The short answer
Start with the complete reconciliation population for the period, reconcile its status counts, and stratify by consequence and failure risk. Select samples reproducibly, review the exact signed version and supporting evidence, expand testing when findings indicate a wider defect, and state precisely what the sample does and does not support.
Key takeaways
- No sample is interpretable without a denominator.
- Select high-risk and random items separately.
- A finding may require population reconstruction, not just one correction.
Reconcile the population before selecting workpapers
List every required property, bank account, entity, period, preparer, reviewer, status, completion date, exception count, and version. Resolve missing, duplicated, reopened, and late records before sampling.
Do not build the denominator from a folder of completed PDFs; that method excludes work that never reached the folder. Use the governed close population and compare it with the evidence repository.
Separate coverage from risk selection
Document the reason each stratum exists and keep random selection reproducible.
| Stratum | Selection signal | Question answered | Expansion trigger |
|---|---|---|---|
| High consequence | Large cash or owner release | Were consequential conclusions supported? | Any unsupported balance or authority |
| Known exception | Reopened or qualified sign-off | Was the exception resolved as approved? | Same cause elsewhere |
| Change | New preparer, bank, mapping, or process | Did the change preserve the control? | Control-version defect |
| Aging | Late or repeatedly carried item | Is closure evidence genuine? | Hidden backlog pattern |
| Random | Seeded selection from remaining population | What do ordinary items show? | Failure above approved tolerance |
Write a conclusion no broader than the evidence
Record the population date, strata, sample IDs, workpaper versions, procedures, findings, corrections, expanded review, unresolved exceptions, and reviewer. Distinguish design failures from isolated execution errors.
Do not convert a convenience sample into an error rate or performance claim. If the population is incomplete, the selected workpaper changed after review, or findings cannot be bounded, the portfolio sign-off remains qualified.
Operational checklist
Mark your progress, then save a working copy. Selections reset when you leave this page. A checked box is not an approval or evidence of completion.
☐
Population denominator reconciled
☐
Status and version fields complete
☐
Risk strata documented
☐
Random method reproducible
☐
Exact workpapers retained
☐
Findings mapped to population
☐
Conclusion scope and limitations stated
0 of 7 marked
Edge cases
- One workpaper covers several bank accounts: define the review unit before counting.
- A sampled item changes after sign-off: invalidate or repeat the affected procedure.
- A finding points to a shared import rule: test the rule’s full affected population.
Sources and references
Follow each source to check the underlying claim. Access checks and professional review are different steps.
Continue the workflow
Three-way reconciliation for property managersProperty-management close calendar dependency registerHow to reconcile an owner statementRevision history
2026-09-18
Initial Phase 4 operational article with a distinct evidence artifact, failure states, source limits, and AI-assisted technical review.