Name the outcome.
Start with the commercial or operational requirement. State why it matters and which role can influence it. Avoid adding a measure solely because the system can collect it; collection is a cost as well as a capability.
Capabilities / Performance-management systems
Every measure
needs an owner.
Connect strategic outcomes to line-level definitions, accountable owners, review cadences and controlled responses.
Operate the decision lens
Reading now 1 of 6
Production
Output ÷ planned time. Owner: line supervisor.
Find the decisionProcedural visualisation. Schematic, not a facility.
The decision lens
Follow a strategic outcome down to an action.
The programme’s performance design organises measures across production, quality, delivery, cost, people and sustainability. The point of the cascade is to connect these families to controllable line-level decisions.
Select a family to inspect an example chain. These are conceptual measurement contracts, not live factory readings or approved targets.
Strategic outcome→Production→Output ÷ planned time→Line supervisor→Investigate defined loss
Output needs a declared denominator.
Connect accepted output to planned time and the actual product basis. A supervisor needs to know whether a gap reflects lost time, rate, quality or an unrealistic plan.
- Owner
- line supervisor
- Review
- shift
- Response
- investigate the defined loss
Illustrative cascade. Node labels are example measurement contracts, not live readings. Illustrative owner roles and review routines. Actual accountability and cadence must be agreed with the operating team.
Keep the source boundary
A reported percentage is useful only with its definition.
The programme annex reports CICAM capacity utilisation from 10.83 to 48.9 percent across the sites in a July 2023 report. That is capacity utilisation, not the availability, performance and quality factors of an OEE calculation.
CICAM 2 · July 2023 report (R12)
10.83% Capacity utilisation
3 OEE factors (availability, performance, quality) not reported in annex10.83CICAM 233.7Garoua48.9CICAM 1
Definition first, hung from zero
Field 1 of 4 · Contract sequenceName the event being counted and the unit it is counted in, before any percentage is shown.
Reported metric vs missing OEE factors
1 of 4 quantities in R12Capacity utilisation10.83%
Availabilitynot reported
Performancenot reported
Qualitynot reported
The quantity the July 2023 report publishes for the three sites.
CICAM 2 reported 10.83 percent capacity utilisation in the July 2023 report. The same report puts the cost of underutilisation at about CFAF 2 billion in 2020.
Historical utilisation context is R12. The displayed chain is a GFT method illustration, not a measured operating result.
The same discipline belongs inside a factory: define the event, unit, time window and exclusions first, keep missing data visible, and keep the reason when a correction changes an issued result.
Step 1 of 3 · Name the outcome
The design sequence
Build the cascade around the decision.
A performance system is an operating routine supported by records. The tree below changes as the argument moves from strategic outcome to definition and then action.
Each layer has a distinct role; adding a dashboard does not remove the need for the layers beneath it.
Define the observation.
Specify the unit, denominator, capture point and validation. Record what happens to missing, delayed and corrected events. Test whether different people would interpret the same event in the same way.
Close the response loop.
Give an exception an owner, a review point and a response. Check whether the intervention changes the original outcome without moving loss elsewhere. Retain the failed action as evidence for the next review.
Warp and Signal
A dashboard cannot take responsibility.
The measure ends in a decision, or it ends too early.
A figure becomes useful when its meaning is shared and its exception leads somewhere.
Why this protects the people doing the work
This also protects the people doing the work. A poorly defined target can reward incomplete output, conceal quality loss or move an interruption outside the reporting window. Review incentives and unintended effects before attaching judgement to a metric. A measure should help explain the process and support improvement, with context retained.
Illustrative register row, Production family
- Purpose
- Explain a gap in output
- Unit
- Accepted units
- Numerator
- Accepted output
- Denominator
- Planned time
- Source
- Count at the capture point
- Exclusions
- Declared, with the reason
- Owner
- Line supervisor
- Target status
- Proposed, not approved
Every change to a definition is kept with its date and reason, so an improvement cannot be created by silently revising it.
PurposeExplain a gap in output
A reviewable scope
Deliver the definitions and the routine together.
The work should leave enough structure for a team to operate and revise the system. Keep the measure register close to the capture form and dashboard.
If definitions live in a document nobody can find during a shift, disagreements will be settled by memory rather than evidence.
Deliverable 01 · Definition contract
Measure register
For each measure, record its purpose, unit, numerator, denominator, source, exclusions, owner and target status. Preserve changes so an apparent improvement cannot be created by silently revising the definition.
Register contract fields8 of 8 specified
Inspecting field: Purpose: Explain a gap in output
Deliverable 02 · Boundary validation
Capture and reconciliation design
Describe the source event and the checks that make it usable. Include duplicate handling, late entry, missing records and corrections. Reconcile totals at the boundary where the team actually makes a decision.
Capture check readiness (illustrative)2 of 4 rules defined
Toggle any rule in the left register to update the boundary count.
Deliverable 03 · Operating routine
Management and improvement cadence
Connect shift review, escalation and broader management review. Record the action, accountable role and result. The improvement record should show whether the intervention worked under comparable conditions, including its side effects.
Connected review cadence3 of 3 tiers linked
Active record: Management review: Whether the action worked under comparable conditions
Conditions that change the answer
Do not let the summary hide the uncertainty.
A polished chart can make a weak observation look authoritative. Keep the uncertainty in the operating view, where it can affect the decision.
Why the manual routine comes first
The design should support a reliable manual routine before more advanced analysis depends on its history. That sequence creates evidence for models instead of expecting models to repair undefined records.
Illustrative record line, ten days, no data
Is missing being treated as zero?
An absent stop record is not proof of uninterrupted running. Use an explicit missing state and a reconciliation route. Give the owner a way to resolve uncertainty without overwriting the original event.
Are the shifts comparable?
Check product mix, planned time, rework treatment and acceptance rules. Keep different boundaries visible. A comparison that changes the denominator without saying so should not support a performance judgement.
Does the action improve the whole process?
Look beyond the local indicator. A rate improvement that causes defects, excess stock or unsafe work has not established a better operating result. Review the connected measures and retain the trade-off.
Missing days stay visible as gaps. The average uses only the days that were recorded.
The next controlled step
Complete the accountability chain.
Select the weakest link in the current definition. The closing tree points to the next practical investigation.
Programme
The performance-management workstream
The role of the measurement routine in the Cameroon textile programme.
Web edition availableFollow this thread
Move from the cascade into a complete measure definition, the programme design and the data architecture that preserves meaning. Operate the models with explicit inputs before applying their outputs to a real line.
Every measure has an owner, a unit and a response.
Keep the question with the responsible role so a dashboard change does not become a substitute for an operating decision. Source context on this page comes from the programme annex. Source-domain references have not been independently reverified here. Related paper status is determined by the available web edition, not by a promised PDF.
Programme
The performance-management workstream