Marketing Analytics Checklist

List every active source feeding the reporting stack: GA4, GSC, the MAP (HubSpot/Marketo/Pardot), CRM, ad platforms (Google Ads, Meta, LinkedIn), and any CDP like Segment. Confirm each is sending to the right property and that no orphaned pixels from a sunset campaign are still firing.

Use GTM Preview mode and the Tag Assistant to confirm tags fire on the expected triggers. Watch for duplicates — a Meta Pixel firing twice is the most common reason CPA looks half what it should be.

Walk every key conversion (demo request, signup, purchase) end-to-end and confirm the event fires only on actual submit — not on email-blur or page-view. Mis-mapped events are the leading cause of inflated conversion counts and distorted budget allocation.

For each broken event, file a Jira ticket with the page URL, expected trigger, observed behavior, and a GTM Preview screenshot. Block the monthly report until fixes are deployed and re-validated.

Pull the source/medium/campaign report in GA4 and scan for convention drift — capitalized vs lowercase mediums, freeform campaign names, missing utm_content. Drift breaks period-over-period comparison. Update the link-builder template if you find a gap.

Compare GA4 form-submit counts against MAP form fills and CRM lead inserts for the period. Variance over 5% usually means a routing rule, dedupe rule, or hidden-field mapping is dropping records. Document the variance source before reporting.

Confirm GA4 internal-traffic filter excludes office IPs and known QA tools. Check for sudden spikes from a single referrer or country — bot traffic skews engagement metrics and can mask real performance shifts.

Pull CPC, CPA, ROAS, and MQL-to-SQL conversion rate by channel. Compare to the prior period and the trailing-90-day baseline. Flag channels where CPA is 20%+ above baseline for the paid-media review.

In GA4 Attribution, compare data-driven against last-click for the period. Large gaps point to channels (typically display, paid social) that are over- or under-credited under last-click. Note implications for budget allocation in the report.

Surface the three findings that should drive a decision next month — channel reallocations, content refresh candidates, funnel-stage drop-off fixes. An insight without a recommended action is a metric, not an insight.

Update the Looker Studio or Tableau dashboard with the closed period. Sanity-check pipeline, MQL volume, MQL-to-SQL rate, CAC, and payback period against the source-of-truth in CRM before the dashboard goes to the CMO.

One row per channel: spend, leads, MQLs, opps, CPA, ROAS, period-over-period delta. Color-code deltas; add a one-line narrative for any channel moving more than 20%.

Lead with the three insights and recommended actions, then performance against OKRs, then channel detail. Avoid leading with a wall of charts — the executive audience reads the first paragraph and skims the rest.

Walk the director through the draft before distribution. Capture edits, narrative changes, and any data points the director wants pulled or added before the WBR.

Incorporate the director's edits, re-validate any changed numbers against source-of-truth, and route back for final sign-off before distribution.

Send to the CMO, sales leadership, and the exec team with a 3-bullet summary in the email body. Post the dashboard link in the #marketing-reporting Slack channel for the broader team.

Load the site from an EU IP and confirm OneTrust (or your CMP) blocks GA4, Meta Pixel, and LinkedIn Insight Tag until consent is granted. Pre-consent firing is a recurring GDPR finding and contradicts the banner's promise to users.

Confirm the MAP suppression list synced to every sending platform — including ad-platform custom audiences. CAN-SPAM requires honoring unsubscribes within 10 business days; CCPA opt-outs require removing addresses from data-broker audience uploads, not just from email.

Verify GA4 user-data retention matches the privacy policy commitment (commonly 14 months). Mismatch between policy and platform is a documented GDPR finding even when no breach occurs.

For each data source, name the owner, the system of record, and the downstream consumers. Without this, a deletion request or schema change becomes an incident hunt across teams.

Final sign-off from marketing ops and the privacy reviewer (legal or DPO contact). Note any open items that roll into next month's run.