An incomplete inventory can favor well-documented systems when assessment or selection depends on fields that poorly understood applications lack. The resulting shortlist can reflect ease of description as much as business value or risk, even when no reviewer intentionally prefers better documentation.

This is a conditional selection mechanism. It does not establish that every portfolio is biased or that undocumented applications should automatically receive the highest priority. Missing evidence changes what can be concluded, not the underlying facts by itself.

Selection can happen before scoring

Suppose a candidate query requires an owner, support date, and recorded database relationship. Applications missing one of those fields can disappear before the prioritization model compares their consequences.

The surviving records may be accurately scored. The hidden problem is the population excluded from consideration. A clean ranking cannot explain applications that never reached it.

This constructed example suggests inspecting inclusion rules alongside scoring rules. Report the population considered, the records excluded, and the evidence gaps responsible for exclusion rather than presenting the result as a complete estate ordering.

Structural completeness can invite false certainty

An inventory schema can require fields without making their contents true. SHACL conformance checks evaluate data against shapes; passing a shape is a structural result whose meaning depends on the constraints defined.

Filling a missing version or owner with a guessed default can satisfy a field requirement while weakening the evidence. Explicit unknowns preserve the distinction between a required question and an observed answer.

Application records should also distinguish installed versions from product families and planned targets. Otherwise a well-formed record can describe what the organization intends to deploy rather than what currently performs the work.

Query semantics matter to the omission

SPARQL optional matching can preserve a solution when optional information is absent. An ASK result of false establishes no matching solution in the queried graph, not the absence of a dependency in the estate.

A decision query should return unknown prerequisites and missing evidence where possible. No recorded contraindication must remain distinct from a positive finding of suitability.

Test the query against a curated set containing incomplete, conflicting, stale, and superseded records. That reveals whether its behavior matches the intended treatment of uncertainty before it shapes a real portfolio.

Better documentation can be a legitimate advantage

A well-understood application can be more ready for a bounded migration trial. Selecting it to test a delivery method can be reasonable if that objective is explicit and does not replace investigation of consequential unknown systems.

The hypothesis weakens when the portfolio separates readiness from priority and maintains an investigation queue for incomplete evidence. It strengthens when missing records consistently escape scrutiny or receive favorable risk defaults.

The UK legacy-risk framework separates risk assessment from treatment selection and carries an August 2026 review notice. Neither its categories nor a local completeness score should be treated as an automatic rewrite decision.

Make the missing population actionable

AWS portfolio guidance asks assessors to examine CMDB freshness, refresh processes, gaps, and relationships to other evidence. Sample consequential records against deployed observations and retain source identifiers and conflicts.

Sensitivity analysis can then examine whether plausible resolutions of unknowns would alter priority or route. Investigation should focus on uncertainty capable of changing the decision, rather than maximizing filled cells without regard to consequence.

The practical result is a portfolio that can explain both its selected work and its unanswered questions. Documentation readiness remains useful evidence while no longer silently substituting for the business and operational reasons to act.