Anonymized travel-tech marketplace
100+ countries • 60,000+ point-to-point routes
We cut 40% of the ad budget. Revenue and margin never noticed.
Monthly ad spend came down 40%. Revenue held its year-over-year growth and average gross margin went up 9 points. That gap is the whole story: roughly 40% of this budget was buying nothing. Finding it meant scoring all 60,000+ routes on backend profit, PPC performance and SEO demand, then rebuilding a 63,846-campaign account into 127 around those scores.
The Challenge
When I first logged into this Google Ads account, I saw archaeological layers of six years of continuous campaign creation, built on top of each other without a single layer being removed.
For context, a large e-commerce account might have 50 to 100 campaigns. A complex multi-country operation might push to 500 or 1,000. This was 63,846.
Three Critical Problems
1. Algorithmic Concentration
Google's algorithm naturally pushes 80% of spend to the top 20% of performers within any campaign. With 60,000+ routes mixed together, the algorithm decided where money went, not the business. High-potential routes never got enough budget to prove themselves. Strategic business priorities were ignored.
2. No Connection Between Ad Spend and Business Value
The algorithm concentrated spend on proven winners and starved new routes before they could gather meaningful data. New high-potential routes needed conversions to prove themselves, but low initial spend meant no conversions were detected. The algorithm then stopped spending on these routes entirely, making it impossible to discover which new routes could become profitable. Meanwhile, there was no system connecting backend profitability data to advertising decisions. High-margin routes in strategic markets got the same algorithmic treatment as low-margin routes in saturated markets.
3. Google's "Other Search Terms" Black Box
Nearly half of all paid clicks went to searches Google would not show us. Those hidden searches converted worse on both metrics that mattered, and the aggregate was the only view we had. Some campaigns inside "other terms" almost certainly outperformed the known ones, others were pure waste, but with thousands of campaigns there was no way to tell which was which.
Cost per acquisition on those hidden searches ran 24% above the known terms.
The Algorithmic Concentration Problem
Google's algorithm naturally applies the 80/20 rule. You can't fight it-but you can control WHERE it happens.
💡 The Breakthrough
You can't stop the 80/20 rule-it's fundamental to how machine learning works. But by segmenting routes into separate campaigns, you control which 20% gets the 80%. The algorithm's natural behavior now works FOR your business strategy, not against it.
The Solution
The answer wasn't in Google Ads data alone. It required building a custom scoring system that blended multiple data sources to give us a holistic view of route performance and potential.
Route Segmentation Framework
Blending backend profit data, PPC performance, and SEO search volume to classify all 60,000+ routes into strategic segments
Cash Cows
Proven winners • High revenue & margin • Consistent POAS >150%
Maximize investment • Never miss traffic opportunities
High Potentials
Strong margins • Good signals • Limited data but promising
Scaling candidates • Tomorrow's Cash Cows
Low Score
Underperformers • Low margins or poor POAS • Limited demand
Minimal maintenance • Test occasionally
Sleepers
Never advertised • Zero historical data • Unknown potential
Controlled exploration • Discovery pipeline
Routes move between segments based on performance. A Sleeper that performs well gets promoted to High Potential. A High Potential that proves out becomes a Cash Cow. The system continuously adapts to align budget with actual business value.
Transformation Timeline
From unmanageable chaos to strategic control in 4 phases
Immediate Cleanup
Paused 11,000 poor-performing campaigns. Killed phrase match. Split brand campaigns by region. Implemented regional tPOAS targets.
Build Route Scoring
Blended backend profit data (BigQuery), Google Ads performance, and SEO search volume into proprietary scoring system.
Strategic Segmentation
Classified all routes into Cash Cows, High Potentials, Low Score, and Sleepers. Separated into dedicated campaigns.
Domain Migration Launch
Complete rebuild during domain change. Parallel run strategy. New structure: 127 campaigns vs. 63,846 old campaigns.
The Transformation in Scale
From unmanageable chaos to strategic control
Less complexity = cleaner signals = better optimization
The Results
The rebuild launched alongside a full domain migration, so the first week looked alarming. Spend ran well above the old baseline while the algorithm relearned everything. By week two it found its footing. By week four we were past the old performance levels, and the budget kept coming down from there.
The number that matters is the one that did not move. Monthly investment fell by 40% and the business carried on growing year over year at the same rate, on better margin. Nothing was traded away to get the saving, which is the clearest evidence that the removed spend had never been buying anything.
The headline result. Two fifths of the monthly budget came out and stayed out, through a period that also included a full domain migration.
Year-over-year growth continued at the same rate on 40% less spend. No drop during the transition, no recovery period to pay back afterwards.
Margin improved rather than holding flat, because the budget that remained sat on high-margin routes instead of wherever the algorithm had drifted.
Profit on ad spend is the arithmetic of the three above. Roughly the same profit, produced on 40% less spend, at slightly better margin.
Figures are stated as change against the prior period. Absolute spend, revenue and margin are withheld at the client's request.
Key Takeaways
Multi-Source Data Beats Single-Source Data
Google Ads data alone is insufficient at scale. We blended backend profit data (BigQuery), PPC performance, and SEO search volume for holistic route scoring. High POAS but low margin? Not a Cash Cow. Low data but high search volume and good margin? High Potential.
Work WITH the Algorithm, Not Against It
Google's 80/20 concentration is inevitable. You can't fight it, but you can control WHERE it happens. By segmenting routes into separate campaigns (Cash Cows, High Potentials, Low Score, Sleepers), the algorithm's natural behavior now aligns with business priorities.
Most of a Bloated Budget Is Buying Nothing
17M keywords sounds impressive, but 99.85% never converted. If 40% of the spend can come out with no revenue impact, that spend was never producing revenue. It was producing noise that stopped the algorithm from optimizing. Cutting it is not a sacrifice, it is a correction.
Algorithms Optimize for the Past. Strategy is About the Future.
The route scoring system worked because it combined algorithmic optimization (within segments) with human judgment (defining the segments based on business strategy). Don't confuse the two.
Ready to Transform Your PPC Strategy?
Whether you're drowning in complexity or just want to align your ad spend with business value, let's talk about how multi-source data scoring can transform your performance.