Chapter 04SAP / DataArticle 4.1
Data Is a Signal, Not the Decision
SAP gives critical transactional context, but professional judgement is needed to interpret whether a material requirement should be purchased, transferred, held or challenged.
- ProblemSystem data is treated as if it were the decision.
- SignalsStock, PR, Min-Max, consumption, criticality, supplier and certificate status.
- DecisionMade by an accountable person, in context.
- ControlAn audit trail that shows the reasoning.
01Business problem
SAP and similar systems hold a great deal of information: balances, reservations, requisitions, orders, Min-Max levels, consumption, criticality codes, supplier and certificate status. Because the data is structured and always available, it is tempting to let it decide.
The trouble is that every field is a snapshot of something recorded by someone, at some point, for some purpose. A balance can be wrong. A reservation can belong to a cancelled job. A criticality code can be years old. When data is treated as the decision, those errors become decisions too.
02Core concept
Between system data and a decision there are steps the system cannot take on its own:
- Signal: the data shows that something may need attention.
- Context: the people and facts around the data confirm, qualify or contradict it.
- Judgement: an accountable person decides, and records why.
Automation is valuable in the first step. It makes signals visible, consistent and fast. It should not hide its assumptions, or replace accountable judgement where context matters.
03Decision rule
Treat a system value as a signal until its context has been checked. The more critical the item and the less reversible the decision, the more context is needed.
When the data and the context disagree, the disagreement is the finding. Correct the data, or record why the decision departs from it.
04Signals and inputs
- SAP stock
- The recorded balance, which may differ from the shelf.
- PR quantity
- What someone requested, not necessarily what is needed.
- Min-Max
- A setting based on past assumptions.
- Consumption
- History, including one-off events.
- Criticality
- A classification that may be out of date.
- Supplier status
- Promised dates and confirmations.
- Certificate status
- Recorded documents, which may not match the delivered item.
05Model
- 01System dataStock, PR, Min-Max, consumption, criticality, status.
- 02SignalWhat the data suggests needs attention.
- 03ContextWork plans, equipment, history, people who know the item.
- 04Professional judgementThe signal weighed against the context.
- 05DecisionProceed, reduce, transfer, clarify, hold, cancel or escalate.
- 06Audit trailThe reason, recorded where the next person will find it.
06Worked example
Illustrative example
Three signals, three readings
- The system shows 12 units free. The warehouse finds 9: three were issued without a transaction. The signal is corrected before any decision.
- A reservation of 20 units blocks a transfer. The work order behind it was cancelled months ago. Context changes the decision.
- Min-Max suggests no stock for an item with no consumption in six years. It is the only spare for a compressor without standby. Judgement overrides the signal, and the reason is recorded.
07Professional judgement
Good materials management converts system data into an operationally defensible decision, with enough context for another person to understand why.
Data quality matters, but perfect data is not the goal. The goal is knowing which values to trust for this decision, checking the ones that matter, and leaving a trail that shows what was checked.
Accountability cannot be automated. A system can recommend. A person decides, and should be able to explain the decision without saying that the system told them to.
08Failure modes
Stale data trusted
Decisions rest on balances, codes or dates that are no longer true.
ControlCheck the age and source of the values that drive the decision.
Missing context
A correct number leads to a wrong decision.
ControlConfirm work plans and equipment context for significant decisions.
Exceptions forced into rules
Unusual cases are processed as routine.
ControlRoute exceptions to a person with authority.
Automation with hidden assumptions
Nobody can explain why the system suggested an action.
ControlDocument the logic and thresholds behind automated signals.
No audit trail
The decision cannot be reviewed or learned from.
ControlRecord the reason with the decision.
09Field note
10Practical checklist
- The values that drive the decision are identified.
- Their age and source are known.
- Context is checked with the people who know the item or the work.
- Disagreement between data and context is resolved or recorded.
- The decision has an accountable owner.
- The reason is recorded with the decision.
11Connected intelligence
Experience
See this principle in practice
- Demand Review and Excess PreventionOperational experience
- Material Master, BOM and SPIL Data QualityOperational experience
12Evidence and basis
The examples are illustrative, not records of company transactions. Basis: requisition review, stock and Min-Max governance, technical evaluation and master-data work across the CV roles.