operational read.

Variance analysis draws on transaction data from QuickBooks Online, NetSuite or Sage Intacct, and for manufacturers, from whatever inventory or production system holds bill-of-materials and standard cost data. The decomposition itself is typically built and delivered in Google Sheets or Excel so every formula is visible and auditable, or in Causal for businesses wanting the analysis integrated into a live forecasting model rather than a static monthly report. Access needed depends on which lines are being analyzed and is agreed with you before work starts, since a material-cost variance and a labor-cost variance draw on different underlying systems.

A decomposed variance report shows the total swing on a line, then breaks it into a price component, a volume component, a mix component where relevant, and a timing component where relevant, with the components summing back to the total so nothing is left unexplained. Each component that crosses a meaningful size gets a one or two sentence root-cause note tied to what actually happened, not a generic statement that costs went up. A short list at the end of the report names the items worth a direct conversation, so the report ends with an action list rather than just a wall of numbers.