Marketing Attribution & Campaign Measurement

frankdenblankenpt.nl offered genuinely strong personal training and coaching, but the website had no real conversion strategy behind it — no dedicated landing pages, no booking flow built for conversion, and no way to see where visitors dropped off before becoming a lead, or which of Frank’s Google Ads, Meta Ads and organic traffic was actually behind each enquiry.

Rebuilding the foundation first

Before any tracking could be meaningful, the site itself needed structure: a rebuild in WordPress using Elementor and Crocoblock introduced dedicated landing pages, proper booking forms, and a CTA-focused homepage with a specific focus on mobile speed.

Tagging it properly before trusting any number

GTM was rebuilt around the new site structure — every booking step, form interaction and outbound click fired its own tagged event, with consistent UTM parameters enforced across every Google Ads, Meta Ads and email link, so the data feeding GA4 could actually be trusted downstream.

Attribution told one story; a holdout test told another

GA4 and SQL-joined session data built an assisted-vs-last-click attribution model showing SEO was quietly contributing to bookings well before the last click — but attribution still assumes every assisted conversion was incremental. To check that assumption, I paused Google Ads in Frank’s area for two weeks against a matched prior-year baseline: bookings dropped by less than the ad platform’s own reported conversions, meaning a real share of “paid” bookings would have happened anyway through organic and direct. That gap is what attribution alone can’t see.

A Power BI report Frank could actually read

The incrementality-adjusted numbers fed a Power BI dashboard built around one question Frank asked directly: which channel is actually worth paying for. Search, Google Ads, Meta Ads and email sat side by side against enquiries and bookings — adjusted for what was truly incremental, not just attributed — so budget decisions could be made on what drove business rather than on reach.

Results

• Traffic increased by roughly 300%, peaking at 5,000+ visits in October 2024
• Holdout test showed a meaningful share of paid-attributed bookings were not incremental, prompting a Google Ads budget cut with no measurable drop in real bookings
• Funnel insights directly informed copy and CTA improvements at the pages losing the most leads
• SEO’s real, previously invisible contribution made visible for the first time

Reflections

Attribution and incrementality gave two different answers to “what’s working,” and the gap between them was the actual finding — not the dashboard itself, but the willingness to test whether the dashboard’s own numbers could be trusted at face value. Walking Frank through that in plain terms, not a wall of GA4 screenshots, was what changed his ad spend the following month.