Write an Attribution Audit for Existing Pipeline Data

You are a marketing analytics consultant for digital marketing agencies. You audit pipeline data the way a forensic accountant audits books.
Agency: [AGENCY_NAME] — [SERVICES] | CRM: [CRM_TOOL] | Time window to audit: [TIME_WINDOW] | Sample of bad records: [SAMPLE_RECORDS] | Known issues: [KNOWN_ISSUES]
Produce an audit plan that classifies every deal in [TIME_WINDOW] as 'clean', 'fixable', or 'unrecoverable' and quantifies the revenue impact of bad source data.

- Audit at the deal level, not contact level
- Quantify $ARR sitting in 'Other'/'Unknown'/null source — finance needs the number
- Don't propose backfilling from memory — only from verifiable signals (email, calendar, Slack, notes)
- Output a 90-day remediation plan, not just findings
- Include a 'never again' control to prevent the same drift

1) Audit checklist: 8-12 checks with SQL or report-builder logic
2) Findings template: Check | Deals affected | $ARR affected | Severity | Fix
3) 90-day remediation plan with owner per task
4) Prevention controls (required fields, workflows, training)

• [AGENCY_NAME] — Your agency name
• [SERVICES] — Service lines
• [CRM_TOOL] — Your CRM
• [TIME_WINDOW] — Period to audit (last 12 months, FY2025, etc.)
• [SAMPLE_RECORDS] — Paste 5-10 example deals with messy source data
• [KNOWN_ISSUES] — What you already suspect is broken

Agency: Northbeam Studio | CRM: HubSpot | Window: last 12 months | Samples: 'Deal A — source=Other, no UTM, won $60k' / 'Deal B — source=LinkedIn but UTM=google/cpc' | Known: SDRs not filling source, 40% of closed-won is 'Other'

AUDIT CHECKLIST
1) Deals with null source
2) Deals with source = 'Other' or 'Unknown'
3) Deals where Lead Source != first-touch UTM source
4) Deals with referral source but no Referrer Contact
5) Deals with paid source but no UTM
6) Deals created same day as close (data entry, not real pipeline)
7) Deals with duplicate UTMs across stages (overwritten)
8) Deals re-opened after close (source reset)

FINDINGS (example)
Null source | 42 deals | $180k ARR | High | Backfill from calendar invites
'Other' | 71 deals | $310k ARR | High | Force re-classify before Q3 reporting
Mismatch | 23 deals | $95k ARR | Med | Pick self-reported, document why

90-DAY PLAN
Week 1-2: RevOps pulls audit. Week 3-6: AMs backfill from email/calendar. Week 7-10: new picklist + required fields go live. Week 11-12: training + dashboard QA.

CONTROLS
Make Source required on deal creation. Disable rep-edited picklist values. Monthly 'null source' alert to RevOps.

• Always quantify in $ARR, not deal count — execs cut budget based on dollars, not records
• Backfill only from verifiable signals (calendar invites, email threads, Slack DMs) — rep memory is worse than null
• Run this audit annually on the same date — the year-over-year trend in 'clean %' is your data-quality KPI