One product, four URLs, zero ranking power

One product, four URLs, zero ranking power

A SaaS comparison site had strong content but pages that 'should have ranked' sat at position 30+. Backlinks weren't the issue — they had plenty.

The clue came from a link audit. The same comparison page was linked four ways: ?ref=home, ?ref=email, #tab=pricing, and clean. Each variant rendered different client-side state.

The deeper problem: their framework wrote the canonical tag client-side, after reading the current URL. So ?ref=email self-canonicalized to ?ref=email. Google saw four 'canonical' pages, each claiming to be the original, splitting link equity four ways.

Client-rendered canonical tags inherit whatever junk is in the address bar. The crawler trusts what's in the DOM at render time, and the DOM said 'I am the canonical.'

Fix: canonical hardcoded server-side to the clean URL, stripped of all params and fragments, identical across every variant.

Result: consolidated signals lifted the page from position 31 to 7 in 10 weeks. The other three variants quietly de-indexed, and clicks tripled.