Variance analysis: sales, revenue or margin bridge?

Revenue is 6,000 € above budget, margin is 800 € below. Spotting these two variances takes a second. Explaining them is the whole job of variance analysis. Yet in many finance teams, "variance analysis" boils down to the "variance" column of the monthly report. A difference, with no cause. To move from observation to explanation, variance analysis relies on decomposition methods: bridges. A sales bridge explains a volume variance. A revenue bridge explains a revenue variance. A margin bridge explains a profitability variance.

This page defines variance analysis, presents three bridges that explain it, and explains when to use them.

In short

• Variance analysis is the overall approach: it compares an actual figure to a reference (budget, forecast, prior year), measures the difference, then explains it.
• A sales bridge explains the volume variance (units, orders, customers) through commercial levers: number of customers or points of sale, purchase frequency, rate of sale.
• A revenue bridge values those volumes and breaks the revenue variance down into volume, price and mix effects.
• A margin bridge breaks the profitability variance down into volume, price, mix and cost effects.

The three bridges nest into one another and share the same format: a waterfall chart that starts from the reference, stacks each effect, and lands on the actual.

What is variance analysis?

Variance analysis compares the actual value of a KPI to a reference value, over the same period and the same scope, then explains where the difference comes from. It runs in three steps.

1. Measure the variance. It is a simple calculation:

Variance = Actual - Reference

It is expressed in value ("800 € below budget") or in percentage ("-3.3% vs budget").

2. Explain the variance. The variance is broken down into additive causes: customers, volume, price, mix, costs. That is the job of the bridge.

3. Decide. Each effect points to a lever and a team: customers and distribution to the sales force, price to pricing, mix to the offer and category management, costs to procurement and operations.

Stopping at the first step produces a variance, not an analysis. The variance tells you how much. It does not tell you why.

Choosing the reference

The reference sets the question the analysis answers:

Bridges: explaining the variance

A bridge breaks a KPI's variance down into additive effects: together they add up exactly to the measured variance. The principle never changes. What changes is the KPI whose variance you explain, and therefore the equation that breaks it down.

The three bridges below nest into one another. The sales bridge details volumes. The revenue bridge values those volumes and adds price and mix. The margin bridge adds costs.

What is a sales bridge?

A sales bridge explains the variance in volumes sold: units, orders, bookings. It does not deal with price yet. It breaks volumes down by the commercial levers that produce them, and the equation depends on the route to market:

• Direct or B2B sales: Volumes = Active customers × Units per customer
• Retail: Volumes = Distribution × Sales per point of distribution (coverage and rate of sale)
• E-commerce: Orders = Sessions × Conversion rate

It isolates effects such as:

• Coverage effect: the variance due to the number of customers, points of sale or visitors.
• Intensity effect: the variance due to what each of them buys (frequency, units per customer, rate of sale, conversion).
• Mix effect: the variance due to a shift between segments (retailers, regions, customer types) that do not buy with the same intensity.

The sales bridge answers the question of the sales force and trade marketing: why are we selling more, or less, than planned? At innocent, for example, in-store sales are broken down into weighted distribution (on and off promotion), volumes sold per point of distribution, and price. A distribution loss at one retailer shows up immediately (see the use case).

What is a revenue bridge?

A revenue bridge explains the revenue variance. Its simplest form relies on a two-factor equation:

Revenue = Volume × Average price

It isolates three effects:

• Volume effect: the variance due to the number of units sold. This is the effect the sales bridge details.
• Price effect: the variance due to a genuine price change, within each segment.
• Mix effect: the variance due to a change in the composition of sales (more entry-level products, a country gaining weight), with prices unchanged.

The revenue bridge answers the question of pricing teams and sales leadership: where does our revenue variance come from? It adapts to any business model. At TheFork, for example, the equation becomes Bookings × Average booking value, presented every week to the executive committee against budget (see the use case). For an e-commerce site, it is more likely Sessions × Conversion rate × Average basket.

The detailed calculation (formulas, separating mix from price, the double-counting trap) is covered in our complete guide to price volume mix analysis.

What is a margin bridge?

A margin bridge explains the margin variance. The equation brings in costs:

Margin = Volume × (Average price - Average unit cost)

On top of the volume, price and mix effects, it isolates:

• Cost effect: the variance due to a change in unit cost (materials, logistics, labour), with composition unchanged.
• Mix effect on margin: the variance due to the fact that the products or customers gaining weight do not carry the same margin rate as the others.

The margin bridge answers the question of the CFO and financial controllers: why is our profitability deviating from plan? It can be broken down as far as the cost structure allows (variable cost, fixed cost per unit, by site or by channel). The full method is in Stop getting your margin analysis wrong.

Side-by-side comparison

One month, four readings

Take a company that sells two products to business customers. Product A is premium (50 € price, 30 € unit cost). Product B is entry-level (20 € price, 16 € unit cost). No price and no cost changed between budget and actual.

Measuring the variance gives two contradictory signals: revenue beats budget, margin falls short. It does not say why.

The sales bridge explains the +600 units:

• Customer effect: +500 units (100 more active customers, at 5 units each as in the budget).
• Units-per-customer effect: +100 units (each customer buys 5.2 units instead of 5).

The revenue bridge explains the +6,000 €:

• Volume effect: +21,000 € (the 600 extra units, valued at the budget average price of 35 €).
• Mix effect: -15,000 € (product B rises from 50% to 69% of units, the average price drops from 35 € to 29.23 €).
• Price effect: 0 € (no price moved).

