Budget vs Actual, How to Run a Monthly Variance Review That Changes Decisions

A budget vs actual review compares what each line of the budget planned with what really happened last month, explains every material difference, and ends with decisions. Run it within a week of month end, focus on a few material variances, and record an owner for each action.

Plenty of companies produce a budget vs actual report every month. Far fewer get anything out of it. The report lands in an inbox, someone notices marketing went over, and the conversation stops there. A good monthly variance review is different: it's a short meeting with a fixed agenda that leaves the company with changed spending, changed forecasts, or a conscious decision to accept the gap.

What budget vs actual really measures

Variance is simply actual minus budget for each line, usually shown as an amount and a percentage. The sign convention matters, so pick one and stick to it. Most finance teams show variances as favorable or unfavorable rather than just positive or negative, because higher revenue is good and higher costs are bad, and a column of mixed signs confuses everyone who isn't in finance.

In the table above, a positive number is favorable. Revenue came in 24 thousand under budget, which is unfavorable. Payroll came in 6 thousand under budget, which is favorable. Paid marketing ran 13 thousand over, unfavorable. It's a simple convention, but writing it at the top of the report prevents a surprising number of arguments.

The comparison only works if both sides use the same basis. If the budget was built on accrual accounting, compare it with accrual actuals from a closed month. If your books aren't closed yet, label the actuals as preliminary and expect some movement.

Materiality, or how to avoid reviewing 200 lines

Nobody should walk through every account in a monthly review. Set a threshold and only discuss lines that cross it. A common approach is a double test: the variance must be above a fixed amount and above a percentage of the budget line. A 40 percent overrun on a 500 dollar line doesn't matter. A 3 percent miss on revenue might matter a lot.

• Revenue and gross margin get reviewed every month regardless of size.
• Payroll and headcount get reviewed every month, because small differences compound.
• Everything else gets reviewed only when it crosses both thresholds.

This keeps the meeting short and keeps attention on the numbers that move cash and runway.

The four questions for every material variance

For each line that crosses the threshold, the review answers four questions, in this order.

Is it real or is it timing

A large share of variances are timing differences. An invoice was booked in October instead of September. A conference was budgeted in March and happens in April. Timing variances reverse on their own, and the right action is usually to note them and move on. Be honest about it though: "timing" that never reverses is a real overspend.

What drove it

Break a real variance into its drivers. For revenue, that usually means volume versus price: fewer new customers than planned, or the same number at a lower average deal size. For payroll, it's headcount versus rate: hires starting later than planned, or salaries above the plan. For paid marketing, it's spend versus efficiency. Driver analysis turns "we missed" into "we closed nine deals instead of twelve, at the planned price".

Will it continue

This is the question that actually changes decisions. A one-off overspend doesn't need a new forecast. A trend does. If the last three months of paid marketing all ran 25 percent over, the budget isn't a useful benchmark anymore, and either the spend or the budget needs to change.

What are we doing about it

Every material variance ends with one of three outcomes: an action with an owner and a date, a forecast update, or an explicit decision to accept it. Write it down. Next month's review starts by checking those actions.

A monthly variance review agenda that works

1. Headline numbers, five minutes. Revenue, gross margin, operating expenses, net burn and cash, budget versus actual.
2. Last month's actions, five minutes. Were they done, and did they work?
3. Material variances, twenty minutes. The four questions for each line above the threshold.
4. Forecast impact, ten minutes. What the variances mean for the rest of the year, runway and the cash floor.
5. Decisions and owners, five minutes. Read back the action list.

Forty-five minutes, inside the first week after the books close. Department heads attend for their own lines and leave once their part is done.

A worked example from the table above

Let's run the four questions on the example month. Assume a materiality test of at least 5 thousand USD and at least 10 percent, with revenue and payroll always reviewed.

