Conversion Tracking Audit Checklist 2026: 20 Checks to Fix Broken Attribution

User Flows · May 4, 2026 · 12 min read

Nearly half of all ad accounts have conversion tracking problems that silently destroy campaign performance. Double-counted conversions inflate your ROAS. Missing events leave your bidding algorithms flying blind. Incorrect attribution models take credit from the channels actually driving revenue and hand it to the ones that aren't.

The worst part: you won't see it in your dashboards. Everything looks fine until you compare ad platform numbers to your actual bank deposits and realize the gap is six figures wide.

This checklist covers 20 specific checks across GA4, Google Ads, Meta, server-side tracking, and cross-channel attribution. Run through it in about two hours. Most brands find at least three or four problems that are actively distorting their data.

Part 1: GA4 Foundation (Checks 1-5)

GA4 is the single source of truth for most brands. If this layer is broken, every downstream report and optimization is compromised.

1. Verify your GA4 property is receiving data

Open GA4 > Admin > Data Streams. Confirm your web stream shows "Receiving data" with a green indicator. Check the Realtime report while browsing your site in another tab. If the numbers don't move, your measurement code isn't firing.

• Check for duplicate GA4 tags (GTM + hard-coded). This is the number one cause of inflated pageview counts.
• Verify the Measurement ID matches between GTM and your GA4 property.
• Confirm data is flowing for all subdomains if you have them.

2. Audit your key events configuration

Go to Admin > Events and check which events are marked as key events (formerly "conversions"). You should see:

• purchase -- with revenue values populated (not $0)
• begin_checkout -- to measure checkout funnel abandonment
• add_to_cart -- to measure product page effectiveness
• generate_lead or form_submit -- for lead gen sites
• sign_up -- if relevant to your business model

If you only have "purchase" marked as a key event, you're missing the micro-conversions that help Smart Bidding optimize the full funnel.

3. Check for duplicate conversions

This is the silent killer. Navigate to Reports > Monetization > Ecommerce purchases. Export a day's worth of transactions and compare against your Shopify, WooCommerce, or backend order records.

• If GA4 shows 15-20% more transactions than your backend, you have a duplicate firing problem.
• Common cause: the purchase event fires on page load of the confirmation page, and customers refresh or revisit it.
• Fix: use transaction ID deduplication in your dataLayer push. GA4 will automatically ignore duplicate transaction IDs within the same session.

4. Validate revenue accuracy

Compare total revenue in GA4 vs. your payment processor for the same 7-day period. An acceptable variance is 5-10%. If the gap exceeds 15%, investigate:

• Currency mismatches (sending GBP values but GA4 property is set to USD)
• Tax and shipping included in one system but not the other
• Refunds processed in Stripe but not subtracted in GA4
• Missing revenue values (the "value" parameter is null or zero)

5. Confirm cross-domain tracking

If your checkout lives on a different domain (e.g., checkout.yourbrand.com), you need cross-domain tracking configured. Test it: click through your full purchase flow and check the URL bar. You should see a _gl= parameter appended when crossing domains.

Without this, GA4 treats the checkout as a new session from a referral. Your attribution data breaks completely -- direct and referral traffic inflate while paid channels lose credit.

Part 2: Google Ads Tracking (Checks 6-10)

6. Audit imported vs. native conversions

Go to Google Ads > Goals > Conversions > Summary. For each conversion action, check the source column:

• Website (Google tag) -- fires directly from the Google Ads tag
• Import from Google Analytics -- pulls from GA4 key events

If you're importing from GA4 AND running native Google Ads tags for the same conversion, you're double-counting. Pick one source per conversion type and remove the other.

7. Check conversion counting settings

In Google Ads, each conversion action has a counting setting: "Every" or "One." For purchases, use "Every" (each transaction counts). For lead forms, use "One" (multiple form fills from the same person in a session shouldn't count as separate leads).

Misconfigured counting settings are the most common cause of inflated lead counts or understated purchase data.

