P&L variance analysis is the process of comparing actual revenue, costs, and profit with a budget, forecast, or prior period, then explaining the differences. The calculation is simple: actual minus plan. The value comes from what follows, which is isolating the business drivers behind each difference and deciding whether it calls for action. The Association for Financial Professionals describes price and volume variances as the standard way to show whether a revenue gap came from pricing or from units sold (AFP, variance analysis).
A variance is favorable when it improves profit and unfavorable when it reduces it. Watch the sign: actual minus budget is positive both when revenue beats plan and when spending runs over.
Most monthly variance packs list deltas rather than causes. "Marketing expense: $42,000 unfavorable (14%), due to higher than expected spend" restates the number in words without saying what happened or whether it will recur.
The underlying problem is usually time. The 2026 FP&A Trends Survey of 475 finance professionals found that FP&A teams still spend 47% of their time collecting and validating data. When nearly half the month goes to assembling numbers, explanation gets squeezed into the day before the review. A repeatable process protects that diagnostic time.
Confirm which version you are comparing against (original budget, latest forecast, or prior year) and that the period is fully closed. Late accruals and unposted journals create false variances. If budgets and actuals sit in different systems, check that account mappings and hierarchies match. Working from connected ERP and operational data removes much of this reconciliation risk.
Not every variance deserves commentary. Combine an absolute and a relative threshold, for example: investigate any line where the variance exceeds both $25,000 and 5% (illustrative figures; calibrate to your business). Requiring both keeps small accounts with big percentage swings from eating up the analysis. Write the rule down so it applies the same way every month.
This step separates diagnosis from description. If your budget is built on drivers such as units, rates, and headcount, these splits fall out of the model directly.
|
Variance driver |
How to calculate it |
Common root causes |
|
Price (revenue) |
(Actual price minus budget price) × actual volume |
List price changes, discounting, contract renewals |
|
Volume (revenue) |
(Actual volume minus budget volume) × budget price |
Lost or delayed customers, launch timing, demand shifts |
|
Mix (revenue) |
Effect of selling a different blend of products or customers than planned |
Shift toward lower or higher margin lines |
|
Rate (expense) |
(Actual cost per unit minus budget cost per unit) × actual quantity |
Supplier increases, wage changes, FX |
|
Usage (expense) |
(Actual quantity minus budget quantity) × budget cost per unit |
Overtime, waste, headcount above plan |
|
Timing |
Same annual amount recognized in a different month |
Invoice delays, deferred projects, accrual cutoffs |
A volume variance is still just a number. The root cause is the event behind it: a delayed launch, a lost account, a sales role left open for two months. Ask the budget owner a specific question tied to the driver ("Units were 8% below plan in the Northeast. What changed?") instead of a generic "please explain," then check the answer against ledger detail and operational data such as CRM pipeline and headcount.
A timing variance reverses on its own, such as an invoice expected in March that lands in April. A permanent variance will repeat, such as a supplier price increase. A one-time variance will not recur, such as a legal settlement. The classification decides what happens next: permanent variances flow into the forecast, timing variances are tracked until they reverse, and one-time items are footnoted so they do not distort trend analysis.
Strong commentary answers four questions: what moved, why it moved, whether it will recur, and what should be done. Compare these two versions of the same note (hypothetical figures):
Consider a hypothetical distributor that budgeted 20,000 units at $100 each, or $2,000,000 in monthly revenue. It actually sold 19,000 units at $103, or $1,957,000. The headline variance is $43,000 unfavorable, about 2%, which most review meetings would wave through.
|
Driver |
Amount |
Root cause |
Classification |
|
Price |
+$57,000 favorable |
List price increase took effect a month earlier than budgeted |
Permanent |
|
Volume |
-$100,000 unfavorable |
Two mid-size accounts paused orders after the early increase |
Unknown: at risk of becoming permanent |
|
Total |
-$43,000 unfavorable |
|
|
Decomposed, the picture changes. A $100,000 volume problem is hiding behind a price gain, and its cause may threaten two customer relationships. The right response is not a line of commentary. It is a call from sales leadership to both accounts this week and a forecast scenario that assumes they do not return.
Variance analysis is a core capability of corporate performance management (CPM) and enterprise performance management (EPM) software, which hold budgets, forecasts, and actuals in one model. Actuals typically come from ERP reporting, and multi-entity groups must finish consolidation, including intercompany eliminations, before group variances mean anything. Extended planning and analysis (xP&A), and its finance-led form, xFP&A, extends this work to sales, operations, and HR budget owners, shortening the distance between a number and the person who knows why it moved.
The six steps work in a spreadsheet, but steps 1 and 4 absorb most of the effort when data is spread across ERP, CRM, and payroll systems. Platforms that combine planning, reporting, consolidation, and analysis on one data warehouse remove the reconciliation step and let analysts drill from a P&L line to transactions.
Solver is one example: QuickStart integrations bring ERP and operational data into one data warehouse, and the Template Marketplace includes pre-built variance reports such as the P&L Estimate vs Budget report and a Revenue and Expense Exception Report that highlights variances by amount or percent. On top of that data, the Analysis Agent in Solver Copilot supports anomaly detection, trend identification, and root cause analysis.
Technology shortens the path to the cause. It does not replace steps 5 and 6: deciding whether a variance is permanent, and what the business should do about it, remains the analyst's judgment.
P&L variance analysis is useful when it ends in a cause and a decision. Confirm the data, set thresholds, break variances into drivers, trace them to real events, classify them, and write commentary that changes the forecast or triggers action. Done consistently, the monthly review stops being a recital of deltas.