Revenue, 24 thousand under budget. Is it timing? Partly: one annual renewal worth 9 thousand was signed on the 2nd of the following month instead of the 28th. The remaining 15 thousand is real. The driver is volume, three fewer new customers than planned at roughly the planned price. Will it continue? The pipeline for next month looks similar, so probably yes. Action: the sales lead reviews conversion from demo to close, and finance lowers next quarter's new-customer forecast by two per month until conversion recovers.

Payroll, 6 thousand under budget. Always reviewed, below the threshold. The driver is a hire who started two weeks later than planned. It's favorable now and disappears next month. No action beyond updating the start date in the headcount plan.

Paid marketing, 13 thousand over budget. Crosses both thresholds. It's real, not timing: a campaign kept running after launch week. Combined with the revenue miss, it's the most important line of the month, because spend went up while new customers went down. Action: the marketing lead caps the campaign at the budget and reports cost per acquired customer by channel next month.

Software, 4 thousand over. Above 10 percent but under the 5 thousand amount threshold, so it isn't discussed in the meeting. Finance notes it, and if it repeats, it'll cross the line soon enough. Travel, 3 thousand under, is below both thresholds and isn't discussed at all.

Four lines, ten minutes, three actions with owners. That's what a useful review looks like.

Reforecasting without rewriting the budget

A frequent point of confusion: should you change the budget when reality moves? Usually not. The annual budget stays as the original commitment, the thing you measure performance against. What you update is the forecast, a separate version that reflects your current best estimate of the rest of the year.

Mature teams show three columns side by side: budget, latest forecast and actual. Budget vs actual tells you how the plan held up. Forecast vs actual tells you how good your recent predictions are. The variance analysis approach works the same way for both comparisons, and the second one is often more useful for running the company month to month.

A rolling reforecast after each close is the norm in growing companies. It doesn't need to be a big exercise: carry the actuals in, adjust the few drivers that changed and let the model recalculate the rest. When it's connected to scenario planning, you can see in the same meeting what a slower sales quarter does to runway.

Common reasons the review fails

• It happens too late. A review in week three of the following month is discussing history. Close faster or review preliminary numbers.
• The budget was never realistic. If every month shows the same large variance, the problem is the budget, not the team.
• Nobody owns the lines. Each budget line needs a person who can explain it. Finance can't explain marketing's spend on marketing's behalf.
• The chart of accounts doesn't match the budget. If the budget says "events" and the books spread events across four accounts, someone has to map them by hand every month, and that's where errors creep in.
• Actions aren't tracked. If last month's decisions don't appear at the start of this month's meeting, people learn that decisions don't matter.

Explaining variances to people outside finance

Department heads and founders don't want to read a table of accounts. Give them a sentence per variance: what happened, why, whether it continues and what you're doing. For example: "Paid marketing ran 13 thousand over budget because we kept the search campaign running after the launch week; it continues unless we cap it, so we're capping it at the budgeted 48 thousand from November."

Writing those sentences is tedious, which is exactly why they often don't get written. It's also one of the most practical uses of AI in finance today: a model reading the ledger, the budget and the drivers can draft the explanation for each variance, and a person reviews and corrects it. Budgeter's budget vs actual view does this for every line above your threshold, with the transactions behind each number one click away, so the explanation is checkable rather than a black box.

Tools that keep the review fast

You can run the whole process in spreadsheets. It just takes time every month: export the trial balance, map accounts to budget lines, paste into the report, rebuild formulas, update charts, write the commentary. Most of that is mechanical.

FP&A software removes the mechanical part. It syncs actuals from your accounting system read-only after close, maps them to budget lines once, calculates variances against both budget and forecast, and keeps the history of every version. That leaves the meeting for the part people are actually good at, deciding what to do. If you're weighing that step, the plans and their limits are on the pricing page.

The format matters less than the rhythm. Same agenda, same week, same thresholds, same owners. After three or four months, the review stops feeling like an audit and starts feeling like steering.