The tell
A monthly pack of twenty charts, presented for forty minutes, followed by no decision. Everyone leaves informed and nothing changes — which means the reporting is a ritual, not a control.
The second tell: leadership asking questions the report cannot answer, and analysts building a new chart each time rather than fixing the underlying model.
Start from the decision, not the data
Write down the recurring decisions the leadership team actually makes. There are rarely more than eight. Then work backwards to the smallest number of measures that would change each decision.
| Recurring decision | Number that changes it | Cadence |
|---|---|---|
| Where to move budget | Cost per qualified opportunity by source, from the CRM | Monthly |
| Whether to hire in delivery | Capacity used against committed work, with trend | Monthly |
| Which process to fix next | Exception volume by boundary | Weekly |
| Which accounts need intervention | Health rating changes and breached SLAs | Weekly |
| Whether the forecast is credible | Stage conversion and pipeline age versus history | Weekly |
Report exceptions, not totals
A total describes a population; an exception names a case. Operating reviews run on exceptions: work past its stall limit, responses past their standard, milestones past their target, records without owners. The team can act on each line, which is what makes the meeting produce decisions.
State the limits inside the report
- Name the source system for each number. Two systems will disagree, and the disagreement should be documented rather than rediscovered every month.
- State the window and the model for anything attributed. A performance figure without a window invites an argument nobody can settle.
- Show the denominator. Rates without counts hide sample sizes small enough to be noise.
- Mark estimates as estimates. Credibility is cheaper to keep than to rebuild.
The four reports most businesses actually need
Before building anything new, check whether these exist and are trusted. In most companies two of the four are missing, and the missing ones explain the recurring surprises.
- The exception list
- Everything currently breaching a standard: response times, stage stalls, missed milestones, unowned records. Produced daily, reviewed weekly, and the only report that reliably produces action.
- The flow report
- Volume and conversion between stages, with age. Shows where the process loses work rather than how much work exists.
- The capacity report
- Committed work against available capacity, by team, with a trend. This is the report that turns a hiring argument into a decision.
- The commercial report
- Revenue and cost by source from the system of record, with the attribution boundary stated. Used for allocation, never for channel optimisation.
Anything beyond these four should be able to name the decision it supports. If it cannot, it is a chart somebody enjoyed building.
Put it in a cadence
Daily: exceptions only
An automated list of breaches. No meeting; a channel and an owner. Anything that needs discussion escalates.
Weekly: pipeline and delivery
Fixed agenda, live data, thirty minutes, ends with named decisions and owners. Status without a decision is homework, not a meeting.
Monthly: capacity and allocation
Where money and people are going, against what the operating model can absorb. This is where the budget decisions belong.
Quarterly: structure
Do the operating model, definitions and standards still fit the business? If nothing has changed in a year of quarterlies, they are not being reviewed honestly.
Related work
Reporting integrity depends on the record layer — see the CRM nobody trusts. The measurement boundaries for marketing numbers are covered in marketing operations, and the term itself in operating cadence.