The margin bridge explains the -800 €:

• Volume effect: +7,200 € (the same 600 units, valued at the budget average unit margin of 12 €).
• Mix effect: -8,000 € (product B, which earns only 4 € per unit, takes weight from product A, which earns 20).
• Price and cost effects: 0 €.

The three bridges do not contradict each other, they nest. The sales bridge shows that customer acquisition works: 100 more customers, buying slightly more than planned. The revenue bridge shows that this volume goes to entry-level products. The margin bridge shows what that mix costs in profitability. Read alone, none of them leads to the right decision. Read together, they do: the volume is there, the lever is the mix. What remains is to check whether the new customers are the ones pulling the mix towards entry-level, by breaking the analysis down by customer type.

Note also that the same volume movement takes three values: +600 units in the sales bridge, +21,000 € in the revenue bridge, +7,200 € in the margin bridge. That is expected: each bridge expresses it in the unit of its own KPI.

Which bridge to use, and when?

The general rule: in variance analysis, measurement is for triage, the bridge is for explanation. Measure every variance first to know where to look. Open a bridge only where the variance is significant. And choose the bridge according to the KPI your audience manages: volumes for a sales force, revenue for sales or pricing leadership, margin for finance leadership.

Three mistakes that undermine a bridge

1. An "other" or "unexplained" residual. If the effects do not add up exactly to the variance, the committee's first question will be about the line nobody can explain. That residual almost always comes from double counting between effects. A well-built bridge is additive, to the euro (the trap in detail).

2. A price effect that contains mix. If the average price falls because you sell more entry-level products, that is not a price cut. Presenting that variance as a "price effect" sends the pricing team looking for a problem that does not exist. In the example above, 100% of the average price variance is a mix effect.

3. A bridge in euros for a committee that thinks in percentages. If the executive committee reads its variances "in % vs budget", a bridge in euros forces them to redo the maths in their heads. Expressing each effect in percentage points of the variance makes the chart readable without commentary. That is what TheFork set up for its weekly review.

How Datama builds a variance analysis

In Datama Compare, all three bridges run on the same engine. Two parameters define the analysis:

• The equation (or market equation): Customers × Units per customer for a sales bridge, Volume × Average price for a revenue bridge, Volume × Average price × Margin rate for a margin bridge, or any other breakdown specific to your business (how to build yours). Nothing stops you from combining them into a single equation, for example Customers × Units per customer × Average price, to read the customer, intensity, mix and price effects in one waterfall.
• The comparison: what the actual is compared to (budget, forecast, prior year, or two segments). It can be defined in two ways, depending on the shape of your data.

Compare by dimension

Actual and reference share the same metric columns. A dimension column tells them apart: a "Type" (Actual, Budget, Forecast), a period (2025, 2026), or any other axis. You simply pick the two values to compare.

This is the natural shape of a transactional database or a history table. It also lets you compare two countries, two channels or two stores, not just two periods.

Compare by metric

Actual and reference sit in two separate columns: Revenue actual and Revenue budget, Sales and Sales prior year. This is the most common shape of FP&A exports and finance dashboards.

With the "Compare by Metric" option of the Datama extensions, these columns are compared pair by pair. No need to pivot the data to build a "Type" column. For a multi-step bridge, you create one pair per metric (Units budget / Units actual, Revenue budget / Revenue actual), in the order of the market equation.

What Datama calculates next

Whatever the shape of the comparison, Datama:

• reconstructs the variance exactly, with no residual;
• automatically separates the mix effect from the performance effect at each step;
• ranks dimensions (country, product, channel, customer) by explanatory power, so you know immediately where the variance sits;
• displays effects in value or in percentage points;
• exports the waterfall chart to PowerPoint for the committee presentation.

The only real preparation is the data. Budget often lives in Excel or an FP&A tool, actuals in a transactional database. They need to be aligned at the same grain (month, country, product), with additive metrics: customers, units, revenue, margin. No pre-computed ratios, Datama calculates them. Stacked or side by side, both shapes work.

Datama Compare is available where your data already lives: Excel, Google Sheets, Skill (Claude, ChatGPT) or your usual visualisation tools (Power BI, Tableau, Looker Studio, Qlik).

Frequently asked questions

What is the difference between a sales bridge and a revenue bridge? The sales bridge explains volumes (customers, distribution, frequency), without dealing with price. The revenue bridge values those volumes and adds the price and mix effects. The first serves the sales force, the second pricing and sales leadership.

Should you compare against budget or prior year? They answer different questions. Budget measures delivery against a commitment. Prior year measures business momentum. Finance teams often present both, budget first.

My budget and my actuals are in two separate columns. Do I need to reshape the data? No. Comparing by metric compares two columns directly, for example Revenue budget and Revenue actual, with no pivot beforehand.

What is the mix effect in a bridge? It is the share of the variance due to a change in the composition of sales, with unit prices and costs unchanged. It hides inside averages: average price, average cost, average margin rate. The price volume mix analysis guide covers it with a worked example.

Can you build a bridge in Excel? Yes. The Datama add-in for Excel runs the calculation directly in the workbook.

In summary

Variance analysis is a complete approach: measure the variance, explain it, decide. Measurement tells you how much. Bridges tell you why: the sales bridge for a volume variance, the revenue bridge for a revenue variance, the margin bridge for a profitability variance. They nest into one another, and you choose the one that matches the KPI your audience manages. When volumes, revenue and margin do not tell the same story, present them together: it is almost always the mix talking.

→ Learn more about the solution: datama.io