Management Reporting That Supports Decisions: Designing the Information Layer for CFOs

Statutory and management reporting answer different questions

Statutory reporting organises accountable financial information around accounting standards, legal entities, prescribed classifications, external disclosure and applicable regulatory or audit requirements. Controlled accounting policies establish the basis for recognised historical position and performance. That discipline remains essential; an internal presentation cannot substitute for it.

Management reporting organises information around internal decisions: business models, products, customers, channels, operational processes and responsibility centres. Plans, forecasts and forward implications sit beside historical results. The organising question becomes what management can protect, improve or change, within a defined horizon.

A commercial repricing view may need contribution margin by customer. A capacity decision may need bottleneck contribution and avoidable cost. Functional accountability may need controllable costs rather than allocated corporate overhead. Normalised performance, alternative segments and scenarios require explicit bridges to accounting results, approved definitions and retained adjustments. Neither a different cadence nor a different materiality permits an unexplained reconciliation difference.

Specify the decision before selecting the information

For each section, write a short decision contract: owner, question, horizon, financial outcome, operational drivers, materiality, uncertainty tolerance, required evidence, available actions and review cadence. “Explain performance” is too broad. “Approve targeted repricing for the next quarter without losing profitable constrained capacity” establishes a useful boundary.

The same validated model can serve several audiences, but not through one universal pack. The board needs capital and risk implications; the operating team needs intervention detail; the controller needs accounting and evidence exceptions. Shared definitions prevent contradictory truth without forcing identical information density.

Remove a measure from the executive layer when its movement cannot change a decision, trigger investigation or test an important assumption. Retain it in diagnostic evidence if necessary. A report cannot be optimised until the decision it supports is explicit.

Connect the objective to an accountable response

The five layers are objective, financial result, operational drivers, exception and interpretation, and decision and action. They are relationships, not five mandatory report pages. One finding may connect a margin objective to realised price, a specific discount exception and a commercial approval.

Start from a model whose source population, mappings, adjustments, period definitions and reconciliations support the intended use. FIR-03 financial data validation establishes that boundary. Then ask, in order: what happened, why, why does it matter, what must be decided, and who must act? Stopping at the first question leaves a descriptive report.

A KPI hierarchy makes compression reversible

Govern each KPI through a definition, formula, numerator and denominator, unit, source, owner, frequency, target or comparison, materiality logic, drill-down path, permitted interpretation and limitations. Record effective versions. Changing the definition while retaining the label creates a false trend.

For example, DSO requires a declared receivables population, credit-sales denominator, averaging method and day convention. The approved interpretation is collection intensity within that scope; it does not independently prove overdue debt or collectability. Its evidence path must reach ageing, invoices and subsequent receipts. The owner of the definition may be Finance while collections owns the response.

A lower-level metric earns executive prominence temporarily when it threatens an enterprise outcome. Yield may enter the decision page during a material production problem and return to operational reporting after validated recovery. Retirement criteria prevent yesterday’s exception becoming tomorrow’s permanent KPI.

Each comparative view has a different analytical job

Actual versus budget tests delivery against plan. Actual versus the forecast frozen before the period tests forecast error or new information. Actual versus prior year supplies historical context. Forecast versus budget tests whether the approved objective remains attainable; run rate versus forecast exposes the execution required to get there. A forecast updated after actuals arrive cannot fairly measure forecasting accuracy for that interval.

Align period duration, entity scope, currency, units, classification, accounting treatment, mapping, sign, seasonality and version before comparison. Clearly separate year-to-date flows, full-year expectations and closing positions. Use percentage points for changes in margins or ratios. FIR-06 comparative analysis explains denominator and structural effects.

Do not place all six comparisons beside every line. Preserve them in the model and choose the relevant view. Excess columns encourage readers to select whichever baseline makes their preferred narrative look strongest.

Financial relationships explain contribution before causality

Separate observed movement, deterministic contribution, supported explanation, unresolved hypothesis and decision consequence. A reconciled price-volume-mix bridge establishes how revenue changes under its decomposition convention. It does not establish why customers accepted discounts or whether the volume is sustainable. Operational evidence must support those explanations.

Higher production can reduce absorbed unit cost while building unwanted stock. Positive revenue mix can lower contribution margin when growth favours higher-priced but less profitable products. Neither correlation nor an unexplained residual proves operational causality. FIR-07 variance analysis supplies the controlled driver bridge; management reporting determines which resulting finding needs attention.