8. Verify conversion windows

Check the click-through and view-through conversion windows for each action. Defaults are 30-day click-through and 1-day view-through. For high-consideration products (furniture, B2B software), a 30-day click window makes sense. For impulse purchases, 7-14 days is more accurate.

Overly long windows give Google Ads credit for conversions it didn't actually influence. Overly short windows miss legitimate assisted conversions.

9. Test Enhanced Conversions

Enhanced Conversions use hashed customer data (email, phone, address) to improve conversion match rates, especially as cookies continue to degrade. Check:

• Is Enhanced Conversions enabled in your conversion settings?
• Open Google Tag Assistant and complete a test purchase. Verify the tag sends hashed user data alongside the conversion event.
• In the conversion action diagnostics, check the "adjusted conversion rate" -- this shows how much Enhanced Conversions is improving your match rate.

10. Audit offline conversion imports

If you import offline conversions (CRM data, phone calls, in-store visits), verify the upload cadence and match rates. Go to Conversions > Uploads and check:

• When was the last successful upload? If it's been more than a week, your bidding algorithm is optimizing on stale data.
• What's the match rate? Below 50% means most of your offline conversion data is wasted.
• Are upload errors being monitored? Failed uploads fail silently -- no one gets an alert unless you set one up.

Part 3: Meta (Facebook/Instagram) Tracking (Checks 11-14)

11. Verify Pixel and Conversions API setup

Browser-side Pixel tracking alone is unreliable in 2026. Between ad blockers, iOS privacy changes, and cookie restrictions, you're losing 20-40% of conversion data if you only use the Pixel.

Check Events Manager > Data Sources > your Pixel. You should see events arriving from two sources:

• Browser -- the standard Pixel
• Server -- the Conversions API (CAPI)

If you only see Browser events, you need to implement CAPI. Most e-commerce platforms (Shopify, WooCommerce) have native CAPI integrations that take under an hour to set up.

12. Check Event Match Quality score

In Events Manager, each event shows an Event Match Quality (EMQ) score from 1 to 10. This measures how well Meta can match your conversion events to user profiles.

• Good: 7-10. Your data is feeding the algorithm well.
• Needs work: 4-6. You're sending some customer parameters but missing others.
• Poor: 1-3. Meta is guessing on most conversions. Your CPA data is unreliable.

To improve EMQ: send email, phone, first name, last name, city, state, zip, and country with every event. Hash everything before sending (Meta's SDKs handle this automatically).

13. Audit event deduplication

If you're running both Pixel and CAPI (as you should be), you need deduplication to prevent double-counting. Check that each event sent from both sources includes the same event_id parameter.

Go to Events Manager > Test Events. Fire a test purchase. You should see one deduplicated event, not two separate ones. If you see duplicates, your event IDs aren't matching between browser and server.

14. Validate Aggregated Event Measurement priority

Under Events Manager > Aggregated Event Measurement, verify your event priority ranking. Meta limits tracking to 8 events per domain for iOS users. Your highest-value event (usually Purchase) should be ranked #1.

If your priority order is wrong -- say, PageView is ranked above Purchase -- Meta's optimization for iOS traffic will optimize for the wrong action.

Part 4: Server-Side and Consent (Checks 15-17)

15. Audit your consent mode implementation

If you serve users in the EU, UK, California, or any region with privacy regulations, your tags must respect consent signals. Check your Google Consent Mode v2 implementation:

• Before consent: tags should fire in "denied" mode (no cookies, limited data)
• After consent: tags should update to "granted" mode and begin full tracking
• Test by declining cookies on your site and verifying that conversion tags don't fire full events

Non-compliant tracking isn't just a legal risk. Google and Meta will increasingly penalize accounts with poor consent signals by limiting their access to bidding features and audience data.

