Variance analysis is the process of comparing actual financial results to expected results — budget, forecast or prior period — and explaining the drivers of difference. It is the single most common FP&A activity and the foundation of performance conversations across every organization.
This guide covers what variance analysis is, why it matters, the types of comparisons, how to decompose variances into actionable drivers, common mistakes and what good variance analysis looks like in practice.
What Is Variance Analysis?
At its simplest, variance analysis answers the question: why are results different from what we expected? The "expected" baseline can be a budget, a forecast, a prior period or a target. The variance is the difference. The analysis is the explanation.
The value of variance analysis is not in calculating the number — any spreadsheet can subtract two cells. The value is in explaining the why, quantifying the drivers and recommending what to do about it.
Types of Variance Comparisons
Budget vs Actual
The classic comparison. How did we perform against the plan we set at the start of the year? Best for accountability.
Forecast vs Actual
How did we perform against our most recent expectations? Better for understanding predictive accuracy and identifying blind spots.
Prior Period / Prior Year
How are we trending compared to last month or last year? Reveals trajectory and seasonality patterns.
Forecast vs Forecast
How are our expectations changing over time? Shows whether the business is improving or deteriorating relative to prior outlooks.
Decomposing Variances Into Drivers
Stating that revenue was $200K below budget is reporting. Explaining that $150K was driven by lower volume in the enterprise segment and $50K by a pricing concession on a renewal is analysis. The goal is to decompose every material variance into its constituent drivers.
Common decomposition frameworks include price vs volume (was the variance driven by how much you sold or what you charged?), mix effects (did the product or customer mix shift?), timing (did revenue or expense shift between periods?) and one-time vs recurring (is this a structural change or a one-off?).
Each decomposition should lead to a "so what" — an insight that either confirms the plan is working, identifies a risk that requires action or reveals an opportunity worth pursuing.
What Good Variance Analysis Looks Like
Materiality-driven
Focus on variances that matter. Use thresholds (both dollar and percentage) to avoid spending time on noise.
Driver-based
Decompose into root causes rather than describing the math. 'Revenue was down $200K' is a fact. 'Enterprise volume declined 12% due to delayed pipeline conversion' is analysis.
Forward-looking
Every variance explanation should include an outlook: is this expected to continue, reverse or accelerate? Connect backward-looking analysis to forward-looking forecast updates.
Action-oriented
The best variance commentary recommends action. 'We recommend accelerating marketing spend in Q3 to recover pipeline coverage' is more useful than 'we will continue to monitor.'
Common Mistakes
• Restating the numbers — 'revenue was $200K below budget' is not analysis.
• Explaining every line — focus on material variances, not every row in the P&L.
• Missing the 'so what' — every variance should connect to an action or outlook.
• Late delivery — analysis that arrives after decisions are made has limited value.
• Only looking backward — variance analysis should update the forecast, not just explain the past.