Method · 5 min read

Actual vs budget: explaining the gap, not just measuring it

Published by Variance Works - Updated 20 August 2026

Revenue came in under budget. Everyone in the room already knows that. The useful question is which part of the shortfall is price, which part is volume, which part is mix - and which part is a product that was budgeted and never sold.

The question it answers

A budget variance is a subtraction: planned minus actual. It is true, it takes one formula, and it settles nothing. Two businesses can miss budget by the same amount for opposite reasons - one sold the planned volume at a lower price, the other held price and sold less - and the two need different decisions.

Actual vs budget analysis walks from the plan to the outcome one explanation at a time, on the same period. It is the same arithmetic as a price, volume and mix bridge, applied to two columns of one month instead of two months of one column.

What makes it trustworthy is that the effects add up to the observed variance, exactly. If they do not, the gap is shown as its own bar rather than folded into another effect.

The columns you need

Three columns are required. Without them there is no variance to measure:

• A scenario column - the one that tells budget, forecast and actual apart. Most planning exports have it, under a name like Scenario, Version or Data type. It is the column this analysis is built around.
• A period label - the month, quarter or year being compared.
• A net revenue amount.

Two more columns change what you get rather than whether it works:

• A product identifier and a quantity. These two are what make the gap explainable. Without them the variance can be measured but not split.
• A unit cost. Adds the margin variance next to the revenue one, with its own cost effect.

Choosing the two scenarios

You choose which label is the reference and which is the one being measured. Nothing is inferred from the column name, and nothing is guessed from the words in it.

That matters more than it looks. A file often holds three scenarios - budget, a mid-year forecast, and actual - and the interesting comparison is not always actual against budget. Actual against the latest forecast asks whether the re-forecast was any good. Forecast against budget asks how far the year drifted before anything happened. Same file, same tool, three different questions.

Two modes: bridge and total

A budget rarely arrives at the same grain as the sales ledger. Plans are often built by account, by family or by region, with no product line and no unit count.

When the file carries a product and a quantity, you get the full bridge: price, volume, mix, budgeted-but-never-sold, and sold-without-being-budgeted.

When it does not, you get the two totals and their difference - and the tool says why it stopped there. That refusal is deliberate. A price, volume and mix split computed from summed amounts would be a plausible number with nothing behind it, and a plausible wrong number is worse than no number: it does not look wrong.

The screen tells you which mode you are in before you compute, while you can still go and map a column you have.

What it will not do

It will not tell you whether the budget was reasonable. A variance measures the distance between two declared numbers; it has no opinion on either.

It will not merge two files. If your actuals and your plan live in separate exports, they have to be brought together before the analysis, because deciding how to match their rows is a business decision and not an arithmetic one.

It will not reallocate a gap it cannot explain. When the effects do not add up to the observed variance, the remainder gets its own bar and keeps its name.

Questions

Can I compare a forecast against a budget?

Yes. Any two labels in the scenario column can be compared, in either direction. The one you pick as reference is the one the effects are measured from.

What if a product was budgeted but never sold?

It gets its own effect, separate from price and volume. A product with no actual sales has no price change and no volume change, so putting it in either bar would distort both.

Does my file leave the browser?

No. The whole calculation runs in the page, on your machine. There is no upload, no account, and no server that could hold your figures. How to verify that in two minutes.

Can I keep the result next to my other analyses?

Yes, in a monthly review: one export mapped once, then price volume mix, trends, budget variance and scenarios read the same file, and each result can be added to a single review pack.