← Back to the demo
This document was generated by the close itself and is shown exactly as produced (draft status and all). How it fits in the workflow is explained on the main page. Download variance_table.csv → the full 840-row variance grain, open in Excel or Sheets

Variance Report: EventCo Budget vs Actual

Generated by agents/variance_agent.py from data/eventco_monthly_cleaned.csv (the Data Ingestion Agent's output) and the internal business notes log data/business_notes.csv. This agent never reads data/ground_truth.md.

Method

Variance = actual − budget, computed for every line item, BU and month (28 BU/line series across 30 months; full grain in output/variance_table.csv). Direction is F (favorable) or U (unfavorable): revenue above budget is favorable, cost above budget is unfavorable.

Materiality:

Root-cause discipline: this agent never invents a cause. Every material row carries exactly one of three explanation types, and the type is always visible:

  1. Corroborated by the business notes log: a dated, BU-relevant note found automatically (right BU or group-wide, dated within the evidence window of 10 days before to 20 days after the variance period, and actually about that kind of spend).
  2. Analyst input: no note corroborates the variance, so it went to the FP&A analyst as a follow-up; the explanation shown is what the analyst wrote back in data/analyst_commentary.csv after investigating with the business. Labeled as manual input so it can never be mistaken for machine-found evidence.
  3. No clear driver identified: no note and no analyst input yet. The item stays on the follow-up list; the number is reported without a story.

Excluded from analysis: suspected data entry errors

Same magnitude rule as the ingestion agent (actual >4x or <0.25x budget is not a plausible business event). These rows are withheld from variance analysis and materiality testing entirely, pending correction at source. Explaining them as business variances would be inventing a story about a typo.

Business Unit Line item Month Actual Budget
Brand Events Revenue 2025-11 EUR52,243,584 EUR5,238,445
All 20 material variances as P&L impact, hatched where no documented note exists and the item went to the analyst

Material variances

Sorted by absolute EUR impact. Period spans of more than one month are sustained episodes; Actual/Budget/Variance are cumulative over the period. Evidence cites note IDs from data/business_notes.csv.

Brand Events · COGS · 2025-04..2025-06 · +2,077,456 · +26.6% · Unfavorable

Corroborated by the business notes log: N11 (2025-04-14, Brand Events, Project Director): "Falcon product-launch event (Riyadh): client requested a major on-site scope expansion during build week. Overtime crews, additional staging and expedited freight all booked at premium rates. Expect a production cost overrun through the end of Q2 while the change-order negotiation with the client is still open." / N13 (2025-06-30, Brand Events, BU Controller): "Q2 close: Falcon launch cost overrun confirmed at roughly EUR2.0m above plan across April-June, concentrated in external production costs. Change order signed for partial recovery only; margin impact flagged to group controlling."

Evidence: N11, N13

Brand Events · COGS · 2026-04 · +407,217 · +15.7% · Unfavorable

Analyst input (FP&A Analyst, 2026-05-12): "An unplanned client project was won in late March and delivered inside April; external production was booked at short-notice premium rates. Together with the matching revenue line the net margin impact is slightly negative. Change-order discipline flagged to the project director." Written manually after follow-up; not derived from the business notes log.

Brand Events · Revenue · 2026-04 · +388,823 · +8.3% · Favorable

Analyst input (FP&A Analyst, 2026-05-12): "Unbudgeted revenue from the same late-won April project; see the matching COGS variance. Recognized on delivery." Written manually after follow-up; not derived from the business notes log.

Brand Events · COGS · 2024-11 · -294,277 · -10.6% · Favorable

Analyst input (FP&A Analyst, 2024-12-10): "November events were delivered with in-house crew and staging stock instead of the subcontracting assumed in budget; Q4 freight was also renegotiated. Genuine savings, not a timing shift." Written manually after follow-up; not derived from the business notes log.

Brand Events · Revenue · 2025-07 · +247,346 · +7.3% · Favorable

Analyst input (FP&A Analyst, 2025-08-11): "A client brought part of its September roadshow forward into July. Timing between months rather than new business; the September shortfall stayed within tolerance." Written manually after follow-up; not derived from the business notes log.

Digital/Influence · Revenue · 2025-01 · -224,810 · -14.8% · Unfavorable

Analyst input (FP&A Analyst, 2025-02-10): "Two platform go-lives slipped from January into February on client-side content delays. Revenue is recognized on delivery, so January carries the gap; both projects were invoiced in February." Written manually after follow-up; not derived from the business notes log.

Corporate Events · Revenue · 2024-10 · +215,565 · +10.5% · Favorable

Analyst input (FP&A Analyst, 2024-11-08): "A brand activation sold for early 2025 was pulled into October at the client's request. The matching cost variance appears on the COGS line the same month." Written manually after follow-up; not derived from the business notes log.

