# ReportPatch Report Check specimen

SPECIMEN · SYNTHETIC DATA · NOT A CLIENT RESULT

Case RP-SYN-2026-10-01. One fictional monthly marketing report, intended October 1-31, 2026, America/New_York. This is supplied synthetic aggregate data, not a customer result. USD 99 offer: findings and Acceptance Record; implementation and fixes are excluded. No delivery SLA is represented here.

## Whole-case result

12 of the maximum 50 pairs: 7 within tolerance (6 exact, 1 nonzero), 3 discrepancies, 1 not comparable, 1 missing evidence. These counts cover the entire specimen, not only the findings below. HOLD; actual report approval is unresolved. No signature or actual owner acceptance.

## Method and trace

source-inputs.csv contains 23 fictional supplied rows (11 source, 12 report). checked-pairs.csv traces every checked pair to row IDs. sample-case.json carries the complete declarations and numeric values. All metrics are additive counts, not rates, averages, ratios or distinct-person metrics. Each pair uses a synthetic absolute tolerance of 1 unit and relative tolerance 0.001 (0.1%); allowed=max(1,abs(source)*0.001). These are demonstration choices and have not been approved by an actual report owner. Difference means report minus source. A metadata conflict stops numeric comparison; a missing source is not zero.

All comparable pairs match metric, units, dates, timezone, filter, attribution convention and refresh declaration. Default declarations: October 1-31 inclusive; America/New_York; all campaigns with test traffic excluded; event occurrence date with no cross-channel credit allocation; refresh November 1, 2026, 9:00 AM ET. Source/report declarations are fictional and have not been authenticated against source systems.

## Five findings

1. **P08 / S08 → R08: Email link clicks.** Source 200; report 260; difference +60. Allowed=max(1,200×0.001)=1 click. Absolute difference 60 exceeds 1. Obtain the matching extract and verify the report total.
2. **P09 / S09 → R09: Download events.** Source 80; report 100; difference +20. Allowed=max(1,80×0.001)=1 event. Absolute difference 20 exceeds 1. Verify the matching aggregate and reconcile the reported count.
3. **P10 / S10 → R10: Video start events.** Source 1,500; report 1,400; difference -100. Allowed=max(1,1500×0.001)=1.5 events. Absolute difference 100 exceeds 1.5. Verify the matching aggregate and reconcile the reported count.
4. **P11 / S11 → R11: Landing-page views.** Source ends October 31; report declaration ends October 30. Both supplied values are 2,100 views, but the date mismatch makes the pair not comparable. No numeric verdict is issued. Confirm the intended window and provide a matching report declaration or extract.
5. **P12 / R12: Phone-click events.** Report 37 events; no source row supplied. Difference and allowed difference cannot be calculated. Supply a source total with the complete matching declarations.

P05 is the nonzero pass: S05 10,000 display impressions versus R05 10,005; difference +5; allowed=max(1,10000×0.001)=10. P01-P04, P06-P07 match exactly. A within-tolerance result only applies to these supplied aggregates and declarations.

## Limitations and next checks

No live access, raw-event review or source authentication occurred. The specimen does not establish completeness, causation, correct tracking, financial assurance or approval to send a report. Rates, ratios, averages, distinct people, deduplication and implementation/fixes are outside scope. No explanation of the numeric differences is asserted.

Resolve the three numeric differences; align P11 dates; obtain P12 source evidence; agree tolerances with the real report owner; rerun affected pairs; then request the authorized owner's actual approval. Keep HOLD until that review is resolved. Reconciliation alone does not grant approval.