Where Are We Making and Losing Money — and Why?
Most mid-market organisations can tell you what happened. Revenue was up 8%. Margin dropped two points. Costs ran over.
The question is whether anyone understands why — and whether that understanding arrives in time to change the outcome.
When performance drivers are unclear, the management meeting becomes a debate. Each function has an explanation; none has the decomposition to prove it. By the time the analysis is complete, the corrective-action window has typically closed. The same variance recurs next quarter.
What Good Performance Analysis Produces
- Driver clarity: Management knows which specific factors create or destroy value — not just that results changed.
- Explanation speed: Deviations are understood within the reporting cycle, not weeks after the corrective-action window has closed.
- Decision connection: Analysis reaches a named decision and a named owner — not a slide deck that circulates without response.
Key Business Questions
- Where are we making and losing money? Aggregate profitability masks segment variation. If the answer changes depending on which spreadsheet you look at, performance analysis has failed.
- Which factors drive our results? Volume, price, mix, cost structure — without decomposition, management debates opinions rather than evidence.
- Why did results deviate from plan? A variance without a root cause is not an explanation. It is a restatement of the problem.
- Which levers can management actually pull? Analysis that attributes performance to external factors and market conditions produces no corrective action.
- Is current performance sustainable? One strong quarter can mask structural margin erosion. Trend and driver analysis make the difference visible before it becomes irreversible.
The Performance Analysis Operating Model
Effective performance analysis is not a set of reports. It is an operating model — a repeatable sequence that turns outcomes into understood drivers and drivers into decisions. Five components define the model.
1) Profitability decomposition
Results disaggregated by product, customer, channel, and geography. Not total revenue — contribution by segment. Not total cost — cost behaviour by driver. Most companies discover that a small number of customers or products generate the majority of margin, and that the rest absorbs most of the management attention. Visibility into this structure is the precondition for any resource allocation decision.
2) Cost structure understanding
Fixed vs variable, direct vs indirect, driver-linked vs allocated. Without a documented cost structure model, any variance analysis produces contested conclusions — every department disputes the methodology before engaging with the finding. The analytical logic must be agreed, documented, and applied consistently.
3) Variance analysis
Plan vs actual, with decomposition by price effect, volume effect, and mix effect. A 5% revenue shortfall driven by volume is a sales capacity problem. The same shortfall driven by mix is a portfolio management problem. The same shortfall driven by price is a competitive positioning problem. Each requires a different response. Analysis that stops at “revenue missed plan” informs nothing.
4) Driver identification
Which factors does management control, and at what leverage? Revenue drivers, cost drivers, capacity drivers — the goal is the short list of high-leverage variables that, when managed, shift performance materially. Analysis that surfaces forty drivers produces the same paralysis as no analysis. Fewer drivers with owners beat more drivers without them.
5) Insight-to-decision loop
Analysis only creates value when it reaches a decision. The loop is mechanical: identified driver → named owner → agreed action → due date → expected impact tracked next cycle. Without this structure — where each driver has an owner who commits to a response — performance management becomes reporting on performance rather than managing it.
Driver Attribution Layer
The practical test of performance analysis is whether it can answer: “Margin dropped two points — why, and who acts?”
A structured decomposition works through the waterfall:
- Revenue: Did total revenue change? Decompose into volume and price
- Volume effect: Did we sell more or less? Decompose by product, customer, channel
- Price effect: Did realised price change? Separate mix shift from rate change
- Mix effect: Did the composition of revenue shift toward lower-margin segments?
- Cost structure: Did input costs change? Separate fixed cost absorption from variable cost per unit
- Margin: Net effect — attributed, owned, actionable
Minimum Viable Decomposition — four analyses cover most performance questions:
- Price / volume / mix: what drove the revenue change
- Margin bridge: what drove the profit change (revenue → gross margin → EBITDA)
- Opex drivers: what drove the cost change (fixed load, variable rate, structural shifts)
- Working capital drivers: what drove the cash change (debtor days, inventory turns, creditor terms)
The analysis cycle:
- Day 1 (flag): Exception view surfaces deviations above threshold — signal only, no narrative yet
- Day 2–3 (decompose): Finance completes attribution by driver, names an owner per finding
- Day 3–4 (commentary): Owner confirms attribution, adds operational context
- Day 4–5 (decide): Management reviews; actions logged with owner, due date, and expected impact
- Next cycle (close loop): Prior actions reviewed; impact confirmed or escalated
Each node has an owner. Each explanation is traceable to a specific operational or commercial decision. When this structure exists, “margin dropped two points” becomes a specific, attributed finding — not a debate.