Materiality governs attention, not reconciliation

An exception is a departure requiring attention, decision, investigation, correction, escalation or monitoring. A difference from budget alone is insufficient. Rank absolute financial impact and percentage or margin effect alongside persistence, trend, volatility, controllability, decision sensitivity, liquidity or covenant relevance, concentration, strategic consequence and evidence quality.

Financial materiality concerns the amount; operational materiality concerns a critical process; strategic materiality concerns direction or a key relationship; risk materiality concerns exposure and constraint; information-quality materiality concerns whether management can support the conclusion. A small unexplained movement can be material if it invalidates a financing assumption.

Approve thresholds for the decision and reassess them as headroom changes. There is no universal percentage. Group similar exceptions before ranking them, and prevent favourable and adverse amounts from hiding each other through netting. Uncertainty increases the need for investigation; it does not automatically make every weak signal an executive emergency.

A finding ends with ownership, not concern

Mark fact, calculation, inference and hypothesis explicitly. “Revenue grew” is an observation; the price contribution is a calculation; temporary customer stocking is an explanation requiring order evidence. “Costs should be monitored” supplies none of the decision contract. Draft commentary from validated findings and bind it to the numerical version; changed dependencies require renewed review.

Separate result owner, driver owner, response owner and decision approver. Commodity prices may be external; Procurement owns sourcing, Commercial repricing, Operations consumption, and the CFO the forecast consequence. Give cross-functional actions one accountable coordinator rather than a shared label that leaves nobody responsible.

The action register retains finding, decision, owner, due date, expected financial effect, operating dependency, risk, status, completion evidence and follow-up result. Closing a task requires proof of the response; closing the exception also requires testing the outcome. An approved price list is not proof of improved realised margin.

Keep the detail inspectable without keeping it on page one

The evidence path is executive conclusion → finding → KPI or bridge → validated canonical value → mapping and adjustment → extracted source value → source location. Each step retains the relevant segment, calculation logic, source and mapping provenance, adjustment history, validation status and unresolved limitations.

A dashboard is a presentation surface, not the evidence source. Preserve the source snapshot and model version behind it. Drill-down may be a linked schedule or controlled evidence package; it does not require transaction tables in the executive pack. Access restrictions must follow the underlying information, so an executive summary does not accidentally disclose customer or employee detail.

When challenged, the analyst must reproduce both the number and its interpretation boundary. Executive compression is trustworthy only when the underlying detail remains inspectable. Removing detail without a return path creates brevity at the expense of control.

Report when information can still change action

Available liquidity and overdue receivables may need daily or weekly review; revenue and order intake weekly or monthly; the full management P&L monthly; product profitability monthly or quarterly; strategic capital allocation quarterly; and the long-term plan annually with event-driven revision. These are candidate rhythms, not universal rules.

Real-time or exception-triggered reporting is useful only when sources are stable enough, the decision can occur at that frequency, an owner can act, noise does not dominate and the evidence chain remains controlled. Distinguish provisional operational signals from closed financial outcomes. A daily shipment signal can trigger investigation without pretending to be recognised monthly revenue.

Every report should identify its cut-off, next review and early escalation condition. FIR-10 month-end workflow governs the recurring release. Between releases, refresh only the decision-sensitive dependencies and disclose their different dates; a new cash balance does not silently refresh an older profitability finding.

Design six connected pack layers, not six executive summaries

Page one shows the smallest complete view of the decision, including adverse and favourable evidence. Supporting layers hold explanations, alternatives and challenge material. The executive page links to them rather than reproducing them. The six layers are an information architecture, not a mandatory page count or software interface.

Growth that consumes margin and funding

Consider a wholly fictional mid-sized manufacturer and distributor. Amounts are EUR millions unless stated. Flow comparisons cover January–June 2026; positions are at 30 June. The latest forecast, frozen on 8 July, covers the full year with positions at 31 December. It is not a like-for-like H1 comparator. All historical source and model controls pass.

DSO and DIO use average trade balances, credit revenue and cost of sales over 181 H1 days or 365 forecast days; all sales are on credit. They describe period intensity, not endpoint cash movements. For example, actual average receivables are 31.8 × 61 / 181 = 10.717m and average inventory is 23.0232 × 88 / 181 = 11.193m, rounded. Closing movements below come from separate reconciled schedules.

The sequential volume–mix–price convention gives volume +1.5m, mix +1.0m and Core price −0.7m: budget revenue 30.0m becomes 31.8m. Volume explains 83.3% of the net upside. Specialist has a higher selling price but a lower margin rate. Its greater weight therefore increases revenue while diluting gross margin.

