Reconciling a GA4 Collapse With CRM Growth
Verdant Analytics sells subscription software. After implementing Consent Mode v2 in March, GA4 reports organic conversions down 31% month over month, while GSC shows organic clicks up 6% and the CRM shows 14% more inbound demo requests from organic landing pages. The CMO sees a dashboard with a single 'organic conversions' tile and asks whether SEO should be defunded. The marketing ops lead changed the GA4 conversion event in February from 'form_submit' to 'qualified_lead,' and the attribution model from last-click to data-driven. Sales cycle is 90 days; 40% of demo requests come from branded search. The team has GSC, GA4, HubSpot CRM, and a Looker Studio dashboard built by an intern. The dashboard has no annotations and compares March to February.
1. Reconcile the three conflicting signals. What is the most likely explanation, and what would you check first?
Read the worked answer
These numbers are not necessarily contradictory. GA4 organic conversions fell 31% after Consent Mode v2; consent banners reduce observable events, especially in the EU, and GA4 models some but not all. The event definition changed from 'form_submit' to 'qualified_lead' in February, which raises the bar for what counts and can make the conversion rate look worse. The attribution model also changed from last-click to data-driven, redistributing credit away from organic landing pages. Meanwhile GSC clicks up 6% and CRM organic demo requests up 14% point to stable or improving demand.
Check three things first: GA4 organic sessions versus GSC clicks (if sessions fell but clicks rose, tracking is the likely cause); consent rate and event counts by region; CRM original source and landing page for the +14%. The 40% branded share means some 'organic' demand is influenced by sales and ads, so separate brand and non-brand. Do not accept the CMO's framing that SEO collapsed-the measurement system changed, not necessarily the channel.
2. Design a reporting framework that the CMO can trust. Which KPIs belong in each layer, and how do you handle the 90-day lag and brand/non-brand split?
Read the worked answer
Build a three-layer scorecard. Demand: GSC non-brand impressions and clicks, branded clicks separately, and query-level rankings for priority clusters. Engagement: GA4 organic sessions, engagement rate, product-page views, and demo-request rate, segmented by landing page and brand/non-brand. Outcomes: HubSpot CRM contacts, MQLs, SQLs, and closed-won with original source = organic, plus influenced pipeline. Use a 90-day lag and rolling three-month windows because of the sales cycle.
Annotate every dashboard change: consent mode, event definition, attribution model, campaigns. Replace the single 'organic conversions' tile with a funnel and a confidence note. Add self-reported attribution and, if volume allows, a holdout or geo test for brand campaigns. Define one source of truth per metric: GSC for search demand, GA4 for on-site behavior, CRM for revenue. Reconcile monthly with a short variance note. The goal is not to make the numbers agree; it is to make the disagreement explainable.
3. How would you communicate this to the CMO without overclaiming, and what decision criteria would you propose before any budget cut?
Read the worked answer
Say: 'Observable GA4 conversions fell because we changed consent, event definition, and attribution-not because organic demand fell. In the same period, GSC clicks rose 6% and CRM organic demo requests rose 14%. The honest conclusion is that we have a measurement gap, not an SEO failure.' Then propose a 60-day review with pre-agreed decision criteria: if non-brand organic clicks and organic-sourced SQLs both decline for two consecutive months, revisit budget. Until then, do not defund; instead fund a tracking fix and a CRM-SEO reporting loop. Avoid overclaiming causality from the +14%-it could be sales outbound or brand. Present a range and the assumptions. Offer a small test: allocate 10% of budget to a bottom-funnel content sprint and measure SQLs. This gives the CMO a decision framework rather than a defense.