Attribution Audit Template (2026): 7-Step Checklist to Fix Broken Tracking
An attribution audit walks through 7 layers: UTM naming consistency, tracking pixels, conversion events, attribution model alignment, conversion windows, channel definitions, cross-platform reconciliation. Copy-ready checklist that catches the errors that distort 20%+ of analytics decisions.
Attribution Audit Template (2026): 7-Step Checklist to Fix Broken Tracking
An attribution audit is a systematic review of how marketing performance gets measured - from the moment a user clicks an ad to the moment that user appears in the CRM as a converted customer. The version that catches the errors costing 20%+ of analytics accuracy covers seven layers: UTM naming consistency, tracking pixels, conversion events, attribution model alignment, conversion windows, channel definitions, cross-platform reconciliation. Regular monthly audits typically improve ROI visibility by 20% - not by changing the marketing, but by catching the data errors that were distorting decisions.
Key takeaways
- Seven layers in the audit: UTM naming, tracking pixels, conversion events, attribution model alignment, conversion windows, channel definitions, cross-platform reconciliation.
- The most common attribution problem is UTM naming inconsistency - "Google" and "google" treated as different sources, splitting and mismeasuring 5-15% of traffic.
- Monthly audits catch errors before they distort weeks of decision-making. Quarterly is too slow; weekly is overkill for most teams.
- Attribution model alignment matters more than which model is "correct" - picking one consistent model and applying it everywhere beats switching between models.
- Cross-platform reconciliation (Meta vs GA4 vs Shopify discrepancies) requires a documented expected range - 10-20% gap is normal, 40%+ indicates a tracking break.
One root cause of broken attribution is launching without complete instrumentation in place. Tracking added after launch cannot reconstruct lost data; every week of post-launch operation without proper instrumentation produces unrecoverable measurement gaps. For the antipattern, see the add tracking later trap.
Why broken attribution costs more than it appears
A marketing team makes a $50K budget reallocation decision based on Meta's reported ROAS of 2.8× - only to discover three months later that the conversion pixel was misfiring on 30% of orders, making the true ROAS closer to 2.0×. The reallocation was wrong; the team spent two months scaling the wrong channel and pulling back the right one. The decision wasn't bad - the data feeding the decision was broken. Attribution audits surface these breaks before they propagate into budget decisions.
Attribution audit: a systematic review of marketing measurement infrastructure (UTM tags, tracking pixels, conversion events, attribution models) to identify and fix errors that distort the data used for budget and channel decisions.
01 - Layer 1: UTM Naming Consistency
The foundation. UTM parameters identify where traffic came from; inconsistent UTMs split the data and mismeasure channels.
Fill-in-the-blank checklist:
UTM Naming Audit:
[ ] All UTMs lowercase (e.g., "google" not "Google")
[ ] All UTMs use kebab-case (e.g., "spring-sale" not "Spring Sale")
[ ] utm_source values match documented standard:
Expected: google, bing, meta, instagram, linkedin, tiktok,
twitter, pinterest, youtube, email, direct, referral
Audit query:
SELECT utm_source, COUNT(*) FROM sessions WHERE date > [recent]
GROUP BY utm_source ORDER BY count DESC
Look for: variants of the same source (Google/google/GOOGLE),
misspellings, deprecated sources still appearing
[ ] utm_medium values match documented standard:
Expected: cpc, paid_social, organic, email, social, referral,
display, retargeting, video
[ ] utm_campaign values follow naming convention:
Standard format: [year]_[Q]_[channel]_[campaign_type]_[product/audience]
Example: 2026_q3_meta_promo_enterprise
[ ] No active campaigns missing UTMs (find via direct landing page hits
from paid sources without proper UTM parameters)
The most common attribution problem at this layer: case inconsistency. "Google" and "google" treated as separate sources splits 5-15% of traffic and inflates source diversity in reports.
For related operational depth, see UTM governance guide.
02 - Layer 2: Tracking Pixels
The technical layer. Pixels and tags must fire on the right events, in the right contexts, for attribution to work.
Fill-in-the-blank checklist:
Tracking Pixel Audit:
Meta Pixel:
[ ] Pixel installed on all key pages
[ ] PageView event firing on every page load (test with Meta Pixel Helper)
[ ] Purchase event firing on order confirmation pages with value parameter
[ ] AddToCart, InitiateCheckout, Purchase events firing in sequence
[ ] Conversion API (CAPI) configured for server-side backup
[ ] Match quality score above 6.5 (Meta Events Manager)
Google Ads / GA4:
[ ] Google Tag Manager container correctly installed
[ ] GA4 measurement ID matches active property
[ ] Conversion events configured: form_submit, demo_request, purchase, etc.
[ ] Enhanced conversions configured (email hash passed back)
[ ] Cross-domain tracking configured if multiple domains
LinkedIn Insight Tag / TikTok Pixel / Pinterest Tag:
[ ] Tag installed and firing
[ ] Conversion events configured per platform
[ ] Enhanced conversions / first-party data passing where supported
Server-Side Tracking:
[ ] Conversion API for Meta, Conversions API for Google, etc.
