Count accepted changes, not generated proposals
Treat every generated diff as a proposal until the business's normal acceptance process binds it to an intended outcome, tests, assumptions, known limitation, and accountable maintainer. A busy proposal queue may be valuable experimentation. It is not the same thing as a supported capability the business can sell or operate.
NIST's Secure Software Development Framework is relevant only as common language for documented practices between producers and acquirers. It does not say how much review is enough or that a documented process produces a moat. The strategic question is narrower: can the business account for the changes it chooses to depend on?
Keep five evidence fields visible
Use the accompanying review-capacity ledger to distinguish proposal identity, accountable acceptance, review evidence, recovery ownership, and accepted versus queued work. The last field matters because a team can generate a large backlog without increasing the number of changes that a maintainer will stand behind.
Ask a simple counterquestion before treating output as an advantage: if the generator stopped today, which accepted changes could a qualified maintainer still explain, test, update, and recover? The answer may expose a real operating asset, a concentrated person-dependency, or an unaccepted queue. None of those outcomes is a valuation by itself.
Make the investment decision narrower
If acceptance and recovery ownership are the bottleneck, invest in explicit review boundaries, test evidence, or maintainer context before expanding the proposal queue. If accepted changes remain legible across releases, investigate whether that accountable operation supports a durable customer outcome. Do not infer it merely from throughput.
This model does not measure productivity, code quality, security, revenue, or defensibility. It makes one often-hidden denominator visible. For the broader economic question, see Cheaper Implementation Shifts Value to Accountable Operation and Why Releases Two Through Ten Determine Software Economics.
Is a passing test suite enough to count a proposal as capacity?
No. Tests can be important evidence, but the business still needs a named acceptance decision, known boundary, and recovery owner for the exact change.
Does a large review queue prove future advantage?
No. It may represent useful options or unmanaged work. Record accepted and unaccepted proposals separately before drawing an operating conclusion.