The Permutation Trap: Why Your Shopping Titles Are Destroying Impressions
Ruslan Galba
Google Ads + AI
We stopped permuting Google Shopping titles 18 months ago. Not because it was hard. Because it was destroying our impressions.
Here's the trap nobody warns you about.
The 70-character rule changes everything.
Google displays only the first 70 characters in Shopping ads. Users see just 20-30 without hovering. Words after 70 get reduced algorithmic weight. This is Zone 1—and permutations randomize it completely.
A Search Engine Land study (136 products, 6 weeks) found keyword-rich titles drive +18% CTR. Exact query match in the title? +88% CTR increase. But permutations scatter your keywords across Zone 1 randomly. No strategic front-loading. Just algorithmic chaos.
The Permutation Trap has 4 failure modes:
First, near-duplicates trigger Google's duplicate content detection. Second, random word order breaks Zone 1 optimization. Both kill impressions before you even know there's a problem.
Third, synonym rotation creates awkward phrasing. Fourth—and this is the one that really hurts—all variants cluster around the same keyword pocket. You're not expanding reach. You're cannibalizing yourself with slightly different wrappers.
What we do instead:
Each title gets deliberately crafted for ONE intent. Zone 1 (first 70 chars) gets the primary keyword plus key differentiator. Zone 2 (remaining 80) handles secondary attributes. No permutations. No algorithmic variations. Deliberate.
The shift took 3 weeks to implement across 2,000+ SKUs. But we stopped fighting Google's duplicate detection. Stopped randomizing our Zone 1. Started showing up for distinct search intents. Impressions recovered in 6 weeks.
The permutation trap is tempting because it scales fast. But fast doesn't matter if Google penalizes you for near-duplicates and you're invisible for the queries that convert.
Deliberate beats automated. Every time.
Get the Google Ads Playbook
The same playbook we use across $10M+/mo in managed spend. Covers account architecture, feed optimization, and testing frameworks.