[ ] Server-side dedup with browser-side pixel
[ ] Server-side tracking covers 80%+ of conversions
03 - Layer 3: Conversion Events
What counts as a conversion must be defined consistently across platforms. Different definitions break attribution math.
Fill-in-the-blank checklist:
Conversion Event Audit:
[ ] Documented list of conversion events with definitions:
Lead: [What action triggers? form fill, demo request, etc.]
MQL: [What criteria distinguish from lead?]
SQL: [What's the sales acceptance criteria?]
Purchase: [Order confirmation vs. successful payment?]
[ ] Each conversion event has a deduplication rule
(same email within X days doesn't double-count)
[ ] Conversion values passed back to platforms:
Purchase value, currency, transaction ID
Lead value (if using value-based bidding)
[ ] Conversion events test passed in last 30 days:
End-to-end test from ad click to conversion to platform reporting
[ ] No "phantom conversions" appearing without source:
Audit conversions with utm_source IS NULL - should be near zero
for paid campaigns
04 - Layer 4: Attribution Model Alignment
Which attribution model is being used must be documented, and the same model must apply across reports for the same business question.
Fill-in-the-blank checklist:
Attribution Model Audit:
[ ] Documented attribution model for each business question:
Marketing-sourced pipeline: [first-touch / multi-touch / etc.]
Channel ROI: [last-click / last-non-direct / multi-touch / etc.]
Campaign performance: [view-through / click-only / etc.]
[ ] GA4 attribution model setting matches documented standard
[ ] HubSpot/Salesforce attribution model matches documented standard
[ ] Meta Ads Manager attribution window matches documented standard
[ ] Google Ads conversion attribution matches documented standard
[ ] Same metric (e.g., "marketing-sourced revenue") uses same model
across all reports - board, monthly, weekly
[ ] If multi-touch attribution: weights documented and consistent
[ ] Attribution model changes flagged in reports for at least 90 days
after change (so historical comparison is honest)
For the framework depth, see marketing attribution explained and multi-touch attribution.
Run marketing on one source of truth
Every source in one brief, so the team stops reconciling exports.
14 days free · no credit card
05 - Layer 5: Conversion Windows
The time horizon for crediting a conversion to a marketing touchpoint affects reported performance dramatically.
Fill-in-the-blank checklist:
Conversion Window Audit:
[ ] Documented conversion window for each platform:
Meta Ads: [7-day click + 1-day view / 1-day click only / etc.]
Google Ads: [30 days / 60 days / 90 days]
LinkedIn Ads: [30 days / 90 days]
[ ] Window matches typical customer decision cycle:
B2B SaaS: 30-90 days typically
DTC consumables: 7-30 days typically
Considered purchase: 30-90 days
[ ] Same conversion window used for the same business question
across platforms (no Meta 7-day + Google 30-day comparison)
[ ] Window changes documented and reports flagged accordingly
[ ] For DTC: first-purchase ROAS window separated from
90-day cohort ROAS window - see related guide
For depth on time-window choice, see first-purchase vs 90-day ROAS and marketing attribution windows explained.
06 - Layer 6: Channel Definitions
What counts as which channel must be defined consistently. Otherwise, channel-level comparisons are meaningless.
Fill-in-the-blank checklist:
Channel Definition Audit:
[ ] Documented channel taxonomy:
Paid Social: [Which platforms - Meta, Instagram, TikTok, LinkedIn?]
Paid Search: [Google + Bing? Display included or separate?]
Organic Search: [SEO only? Includes branded vs non-branded split?]
Direct: [Includes/excludes specific referrers?]
Referral: [What counts? Affiliate programs, partner links, press?]
Email: [Owned email? Includes third-party newsletter sponsorships?]
[ ] Channel definitions match across GA4, ad platforms, CRM
[ ] No "Other" / "Unknown" channel exceeding 5% of attributed traffic
[ ] Paid retargeting separated from paid prospecting (different LTV math)
[ ] Organic-search and direct distinguished from branded direct traffic
[ ] Channel renames trigger a 90-day flag in reports
07 - Layer 7: Cross-Platform Reconciliation
The final layer. Numbers from different platforms must reconcile within an expected range.
Fill-in-the-blank checklist:
Cross-Platform Reconciliation Audit:
Expected gaps (within these ranges = normal):
Meta-reported revenue vs Shopify revenue: ±15-25% (attribution windows differ)
GA4 sessions vs ad-platform clicks: ±10-20% (bot filtering, redirects)
HubSpot pipeline vs Salesforce pipeline: <5% (these should match closely)
Stripe revenue vs Shopify revenue: <2% (these should match very closely)
Reconcile each pair:
[ ] Meta Ads reported revenue: $X
Shopify revenue from Meta-sourced sessions: $Y
Gap: Z% - within expected range? [yes/no]
If no: [investigate cause]
[ ] Google Ads reported conversions: X
GA4 conversions from Google Ads: Y
Gap: Z% - within expected range?
[ ] LinkedIn Ads reported pipeline: $X
HubSpot LinkedIn-sourced pipeline: $Y
Gap: Z% - within expected range?