Digital/Influence · Revenue · 2025-09 · -196,953 · -9.3% · Unfavorable

Corroborated by the business notes log: N08 (2025-09-26, Digital/Influence, Account Manager): "NovaTech (US client) roadshow is invoiced in USD. The euro strengthened sharply against the dollar this month, so reported revenue on the contract translates materially lower in EUR. No change in delivered scope or client commitment; delivery costs are EUR-denominated and unaffected."

Evidence: N08

Digital/Influence · COGS · 2024-12..2025-01 · -143,626 · -12.3% · Favorable

Analyst input (FP&A Analyst, 2025-02-10): "External development spend paused over the year-end delivery freeze and resumed in February. This recurs every holiday season; budget phasing to be corrected in the next cycle." Written manually after follow-up; not derived from the business notes log.

Digital/Influence · Revenue · 2025-08 · +135,341 · +10.2% · Favorable

Analyst input (FP&A Analyst, 2025-09-09): "One retainer client commissioned an out-of-cycle campaign microsite in an otherwise quiet August." Written manually after follow-up; not derived from the business notes log.

Corporate Events · COGS · 2024-10 · +125,565 · +17.4% · Unfavorable

Analyst input (FP&A Analyst, 2024-11-08): "Production costs of the pulled-forward October activation; see the matching revenue variance." Written manually after follow-up; not derived from the business notes log.

Government & Institutions · Revenue · 2024-10 · -115,950 · -11.3% · Unfavorable

Analyst input (FP&A Analyst, 2024-11-08): "Two institutional clients pushed their October milestone invoicing into November after a budget-cycle approval delay on their side. Timing only." Written manually after follow-up; not derived from the business notes log.

Digital/Influence · COGS · 2024-07 · -84,412 · -15.6% · Favorable

Analyst input (FP&A Analyst, 2024-08-09): "Contractor spend deferred during the summer staffing gap on the delivery bench; caught up from September." Written manually after follow-up; not derived from the business notes log.

Digital/Influence · COGS · 2025-06 · -82,563 · -10.4% · Favorable

Analyst input (FP&A Analyst, 2025-07-10): "June delivery mix skewed toward license resale rather than custom build, which carries lower external cost." Written manually after follow-up; not derived from the business notes log.

Corporate Events · Opex - Marketing · 2025-04..2026-02 · -81,464 · -16.4% · Favorable

Corroborated by the business notes log: N12 (2025-06-25, Corporate Events, Practice Lead): "Media buying and content creation move in-house from 1 July; both external agency retainers terminated end of June. Marketing spend expected to run 20-30% under budget for the rest of the year." / N16 (2025-09-15, Corporate Events, BU Controller): "In-housing of media buying is tracking ahead of plan - marketing spend well under budget again this month, and the underspend should persist through year-end."

Evidence: N12, N16

Corporate Events · COGS · 2024-06 · -75,009 · -10.5% · Favorable

Analyst input (FP&A Analyst, 2024-07-09): "The June campaign reused spring creative assets instead of commissioning a new shoot." Written manually after follow-up; not derived from the business notes log.

Corporate Events · COGS · 2025-12 · +71,249 · +10.3% · Unfavorable

No clear driver identified: no corroborating note found in the business log for this BU, line item and period, and the analyst has not yet returned commentary from follow-up. The variance is numerically material but unexplained. Recommend follow-up with the BU controller rather than attributing a cause.

Digital/Influence · COGS · 2025-03 · +66,523 · +11.3% · Unfavorable

No clear driver identified: no corroborating note found in the business log for this BU, line item and period, and the analyst has not yet returned commentary from follow-up. The variance is numerically material but unexplained. Recommend follow-up with the BU controller rather than attributing a cause.

Corporate Events · COGS · 2025-03 · +55,202 · +10.0% · Unfavorable

Analyst input (FP&A Analyst, 2025-04-10): "Print and media rates increased at the main supplier from March; a rate-card renegotiation is underway." Written manually after follow-up; not derived from the business notes log.

Government & Institutions · Opex - IT · 2024-11 · +24,116 · +158.1% · Unfavorable

Corroborated by the business notes log: N06 (2024-11-08, Government & Institutions, AV & Systems Manager): "On-site registration and AV control systems failed during a ministry summit engagement on 4 November; emergency replacement hardware and expedited vendor licenses procured outside the normal purchasing cycle. One-off cost; insurance claim filed with the provider."

Evidence: N06

All other BU/line/month combinations are within materiality tolerance and are not individually commented; the full variance grain is in output/variance_table.csv.