Five revenue numbers. One management decision.

context

The situation

Imagine a retail leadership meeting in which commercial, finance and growth teams arrive with different revenue totals. Each source uses a slightly different recognition window, customer definition or adjustment. The immediate problem looks like reporting inconsistency; the underlying problem is that executives cannot tell whether apparent growth improves customer economics.

The management question is specific: should scarce growth funding go to acquisition, repeat purchase or operational fixes? It cannot be answered from a top-line revenue chart alone.

diagnosis

The diagnosis

Five apparently routine KPIs hide different denominators and incentives. Orders versus customers, gross versus net revenue, revenue per order versus order frequency, repeat purchase rate versus retention, and the fully loaded cost of acquisition need separate treatment.

In this synthetic series, revenue rises 52.2% between endpoints while CAC increases 27.3%. That combination warrants inquiry, but does not prove acquisition became uneconomic. Contribution margins, customer mix, channel allocation and payback period are still missing.

intervention

The analytical intervention

Rather than start with a dashboard redesign, the proposed operating model starts with definitions and governance.

• Metric contract: freeze business meaning, denominator, exclusions, grain, refresh time, owner and reconciliation tests for each core KPI.
• Shared calculation layer: calculate revenue, orders, AOV, repeat rate and CAC once, with lineage back to source transactions and campaign cost.
• Management view: show growth and its counter-metrics together; headline revenue without acquisition economics or cohort behavior is incomplete.
• Operating cadence: route material deviations to named owners, with documented causes, next action and closure date.

findings

What the sample actually shows

The synthetic repeat-rate series increases from 52.7% to 65.2%, a 12.5-percentage-point movement. A higher share of repeat purchasers is promising, but it cannot be credited with causing revenue growth without cohort and basket analysis.

The data therefore supports a better question, not a proven turnaround. Are acquisition costs increasing because channels deteriorated, because higher-value cohorts are being acquired, or because expense attribution changed?

A shared number becomes valuable when it changes the meeting from “whose report is right?” to “which driver should we act on?”

decision

The management recommendation

Recommended next decision — not an executed outcome: retain the existing growth envelope temporarily, investigate channel-level CAC payback and repeat cohorts, and only then reallocate marketing spend. Prioritize reconciliation failures that materially alter the allocation decision.

Success criteria for a real rollout would include: 100% ownership of tier-1 metric definitions; zero unexplained reconciliation breaks above agreed thresholds; and evidence that management actions use the same reconciled contribution view. These are suggested acceptance criteria, not measured achievements.

limits

What this case does not prove

• No actual change in profit, campaign ROI, operating effort or decision speed was measured.
• Endpoint changes do not remove seasonality or customer-mix confounding.
• Correlation between repeat rate and revenue does not establish incremental contribution.
• The metric dictionary must be validated against real source systems before use.

sources

Evidence and further reading

ReadSigil: synthetic work sample · Original fictional KPI/MIS report (PDF)

For a documented real-world comparison, see AWS’s case on Together Credit Union, where access to branch data changed materially after a data-lake modernization. That is a separate customer-reported implementation, not evidence for Northwind.