[ ] Total revenue (Stripe): $X
Sum of channel-attributed revenue (all platforms): $Y
Unattributed: $Z
Unattributed % - is it under 20%?
For the DTC reconciliation framing specifically, see why ROAS misleads DTC and the MER framework.
What separates a working attribution system from a broken one
The ICP problem this section addresses: a marketing team trusts their dashboards, makes budget decisions based on reported ROAS, and then discovers (often during a fundraise or strategic review) that the underlying tracking has multiple breaks costing 20-40% accuracy. The cost isn't just the wasted spend on misallocated budget - it's the credibility damage when the discrepancy surfaces.
Analysis of marketing-tracking breakdowns consistently shows that the most-common attribution errors are quiet rather than loud. Loud errors (pixel completely missing) get noticed because reporting shows zero conversions. Quiet errors (UTM case inconsistency, conversion event firing on wrong page, attribution model mismatch between Meta and CRM) get accepted because the numbers look plausible. The result is dashboards that look fine but produce decisions that are 20-40% off optimal.
The mechanism is verification friction. Setting up tracking is one team's job (often growth engineering or RevOps); reviewing tracking is no one's job. Without scheduled audits, the system drifts - minor changes accumulate, small errors compound, and by the time the discrepancy is large enough to notice through performance changes, months of decisions have been made on broken data.
The operational implication: monthly attribution audits are the lowest-cost, highest-leverage operational improvement most marketing teams can make. A 90-minute monthly audit across the seven layers above catches errors before they distort the next month's decisions. The ROI on this 90 minutes is consistently 10× across teams that implement it - not because anything dramatic gets fixed each month, but because the steady accumulation of small fixes prevents the large-scale drift.
Prooflytics surfaces this in the daily briefing as: cross-platform reconciliation is monitored continuously, with gaps outside expected ranges surfaced as alerts. The brief shows when GA4 vs ad-platform conversion counts diverge beyond normal, when UTM source distribution shows unexpected patterns, and when attribution model changes have propagated through reports.
For related operational guidance, see paid media reporting guide and marketing analytics guide.
How Prooflytics supports attribution audits
Prooflytics attribution support joins your full stack: Meta Ads, Google Ads, LinkedIn Ads, TikTok Ads for platform-reported conversion data; GA4 for session-level attribution; HubSpot, Salesforce for CRM-level attribution; Stripe, Shopify for actual revenue data to reconcile against.
The daily briefing monitors cross-platform reconciliation continuously, surfaces UTM consistency issues, and flags when conversion event patterns shift suggesting tracking changes. Monthly attribution audit checklists can be completed against live data rather than as standalone analytical projects.
You can read independent reviews of Prooflytics on G2 and compare it to alternatives in the marketing intelligence category.
Bottom line
- Seven audit layers: UTM naming, tracking pixels, conversion events, attribution model alignment, conversion windows, channel definitions, cross-platform reconciliation.
- The most common attribution problem is UTM naming case inconsistency, costing 5-15% accuracy.
- Monthly audits are the operational standard. Weekly is overkill; quarterly is too slow.
- Attribution model alignment matters more than which model is "correct." Pick one and apply it consistently.
- Cross-platform reconciliation: Meta vs Shopify gap of 15-25% is normal; 40%+ indicates a tracking break.
Book a Prooflytics walkthrough to see continuous attribution monitoring on your own data.
Frequently asked questions
How often should attribution be audited?+
Monthly for most B2B SaaS and DTC teams. Weekly is overkill except for very high-velocity paid teams. Quarterly is too slow - by the time you discover an error in a quarterly audit, two months of decisions have been made on broken data. Monthly is the operational sweet spot.
Who should own attribution audits?+
Marketing operations (RevOps in B2B contexts). The audit requires understanding both the marketing strategy and the technical implementation, so neither pure marketing nor pure engineering owns it well alone. In smaller orgs without dedicated marketing ops, the head of marketing should run the audit personally - it's too important to delegate to someone who doesn't understand the strategic implications.
What's the difference between an attribution audit and an analytics audit?+
Attribution audit focuses on how credit gets assigned across marketing touchpoints - UTM consistency, conversion events, attribution models, cross-platform reconciliation. Analytics audit is broader - includes attribution but also covers data quality, schema, reporting accuracy, and BI tool configuration. Run attribution audits monthly; analytics audits quarterly or semi-annually.
How do I know if my attribution is broken?+
Four leading signals: (1) channel-level ROAS that varies dramatically week-over-week without spend or creative changes, (2) cross-platform numbers (Meta vs Shopify, Google vs GA4) with gaps exceeding 25%, (3) "Other" or "Unknown" channel exceeding 10% of attributed traffic, (4) reported metrics that don't match what your CFO calculates from financial data. Any one of these signals warrants an audit.
What's the most common attribution error?+
UTM naming inconsistency - "Google" and "google" treated as separate sources. Fixing this single error usually improves channel-level accuracy by 5-15%. The second most common: conversion events firing on the wrong page (e.g., on the form page instead of the confirmation page), which inflates conversion counts.
Run marketing on one source of truth
Every source in one brief, so the team stops reconciling exports.
14 days free · no credit card