Blog · Board and management reporting · Finance and FP&A teams
A variance dashboard turns cost-center budget gaps into an investigation list. Use a worked example to set review thresholds, split volume from rate effects, reconcile the drivers, and assign accountable owners.
A variance dashboard compares actual cost with budget, identifies material gaps, and connects each gap to its contributing rows and owner. It should let a finance reviewer move from a cost-center exception to a supported explanation. The most useful result is a review decision with an accountable person and a next step.
State the period, currency, baseline version, cost scope, and sign convention above the table. This example covers September 2026 operating expenses at a US distribution company, in USD, against its approved monthly budget. It uses actual minus budget: positive expense variance is unfavorable, and negative is favorable.
Dollar variance = Actual cost − Budget
Percentage variance = Dollar variance / Budget
The broader variance analysis method explains the underlying comparison. This dashboard concentrates on deciding which cost centers need attention and what explains their differences.
Start with the monthly budget export, the ledger expense detail, a cost-center mapping, and the operational driver summary where one exists. Each actual row needs a posting period, cost center, account, amount, and reference. The budget must use the same scope, including any exclusions such as depreciation or intercompany charges.
Do not join a monthly budget total to every invoice row and then sum it. Aggregate actuals to the review level first and attach each budget amount once. In a spreadsheet, SUMIFS can total ledger amounts by matching cost center and period; Microsoft's SUMIFS documentation describes the function's multiple-criteria behavior.
The finance team sets an illustrative review rule: investigate a cost center when the absolute gap is at least $5,000 and the absolute percentage gap is at least 5%. This is a chosen policy for the example, not an industry benchmark. Unbudgeted spending receives a separate review because a percentage against zero budget is undefined.
| Cost center | Budget (USD) | Actual cost (USD) | Actual − budget (USD) | Variance | Queue | Owner |
|---|---|---|---|---|---|---|
| Operations | 100,000 | 120,000 | 20,000 | 20.0% | Review | Operations director |
| Sales | 80,000 | 76,000 | -4,000 | -5.0% | Watch | Sales director |
| Support | 60,000 | 63,000 | 3,000 | 5.0% | Watch | Support manager |
| Facilities | 40,000 | 40,000 | 0 | 0.0% | Watch | Facilities manager |
| Total | 280,000 | 299,000 | 19,000 | 6.8% | Finance reviewer |
Operations meets both tests: $20,000 exceeds the dollar floor and 20.0% exceeds the percentage threshold. Sales and Support each meet the percentage test but fall below the dollar floor. Their rows remain visible, with no detailed investigation triggered by this rule alone.
The company variance is 299,000 − 280,000 = $19,000. Its percentage is 19,000 / 280,000 = 6.8%, rounded to one decimal place. Never average the cost-center percentages for the company result.
With the displayed table starting in A1, dollar variance in D2 is =C2-B2. Percentage variance in E2 is:
=IF(B2=0,"N/A",D2/B2)
The queue in F2 is:
=IF(B2=0,IF(C2=0,"Watch","Review"),IF(AND(ABS(D2)>=5000,ABS(E2)>=5%),"Review","Watch"))
Fill the formulas down through the cost centers, keeping the total percentage based on summed dollars. Microsoft's IF documentation explains returning different results for true and false conditions. Covirage is not affiliated with Microsoft.
Operations processes customer orders. The operating summary shows completed orders and a comparable cost per order, with the same expense boundary as the ledger total. This example assumes the entire Operations cost is proportional to order volume; fixed costs would need their own bridge component.
| Operations input | Budget | Actual |
|---|---|---|
| Completed orders | 1,000 | 1,200 |
| Cost per order (USD) | 100 | 100 |
| Cost: orders × cost per order (USD) | 100,000 | 120,000 |
Separate the volume effect from the rate effect using a fixed calculation order. Value the volume change at the budget rate, then value the rate change at actual volume.
| Driver | Calculation | Cost effect (USD) |
|---|---|---|
| Volume | (1,200 − 1,000) × 100 | 20,000 |
| Rate | (100 − 100) × 1,200 | 0 |
| Explained Operations variance | 20,000 + 0 | 20,000 |
The explanation ties exactly to Operations' $20,000 unfavorable gap. The director should confirm that the additional orders were required and that the output count uses the agreed definition. Higher expense caused by more completed work calls for a different response from higher expense per unit.
This does not establish whether the added activity was profitable. Revenue and service outcomes belong in the wider FP&A review. For a concise definition of a rate difference, see price variance.
Now keep every input unchanged except actual Operations cost per order, which rises from $100 to $110. Actual order volume remains 1,200. This variation isolates a rate change rather than quietly changing the volume assumptions as well.
| Variation step | Calculation | Amount (USD) |
|---|---|---|
| Revised Operations actual cost | 1,200 × 110 | 132,000 |
| Revised Operations variance | 132,000 − 100,000 | 32,000 |
| Volume effect | (1,200 − 1,000) × 100 | 20,000 |
| Rate effect | (110 − 100) × 1,200 | 12,000 |
| Driver check | 20,000 + 12,000 | 32,000 |
| Revised company actual cost | 132,000 + 76,000 + 63,000 + 40,000 | 311,000 |
| Revised company variance | 311,000 − 280,000 | 31,000 |
The company percentage becomes 31,000 / 280,000 = 11.1%, rounded to one decimal place. Operations' rate effect now deserves a separate explanation: labor mix, overtime, supplier charges, or another supported cost change. Those are investigation questions, not causes established by the ratio.
The bridge assigns the combined volume-and-rate interaction to the rate effect through actual volume. Keep that convention consistent across periods so a change in presentation does not masquerade as a change in operations.
Finance owns the tie between ledger expense, cost-center totals, and the company total. The operating owner confirms the driver evidence and proposes action. Record the cause, supporting references, next action, accountable role, and expected review date alongside the flagged item.
In the original example, the company bridge is 20,000 − 4,000 + 3,000 + 0 = $19,000. That identity must remain true after filtering or reclassifying rows. Preserve a favorable variance such as Sales' lower expense until its timing and completeness have been checked.
Upload your budget, actual-cost export, and supporting driver summary, or bring a Google Sheet, and Covirage maps the columns. Its tools compute and check the differences; the external AI model explains the result with every figure traced to its rows. Create a free Covirage account and bring a sample of your own data.
See cost variance analysis for the supporting investigation workflow.
Include the reporting period, budget version, cost center, budget, actual cost, dollar and percentage variance, review status, owner, and the contributing driver amounts. Keep a visible reconciliation between the summary and the underlying rows so an explanation can be checked before it becomes an action.
Choose a threshold that reflects the business's costs and review capacity. A dollar floor prevents tiny items dominating the queue; a percentage test shows the scale relative to budget. Define whether both tests or either test must be met, and handle unbudgeted spending separately.
Yes. Lower cost can reflect efficiency, delayed activity, vacancies, or incomplete accruals. The sign describes the comparison but does not establish the cause. Preserve favorable items and apply the stated review rules rather than removing them from the report automatically.
A financial dashboard summarizes overall performance across measures such as revenue, margin, and cash. A variance dashboard focuses the review on differences from a baseline, the rows contributing to those differences, and the people responsible for explaining and resolving them.