Gross profit is 8.7768m / 31.8m = 27.6%, down 2.4 percentage points. The positive mix contribution to absolute gross profit must not be called a favourable margin mix: at unchanged prices and costs, margin falls from 30.0% to 9.65m / 32.5m = 29.692%. Operating costs exceed budget 5.4m by 0.2768m, completing the 0.5m EBITDA shortfall.

Invoice and discount records support the Core price movement; purchasing records support the material-price effect. The usage calculation reconciles, but production-order evidence does not yet distinguish physical inefficiency from an outdated consumption standard. The expense remains recognised; its cause and forecast persistence remain qualified. Meanwhile, a 0.15m maintenance underspend reflects delayed work scheduled for H2, not a sustainable efficiency gain.

There are no other non-cash or operating adjustments in this simplified example. Trade working capital absorbs 1.9m against 0.3m planned; the additional 1.6m plus the 0.5m EBITDA miss explains the 2.1m cash shortfall. FIR-08 working capital guides release actions, while FIR-09 cash reconstruction separates flows from balance changes.

Opening net debt is 7.0m. Actual capex 3.9m less operating cash 0.7m increases it to 10.2m; budget capex 3.8m less cash 2.8m gives 8.0m. Opening debt/cash of 9.8m/2.8m becomes 12.0m/1.8m after 2.2m net borrowing. No dividends, currency effects or restricted cash are assumed; real reporting must test these explicitly.

The FY forecast requires H2 revenue 31.2m, EBITDA 2.9m and operating cash 2.2m. Full-year capex 6.4m less operating cash 2.9m gives closing net debt 10.5m. Forecast debt 12.0m and unrestricted cash 1.5m, against a committed available facility of 14.0m, leave available liquidity 3.5m. The fictional board minimum is 3.0m: cushion is only 0.5m, versus 2.0m under the original 5.0m liquidity plan.

A sensitivity delaying 0.8m of collections beyond year-end reduces liquidity to 2.7m. Deferring 0.6m of non-essential H2 capex would restore 3.3m if timing and operational feasibility are approved. The base case already assumes collection improvement and maintenance catch-up; neither can be counted again as incremental upside. A period-by-period cash schedule must also test the interim trough, not just December.

Five findings replace ninety-four executive indicators

Revenue exceeded budget by EUR 1.8 million, but the increase did not convert into planned profitability. Volume contributed most of the upside, while adverse Core pricing and a lower-margin sales mix weakened gross margin; material price and usage variances completed the 2.4-point decline. EBITDA finished EUR 0.5 million below budget. Commercial owns the targeted price response; Operations must validate the remaining usage explanation before its persistence enters the forecast.

The product inputs, individual invoices and all operating KPIs remain available below the executive layer. Page one keeps only the measures needed to approve commercial action, investigate uncertainty, release cash and preserve liquidity. The action register tests completion evidence and realised financial effects at the specified reviews.

A pack can support one decision and block another

The example is decision-ready for targeted repricing within evidenced customer and product scope. Historical usage expense is valid, but its causal explanation is review-required. The liquidity base case is ready with limitations for protective contingency planning; final forecast approval awaits the weekly trough and collection-timing review. These states do not contradict one another.

Record authorised uses, excluded conclusions, accepted uncertainty and the evidence that would reopen each finding. A visually complete pack does not receive a universal green status simply because its totals reconcile.

Professional appearance can conceal decision failure

The information layer earns its place through the decision

For Entimema Financial Intelligence, the implementation sequence is a validated financial model → comparable actual, budget, forecast and prior-period views → deterministic KPI and driver calculations → materiality and exception logic → evidence-linked findings → executive commentary → decision and action layer → traceable deliverable. This is an architecture to scope and validate, not a claim that every integration or approval capability is already deployed.

Model intelligence may interpret semantics, explain relationships, prioritise findings, compare hypotheses, request clarification, draft commentary and adapt the presentation to the decision. Deterministic code owns arithmetic, KPIs, ratios, bridges, variances, comparisons, control totals, fixed materiality rules, reconciliation and drill-down relationships. Human judgement owns KPI selection, materiality approval, material uncertainty, incomplete causal conclusions, action ownership and the final decision.

The FIR-05 end-to-end workflow establishes the wider evidence path. Here, its final deliverable is neither a generic dashboard nor a conversational answer. It is a controlled relationship between a financial finding and an authorised management response.

The next meeting need not begin with fewer available facts. It should begin with fewer unanswered decision questions. Management reporting succeeds when it changes the quality and speed of a decision, not when it maximises the quantity of information presented.