Separate the edit from the accepted baseline
Start with the record that says what was accepted. Then treat the new request as separate. If the team cannot identify the old baseline, request source, or person who may decide its disposition, it cannot responsibly call the work small yet.
NIST defines configuration control as controlling modifications before, during, and after implementation. That frames the ownership boundary; it does not establish cost, contract rules, or universal practice.
Use an ownership ledger, not a universal estimate
The ledger asks whether the change creates work in six layers. Each row requires an inspectable input and avoids assuming a price, rate, or result.
| Ownership layer | Question to inspect | Evidence or assumption to record |
|---|---|---|
| Accepted baseline | What did acceptance include? | Dated acceptance or closure artifact: __________ |
| Definition | Is this correction, request, or ambiguity? | New-request text and requester: __________ |
| Verification | What could change and how is it inspected? | Test or review procedure: __________ |
| Release and records | Which revision, approval, and rollback path apply? | Change record and release owner: __________ |
| Operation | Who supports the result and exception? | Operating owner and escalation route: __________ |
| Learning and exit | Which handoff record diverges? | Updated record list: __________ |
This is a strategic model, not an accounting forecast. It makes disagreements useful: a founder may expect routine verification while an operator identifies a downstream handoff that needs evidence.
Test the counterargument
A mature, well-owned system can absorb a clear, isolated request with negligible coordination when authority, verification, and operating capacity are predefined. The model therefore does not say every post-acceptance request is expensive. A dated sample of completed change records, a demonstrated automatic procedure, or measured effort in one operating context would improve the analysis.
For the wider lifecycle model, see Does AI Make Custom Software Cheaper to Own?. For accountable operation, see Cheaper Implementation, Valuable Accountable Operation. For human-owned work, see What Service Work Should Stay Human Before Productization?.
Does this model show post-acceptance changes are unprofitable?
No. It has no price, effort, revenue, or return dataset. Its purpose is to make ownership assumptions and the counterargument visible before a strategic claim is made.