16. Check server-side GTM container health

If you're using server-side Google Tag Manager (sGTM), verify:

• The container is running on a first-party domain (e.g., track.yourbrand.com), not a third-party endpoint
• Cloud Run / App Engine logs show no errors in the last 7 days
• The server container is forwarding events to all required endpoints (GA4, Google Ads, Meta CAPI, etc.)
• Response times are under 200ms -- slow server-side tags delay your conversion data pipeline

17. Validate UTM parameter consistency

Pull your GA4 traffic acquisition report and filter by source/medium. Look for fragmentation:

• Does "google / cpc" appear alongside "Google / CPC" or "google / paid"? That's broken UTM consistency.
• Are email campaigns tagged consistently? "klaviyo / email" vs "Klaviyo / Email" vs "email / newsletter" splits your data.
• Check your paid social ads -- every ad should have utm_source, utm_medium, utm_campaign, and utm_content at minimum.

Build a UTM naming convention document and enforce it. Inconsistent UTMs make cross-channel attribution impossible.

Part 5: Cross-Channel Attribution (Checks 18-20)

18. Compare platform-reported conversions to GA4

Pull the same date range from Google Ads, Meta Ads, and GA4. Build a simple comparison table:

Source	Conversions Reported	Revenue Reported
Google Ads	--	--
Meta Ads	--	--
GA4 (all channels)	--	--
Backend / Shopify	--	--

If the sum of Google Ads + Meta conversions exceeds GA4 total by more than 20%, you have an overlap problem. Both platforms are claiming credit for the same conversions. This is normal to some degree (each platform uses its own attribution), but a large gap means your combined ROAS reporting is misleading.

19. Audit your attribution model settings

In GA4, go to Admin > Attribution Settings. Check:

• Reporting attribution model: GA4 now defaults to data-driven attribution. If you switched to last-click at some point, consider whether that's still serving you. Data-driven gives a more complete picture for multi-touch journeys.
• Lookback window: 30 days for acquisition, 90 days for all other conversions is the default. Adjust based on your sales cycle.
• Google Ads attribution: In Google Ads, check Tools > Attribution > Model comparison. If you're still on last-click in Google Ads while GA4 uses data-driven, your numbers will never match.

20. Run a conversion path analysis

The final check: go to GA4 > Advertising > Conversion paths. This shows the actual multi-touch journeys your customers take before converting.

• How many touchpoints does the average conversion require? If it's 3+, last-click attribution is lying to you about which channels matter.
• Which channels appear most often in the first position (awareness) vs. last position (closing)?
• Are any channels showing up as assist-heavy but getting zero credit in your reports? Those channels are probably underfunded.

What to Do After the Audit

You've now got a list of problems. Here's how to prioritize:

1. Fix duplicate conversions first. Double-counted conversions inflate your ROAS and cause your bidding algorithms to underbid (they think performance is better than it actually is). This is the highest-impact fix.
2. Implement server-side tracking. If you're only running browser-side tags, you're losing 20-40% of data. CAPI for Meta and Enhanced Conversions for Google Ads are table stakes in 2026.
3. Align attribution models. Pick one model (data-driven is the safest default) and configure it consistently across GA4 and your ad platforms. Mismatched models make cross-channel reporting meaningless.
4. Standardize UTMs. Create a naming convention, build it into a shared template, and audit compliance monthly.
5. Set up monitoring. Don't wait for the next quarterly audit to catch problems. Set up anomaly alerts in GA4 for sudden drops in conversion events -- these usually indicate a broken tag.

Related Audit Checklists

• Google Ads Audit Checklist -- 20 checks for your paid search account
• Facebook Ads Audit Checklist -- 20 checks for your Meta ad account
• Marketing Audit Template -- the complete A-F grading framework
• Ecommerce Marketing Audit -- full-stack audit for online stores
• PPC Audit Checklist -- broader paid advertising checks
• Digital Marketing Audit Template -- channel-by-channel assessment guide
• Marketing Channels Checklist -- 40-point checklist across all five channels
• SEO Audit Checklist -- 25 checks for organic search