All three clients are anonymized. Every figure below is platform reported, and a redacted version of the underlying reporting is available on request. Each case includes the part that did not work, because a case study without one is an advertisement.
Shopify · United States · 16 months
Already profitable and stuck at a spend level nobody wanted to move off. I cut the geographies that never paid for themselves, rebuilt Google as full funnel, and opened Meta and Bing. Spend more than tripled and efficiency went up instead of down. The figures compare December 2024, Google only, with April 2026, Google and Meta combined.
223 → 1,278
Monthly purchases
$2,021 → $6,989
Monthly spend
Read the full case
Lead generation · United States · in progress
The dashboard reported 981 conversions and not one of them was a lead. Conversion actions were firing on pageviews and impressions while the account paid for traffic from other countries. I rebuilt the measurement first, then the account around it.
981
False conversions cleared
11
Tracked contact requests
314 → 194
Keywords on broad
Read the full case
Nineteen locations · Mexico · 24 months
Nineteen branches on a budget that would not cover one campaign in a national account. Measured the way an e-commerce account gets measured, it produces nothing. Measured by what a local buyer actually does, it produces nearly five hundred actions a month, and it has done so for two years.
Read the full case