Campaign Spend vs Daily Plan Evidence Report

Compare your supplied daily media plan with authorized spend exports from several advertising platforms, without connecting an ad account. This report is for analysts and agencies who already keep campaign plans in spreadsheets and need a repeatable handoff showing which comparisons are supported by the export, which days remain unknown, and where arithmetic differs from the declared plan.

What you receive

Each completed run produces one complete report dataset item. Its campaign view shows planned and actual amounts through the declared cutoff, cumulative variance, the full declared plan, and the remaining declared plan amount when the evidence is complete. Its daily evidence view keeps the plan row, observed spend, coverage declaration, identifiers and source references together. Download full JSON, daily evidence CSV, findings CSV and a standalone HTML report with campaign summaries and daily detail.

A positive variance means actual minus the caller's plan is positive. It does not establish an overcharge, a platform spending-limit violation, or poor campaign performance. Google and other platforms have their own budget and billing rules; this tool compares your supplied allocations rather than enforcing those rules.

Input and workflow

1. Choose a snapshotId and planVersion. Declare windowStart, windowEnd and asOfDate as real YYYY-MM-DD calendar dates. The cutoff includes the entire as-of date. Normalize all exports to your chosen reporting timezone before upload; the Actor does not reconcile platform timezones.
2. Supply plans, actuals and coverage arrays, including an explicit empty array for an empty table. Alternatively supply the corresponding CSV and remove that array. CSV supports a UTF-8 BOM, quoted cells and a complete canonical-field to header mapping. Duplicate or unmapped headers, ragged rows and unsupported dates are rejected.
3. Each plan and actual row needs recordId, sourceId, campaignId, date, currency, costBasis and amount. Amounts must be decimal strings, such as "12.3456", with at most twelve integer and six fractional digits. A source/campaign scope has exactly one currency and one literal cost basis, for example billed or served. The Actor never converts currencies or combines different cost bases.
4. Each coverage row needs recordId, sourceId, campaignId, currency, costBasis, startDate, endDate and completeness (complete, incomplete or unknown). Ranges include both boundary dates and cannot overlap within the same source/campaign scope.
5. Set missingPlannedDayPolicy to zero only when an omitted plan allocation means zero in your declared plan. Otherwise use unknown. Review the report before using it in a media-plan handoff.

Full daily actual totals require caller-declared complete export coverage, even when an actual row is present. An observation under incomplete, unknown or absent complete coverage remains a known lower-bound amount; its complete actual total and variance stay unknown. An absent actual row becomes zero only inside complete coverage for that exact scope. Your coverage declaration is not independently verified. Duplicate source/campaign/date rows are rejected, including identical repeats, because accidental duplication and distinct spend components cannot be distinguished safely. Aggregate legitimate components in your export before upload and preserve the correct cost basis.

Interpreting remaining amounts

The report uses exact fixed-point arithmetic, not floating-point money sums. When full-plan and through-cutoff actual evidence are complete, remaining declared plan equals full supplied plan minus actual through cutoff. A negative amount is retained. The optional remaining-per-positive-planned-day value is an exact numerator and denominator, counting future dates with explicit positive allocations. It is an arithmetic expression, not a forecast, budget recommendation or automatically rounded daily setting. If there are no such future dates, the expression is null.

Future actual observations and out-of-window rows remain traceable in the report but do not enter comparison totals. Unknowns are null, distinct from zero. Different source/campaign scopes remain separate; there is no misleading monetary total across currencies or cost bases.

Limits, privacy and pricing

Input is limited to 4 MB and 5,000 combined rows. A window contains at most 366 dates and expansion at most 10,000 daily evidence rows. Large export artifacts are rejected before charging: complete JSON must stay below 8 MB and each download below 9 MB. Shorten the window or split campaign scopes when limits are reached.

Use opaque identifiers and authorized exports only. This Actor makes no external network requests, reads no account credentials, changes no campaigns and calls no paid models. Apify retains cloud input and report artifacts according to its platform settings. HTML escapes supplied content; CSV neutralizes spreadsheet formula prefixes.

Pricing is $0.25 per completed report, with platform usage included. Invalid input or insufficient report budget is not charged a report event. A fresh run is a new billable report. Restarting the same run recovers durable exports and preserves one report item and one event; interrupted exports may temporarily be partial until that run resumes. This evidence report does not diagnose ROAS, attribution or campaign performance and does not replace native platform reports.