Shared Cost Allocation Balance Evidence

Check a monthly shared-cost export against the original cost pools and the fixed shares your organization has already approved. This Actor reports missing submitted outputs, source-to-output residuals, unexpected target identities, duplicate identities and wrong-scope rows. It is useful when an allocation spreadsheet or a separate allocation engine produces a report that needs an independent arithmetic check before operations staff review it.

Supply three normalized tables: original source pools, approved fixed-share rules and independently observed allocation output. Use JSON arrays or strict mapped CSV for each table. No provider account, resource credentials or paid external API is needed. The Actor does not fetch your cloud bill or choose an allocation method. It does not determine fair shares, amortization, accounting treatment, tax obligations or transfer pricing.

Input

Every pool row has rowId, poolId, periodId, currency, costBasis and amountMinor. Declare the basis as billed, effective or custom. Each poolId identifies one unique source amount and scope. Keep separate currencies, periods and bases in separate pools. Never combine billed and effective amounts for the same cost. The currency identifier is exact supplied text; the Actor neither validates a currency standard nor converts currencies. You normalize decimal amounts into your chosen currency's minor units before submission.

Each rule has rowId, poolId, targetId and integer shareBasisPoints. A usable rule set requires unique target and rule row identities and exactly 10000 total basis points. Missing or conflicting rules remain unresolved. They are not repaired or normalized.

Each observed output has rowId, outputLineId, poolId, targetId, periodId, currency, costBasis and amountMinor. Distinct output lines can sum for the same pool and target. Duplicate row or line identities make affected comparisons ambiguous, including collisions across pools. Wrong period, currency or basis rows stay visible but are excluded from the matching pool's sum. No duplicate occurrence is selected or counted twice.

All identifiers are case- and whitespace-sensitive strings, including 0 and leading zeros. Amounts are canonical signed integer strings with at most 30 digits: 0, 100, -100. Negative credits and zero pools are supported. Plus signs, decimals, exponents, separators, leading zeros in amounts and negative zero are rejected. The complete input must be at most 4 MB and contain at most 2000 combined rows.

Declare allocationCoverage as complete, partial or not-provided, and explicitly set toleranceMinor from 0 to 100. Only complete coverage permits absence-in-submitted-export evidence. A missing target under partial coverage is unknown, never an observed zero. Complete coverage is your declaration about submitted data, not proof of live-provider completeness. A complete missing target's 0 is the empty submitted sum, not a zero payment.

Results

One dataset item contains the complete report, original normalized rows, pool and target evidence, source references and findings. Download full JSON, evidence CSV, findings CSV or standalone HTML. JSON preserves exact amounts as decimal strings. Each target shows expectedNumerator / 10000 and discrepancyNumerator / 10000; fractional minor units remain exact. No rounding policy is chosen. The tolerance annotates exact differences without hiding them. A nonzero pool residual means source minus qualified submitted output is nonzero; incomplete coverage or wrong-scope rows can keep its interpretation unresolved.

CSV exports quote cells and prefix potential spreadsheet formulas with an apostrophe, including signed negative strings. Import identifier and amount columns as text if your spreadsheet automatically changes numeric strings. HTML escapes supplied content and has no external scripts. Large retained-reference reports that exceed the 8 MB JSON or 9 MB per-export limit are rejected before charging; split such batches.

The proposed price is $0.25 per completed report, with platform usage included. An insufficient event budget delivers no report and charges no report event. A fresh run is billable again. Resuming the same run checks the saved dataset's input fingerprint and integrity, restores missing exports and avoids another dataset item or report charge. An interruption before completion may leave partial exports; resume the same run to finish recovery.

Raw inputs and reports are stored in Apify storage. Use opaque pool and target identifiers, omit contact details and resource names, and submit only data you are authorized to process. This report provides supplied-data evidence for review. It does not certify FOCUS conformance, provider acceptance, invoice settlement, financial correctness or the quality of your allocation policy.

The prefilled example contains balanced splits, dropped targets, excess output, duplicate identities, wrong scope, a negative credit, a zero pool and fractional expected shares. Replace it with your own normalized data and explicit choices before using the report for a real monthly workflow.