Performance ownership at a glance:
- Definition owner: governs KPI definitions and computation paths (inherits from Governance)
- Data owner: ensures reconciled actuals arrive on time (inherits from Reporting)
- Finance (analyse / attribute): decomposes variance by driver, assigns owner per finding
- Business (confirm / act): confirms attribution, commits to action with due date and expected impact
- Decision owner: management reviews; actions logged and tracked next cycle
Performance Health: Quality Metrics
Performance analysis is only as useful as its speed and connection to decisions. Five metrics indicate whether the operating model is working.
- Explanation coverage: Percentage of material variances with a documented root cause — target: 100% of variances above reporting threshold
- Decision cycle time: Days from deviation identified to management decision taken — above 10 days indicates the analysis-to-decision loop is broken
- Driver identification rate: Percentage of key variances linked to a specific operational driver with a named owner
- Insight-to-action rate: Percentage of analytical outputs that result in a tracked management action
- Proactive flag rate: Percentage of material deviations surfaced before period end — signals whether analysis informs the period or only explains it
No new system is needed to measure these. The variance log, decision register, and exception history contain everything required.
Together, they prove that the reconciled baseline Reporting delivers is being converted into attributed decisions — not circulated as commentary.
Performance Areas
Profitability and Margins
Profitability analysis explains where value is created and where it erodes — by product, customer, channel, or geography. The question is rarely “are we profitable?” It is “which parts of the business are subsidising which other parts, and does management know?”
→ Profitability Analysis Framework · Customer Profitability · Margin Decomposition
Cost Structure and Drivers
Cost analysis explains how costs behave relative to activity — what is fixed, what is variable, what is driven by decisions and what is structural. Without this, cost reduction efforts cut headcount rather than drivers. The underlying cost base often returns within two to three years.
→ Cost Structure Analysis · Cost Driver Identification · Fixed vs Variable Cost Framework
Variance Analysis
Variance analysis decomposes deviations into attributed causes — price, volume, mix, cost — each linked to an owner. The goal is not to explain the past for its own sake, but to identify which levers were managed, which drifted, and which require a different decision next cycle.
→ Variance Analysis Framework · Price Volume Mix Analysis · Root Cause Analysis for Finance
Performance Reporting and Insight
Analysis only creates value when it reaches the right person at the right time. Performance reporting connects driver-based insight to management review cycles — integrated into the standard reporting rhythm, not delivered as a separate analytical exercise that arrives after the decisions are made.
→ Performance Dashboards · Management Commentary · Exception-Based Reporting
Inputs, Controls, Outputs, Decisions
- Inputs: Reconciled actuals from Reporting, governed KPI definitions from Governance, variance attribution from prior cycles
- Controls: Materiality threshold (variances above threshold require decomposition and owner assignment), analysis cycle cadence, insight-to-decision loop
- Outputs: Driver-attributed variance analysis, insight-to-action log with named owners, due dates, and expected impact
- Decisions enabled: Corrective action per driver, resource reallocation, assumption revision for Planning — each logged with owner and close date
What Performance Analysis Is Not
Performance analysis is overloaded. Boundaries matter.
- What happened, on what cadence? — that is Reporting .
- What will happen if conditions change? — that is Planning & Projections .
- Can we trust the underlying data? — that is Governance & Data Trust .
Performance analysis answers one question: why did it happen, and which levers can management pull?
How Performance Connects
Performance analysis depends on reliable reporting. When actuals are inconsistent, delayed, or disputed, driver analysis produces contested explanations — the same debate, with better-formatted slides.
Without trusted performance insight, planning disconnects from operational reality. Plans built on assumed drivers rather than observed drivers produce forecasts management cannot defend when conditions change.
Strong performance analysis is the insight layer of one control system. It converts the baseline that Reporting produces into the driver knowledge that Planning requires. Driver identification feeds Planning directly — each named driver with an owner becomes an observed assumption input for the next planning cycle, not an inherited one.
→ Why Reporting Matters for Mid-Market Companies — the foundation this discipline depends on
Typical Situations
- Variances are reported each month but never explained by driver, so the same issues recur in the next cycle without correction
- Profitability analysis shows “the business is profitable,” so management is surprised when specific products or customers are loss-making
- Cost reduction targets are set at the aggregate level, so cuts are made by headcount rather than by driver — and the costs return
- The management meeting has the numbers but not the decomposition, so each function defends its own narrative — the decision is deferred, and the same variance recurs next quarter
- Growth improves fixed cost absorption and overall margin, masking structural erosion in the underlying product or customer mix
Next Steps
- Explore performance topics in depth — Knowledge Hub
- See how organisations apply performance analysis — Use Cases
- Discuss your situation — Contact