Back to Blog

IA: one page per decision, not per keyword

TL;DR: Key Takeaways

A landing page for every synonym creates the same problem in the AI layer as in classic search: the model cannot tell which URL is the canonical statement.

  • Better: one page per business decision.
  • Swiss multilingualism makes this sharper.
  • Programmatic city pages with no unique fact (90 percent identical, different place name) are more dangerous on a new domain than on a domain with years of trust.

A landing page for every synonym creates the same problem in the AI layer as in classic search: the model cannot tell which URL is the canonical statement.

Better: one page per business decision. What do we offer? Where? For whom not? How does an enquiry work? What proof exists? Internally link the cluster instead of keeping five near-duplicates.

Swiss multilingualism makes this sharper. DE/FR/IT/EN need parallel entities, not machine copies with a different slug. hreflang without content parity is cosmetics.

Programmatic city pages with no unique fact (90 percent identical, different place name) are more dangerous on a new domain than on a domain with years of trust. SOURCE/01 by GlasBox (Search) scores unique-content share, not URL count.

---

Next step

GlasBox measures classic SEO and AI visibility separately. No ranking or mention guarantee. The free short check stays thin on purpose.

Request an introductory call

Frequently Asked Questions

Why do synonym landings fail in the AI layer?
The model cannot tell which URL is the canonical statement. Better: one page per business decision.
Is hreflang enough without content parity?
No. DE/FR/IT/EN need parallel entities, not machine copies with a different slug. hreflang without parity is cosmetics.
Are programmatic city pages safe?
More dangerous on a new domain than on a domain with years of trust. Unique-content share is scored, not URL count.