Variance analysis in management reporting: budget vs. actuals, thresholds and presentation (2026)

Variance analysis in management reporting: budget vs. actuals, thresholds and presentation

The method that makes deviations explainable: comparison axes, the price/volume/mix split, thresholds and the bridge presentation.

What is variance analysis in management reporting?

Variance analysis is the systematic comparison of the actual figures (actuals) against a reference — the budget, last year or a recent estimate — and the explanation of the differences that matter. It is the step that turns a management report from a book of tables into steering information.

‍

The essence sits in the word “explanation”. That revenue is 8% below budget, every reader sees in the table; the analysis must tell where it happened (which product group, which customer, which entity), why (less volume, lower prices, a shifted mix, a timing effect) and what it means for the rest of the year. Only then can management take a decision — and can investors follow the story without having to dig themselves.

‍

Variance analysis is therefore not a separate exercise next to the report, but a fixed section inside it: in the main article on the good management report it is section five of the seven fixed sections, between the cash flow and the KPIs. Its outcomes also feed the commentary: the three to five messages of the month almost always come out of the variance analysis.

‍

Good variance analysis has three properties. It is selective: only deviations above an agreed threshold get explained, the rest is noise. It is causal: every difference is split by what caused it, not merely named. And it is consistent: the same axes, the same thresholds and the same presentation every month, so the reader learns to see patterns instead of starting over each time. The rest of this article works out those three properties.

What do you compare actuals against: budget, last year or the latest estimate?

Against all three — each comparison axis answers its own question, and it is precisely the combination that completes the picture.

‍

Against the budget: are we doing what we set out to do? This is the accountability axis — the budget is the agreement with management and investors, and the deviation against it is the starting point of every monthly conversation. The limitation: a budget is made once a year and ages. Whoever still compares October against a budget from last November partly measures the quality of last year’s planning instead of today’s performance.

‍

Against last year: how are we developing? This axis is insensitive to the quality of the budget and shows the real movement — growth, margin development, cost trends. With seasonal patterns, the comparison against the same month last year is often more informative than against the previous month. The limitation: last year may itself have been a bad year; “better than last year” is not yet “good”.

‍

Against the latest estimate: are we doing what we recently still expected? Whoever maintains an updated full-year expectation next to the budget can measure whether the latest insights are coming true. Revenue that sits 10% below budget but exactly on the latest estimate is a different conversation than revenue that also misses the recent expectation — the first is old news, the second a new problem. How to set up such an estimate is covered in forecasting for SME CFOs.

‍

In practice: put the axes as fixed columns next to the actuals — month and year-to-date — and explain the deviations against the budget as the main line, with last year and the latest estimate as context. One main axis prevents every deviation from having to be explained three times.

How do you split a deviation into price, volume and mix?

The price/volume/mix split decomposes a revenue or margin deviation into three causes: did we sell more or less (volume), at different prices (price), and did the composition of what we sold shift (mix)? Each cause has its own owner and its own action — volume is a conversation with sales, price with pricing, mix with both — and that makes the split more valuable than the total amount.

‍

A worked example. The budget assumes 100 units at EUR 50: revenue of EUR 5,000. Actually sold: 110 units at EUR 47, revenue of EUR 5,170. The total deviation is +EUR 170 — on the face of it a fine month. The split tells a different story: the volume effect is (110 − 100) × EUR 50 = +EUR 500, the price effect is (EUR 47 − EUR 50) × 110 = −EUR 330. The business is growing in units but giving a sizable part back through price — and that calls for a pricing conversation, not congratulations.

‍

The mix effect comes in as soon as there are multiple products, services or customer groups: if volume shifts toward groups with a lower margin, the total falls without price or volume changing within any group. Calculate the mix effect by valuing the actual volumes at the budget margin per group and comparing that against the budget mix — the difference is what the shift cost or delivered.

‍

The same logic works beyond revenue: splitting personnel costs into FTEs times rate, purchasing costs into volume times purchase price. How this plays out per business model is covered in the deep-dive articles — for example the margin split at trade and wholesale and the margin bridge at manufacturing companies.

Which thresholds do you use: when is a deviation worth explaining?

Use a combined threshold: a deviation only gets explained when it exceeds both an absolute amount and a percentage — for example at least EUR 10,000 and at least 5% of the line. The right level depends on the size of the business; the principle is that both conditions must hold at the same time.

‍

The reason for the combination: a percentage alone lets small lines dominate the conversation (a doubling of office costs from EUR 800 is a 100% deviation and zero relevance), and an amount alone makes large lines raise the alarm too early (EUR 15,000 on revenue of EUR 2 million is three quarters of a percent — normal movement). The combined threshold filters out both kinds of noise.

‍

Three refinements make the threshold workable in practice. First: differentiate by level — a lower threshold for margin than for individual cost lines, because margin deviations are more structural in nature. Second: also look cumulatively. A deviation that stays just under the threshold every month but points the same way for seven months is a trend and deserves an explanation after all. Third: fix the thresholds once, together with management, and keep them constant through the year — whoever moves the bar every month makes the analysis incomparable.

