$180K/month. Global campaign.
AMER · EMEA · APAC.

$300K a month in spend. Optimizing for the wrong thing.

The Client was running $300,000+ per month in paid advertising across Google, LinkedIn, and Bing. The numbers looked decent on the surface. But underneath, the account had a fundamental problem that was driving every decision in the wrong direction.

Google was overcounting conversions by 3x. The platform was counting both trial signups and purchases as conversion actions simultaneously — making the algorithm believe campaigns were performing far better than they actually were. The result: Google's AI was scaling spend based on false confidence, pouring budget into campaigns that weren't driving real revenue. We were spending $300K a month to optimize for a metric that didn't exist.

Beyond the tracking issue, the account had three other structural problems: inefficient spend distribution, no coherent MOF engagement strategy for trial users who hadn't converted, and a lead-to-close gap between AMER and EMEA/APAC regions that was leaving revenue on the table globally.

The algorithm was working perfectly. Against you.

When Google's AI is fed bad conversion data, it doesn't slow down — it accelerates. It takes the inflated signal and doubles down, allocating more budget to whatever it thinks is working. In this case, that meant scaling spend on trial signups that never converted to revenue while starving the campaigns that were actually closing deals.

Fix the data first. Then scale what's real.

The first and most important move was switching to offline conversions. Instead of letting Google count trial signups as conversions, we connected actual purchase data from Salesforce — uploading strategic and transactional purchases directly into Google Ads. This cut measured conversions by two-thirds overnight. Budget dropped from $300K to $200K per month because the algorithm stopped scaling campaigns built on false signals.

That sounds like a loss. It wasn't. With accurate data, the algorithm could finally optimize for what actually mattered — real revenue events. The result was a 40% increase in ROAS and 30% revenue growth. We spent less and made more because we stopped paying for phantom performance.

Spent less. Made more. And the LTV compounds it further.

The numbers tell a clean story. $180K in monthly revenue from $200K in spend looks like a thin margin on paper — until you factor in the average customer LTV of 3.5 years. Every customer acquired through this system isn't worth their first month's revenue. They're worth 42 months. That changes the entire math of what a "good" CAC looks like.