‍

The effect of good thresholds shows mainly in what is not there. A variance analysis that explains twenty lines buries the three that matter; an analysis with thresholds delivers a short list of material deviations — and the calm to investigate those properly. The lines below the threshold stay visible in the tables; they just do not get a story.

How do you present the variance analysis?

The strongest presentation form is the bridge: start at the budget, show a plus or minus per explaining item, and end at the actual. The reader sees in one image which causes made the difference and how large each effect is relative to the others — something a table with two columns and a difference column never shows.

‍

The bridge works at every level: from budget EBITDA to actual EBITDA for the group picture, from budget revenue to actual revenue with the price/volume/mix effects as steps, or from last year to this year for the development question. Choose one main bridge per report — usually budget to actual at result level — and use the others as depth in the appendices.

‍

Around the bridge, four presentation rules apply. Label every deviation as timing or structural: an order that slipped a month heals itself; a lost customer does not — and the distinction determines whether action is needed. Show month and year-to-date: monthly figures are volatile, the cumulative shows whether the year is still on course. Keep favorable and unfavorable symmetrical: a windfall deserves the same explanation as a setback, because it too says something about the predictability of the business. And tie every material deviation to the commentary with the fixed triptych: what happened, why, and what are we doing about it.

‍

The place in the document follows the main article: the variance analysis sits after the cash flow and before the KPIs, with the bridge on the first page of the section and the detail tables behind it. In the summary up front, only the conclusion returns: the three deviations that define the picture of the month.

How do you work budget and forecast into the management report?

The fixed column layout is: actual, budget, last year and — where it exists — the latest estimate, for both the month and year-to-date. That layout returns on every page, from P&L to KPI section, so every reader sees every figure in the same context.

‍

The budget stays fixed: it is set once and not adjusted afterwards, even when reality walks away from it. A budget that moves along loses its function as a yardstick — the deviation against it is precisely the information. What does move is the latest estimate: a periodically updated expectation for the current year, built from the realized months plus a refreshed projection for the remaining months. The budget stays the agreement and the estimate becomes the compass — two roles that complement each other as long as they do not blur.

‍

The most important addition for management and investors is the year-end projection: where do we land if the rest of the year runs according to the estimate? That single line connects the monthly deviations to the annual agreement and makes timely steering possible — phasing costs, adjusting financing or resetting expectations toward stakeholders before year-end forces it.

‍

How often you refresh the estimate depends on the dynamics of the business: quarterly is enough for most SMEs, monthly under strong swings. The method behind it — rolling forecasts, scenarios, the liquidity translation — falls outside this article and is covered in forecasting for SME CFOs; for the report, what counts is that a current expectation sits next to the fixed budget.

How do you automate the variance analysis?

The labor-intensive part of variance analysis is not the analyzing but the staging: collecting actuals from the accounting system, reconciling them with the budget model and breaking them down to the levels the analysis runs on. Exactly that part can be automated.

‍

Finstack connects accounting packages such as Exact Online, AFAS and Twinfield — plus, among others, Odoo, Xero, QuickBooks and MS Dynamics 365 BC — directly via the API, read-only and at transaction level. The actuals refresh automatically every day, with a manual refresh at any moment. With multiple entities, the consolidation runs along, including automatic intercompany elimination — so the variance analysis is right at group level and per entity, as worked out in management reporting for a holding company with multiple entities.

‍

The budget and the latest estimate live where they usually already live: in the team’s own Excel or Google Sheets model. Through the 2-way integration, the actuals refresh in that existing model in one click, right next to the budget columns — the monthly copy-paste work disappears, the model’s setup stays. In the Finstack dashboards, the actuals are moreover clickable through to the source booking: the question “where does that deviation come from” becomes a mouse click instead of a search through the ledger. You can also share dashboards and reports with stakeholders such as management, investors and the accountant, by giving them access to Finstack, with access rights set per user.

‍

The effect: the numbers side is ready on day one and the saved days shift to the part that requires human judgment — interpreting causes, formulating actions. AI can deliver a first draft of that explanation; a dedicated article on variance analysis with AI follows within the AI cluster. Finstack starts from EUR 39 per month for the first entity: live in 5 minutes, set up within a day.

The 3 most common mistakes in variance analyses

Three patterns we see again and again in variance analyses. Each makes the analysis weaker than the figures beneath it; each is preventable with the right method.

Sources and provenance

• Finstack — Reporting & insights (dashboards, Excel/Sheets sync): finstack.io/solutions/reporting-insights
• Finstack — Integrations with Exact, AFAS, Twinfield and other packages: finstack.io/solutions/integrations
• Finstack Help Center — Sources & connections (sync frequency, security): help.finstack.io
• Finstack — Pricing (from EUR 39/month): finstack.io/pricing

Last reviewed: 25 July 2026 · Next review: October 2026