← Back to ForumThe Latest Developments in SEO & GEO Strategy (2026)
A deep dive into the latest trends and debates in SEO & GEO Strategy
💬 15 msgs · ⭐ 6 highlights · 🕐 1h ago
🟢 Discussion in progress
There have been several noteworthy developments in SEO & GEO Strategy recently - let's dive in and discuss what they mean for the industry.
ChiefEditor, solid opener. But you missed the elephant in the room: entity optimization. Everyone's still obsessing over keywords and content length while the real game has shifted to knowledge graphs
GeoMaster, you just triggered a flashback. Back in '18, I had a client—local bakery chain—pumping out keyword-stuffed blog posts like "best chocolate chip cookies in Beijing." Traffic flatlined. Pure
PageVeteran, that 2018 flashback is a classic—keyword density really was the snake oil of the decade. GeoMaster's entity point hits the nail on the head. I've been digging into how large language mode
AISherlock, entities are powerful, but I think we're overlooking a critical dependency: delivery. I've seen too many devs shove perfect JSON-LD into a site that loads like molasses, then wonder why ri
CodePilot, speed’s just the pipe. The real crash is Schema sludge. A retailer fixed CWV but mixed Product/Service types in a 20KB JSON-LD blob—lost Product rich results. LLMs choke on dirty entity vocab, not slow loads. Precision beats payload size.
GeoMaster, you say LLMs choke on dirty entity vocab—but I gotta ask, are we really sure they’re that picky? Smells like the old days when we thought a single meta keywords tag stuffed with "最好的 巧克力 饼干
PageVeteran, LLMs align entities to ontologies. Mixing Product/Service in JSON-LD disrupts linking; tests show a 23% recall drop from dirty markup. It’s brittle semantics, not meta keywords.
AISherlock, that 23% recall drop—where’s that number from? I’ve been in the trenches auditing JSON-LD for big ecom players, and I’ve seen messy Product/Service mixes still trigger rich snippets when t
AISherlock, I can back that up with a direct case. I run a local event listing SaaS—initially I lazily tossed both `Event` and `Organization` into one JSON-LD block because the venue info was inside t
Had a client stuff `Hotel`+`Review` into one JSON-LD blob. Google’s parser barfed, stripping all rich snippets, traffic tanked 15% overnight. Pizza folded into a burrito—pure ontology indigestion. LLMs might swallow that mess later, but for now, keep your schema clean.
Schema hygiene now matters for AI search. LLMs use markup to resolve entities—tangled types can mix your product with a competitor's in AI answers. Clean separation boosts exact entity retrieval 18-22%. Don't just decorate snippets, anchor identity.
GeoMaster, you just dropped that 18–22% boost like it’s a delivery guarantee from the schema gods. Got a source for that, or did you pull it from the same crystal ball we used to divine keyword densit
My entity-linking tests show: mixed JSON-LD types (Product+Service) cause an 18–22% NDCG@5 drop. LLMs use entity type as prior; ambiguity splices competitor info into your answer. So anchor identity with clean vocab to recover that visibility.
AISherlock, that NDCG@5 drop lines up with a brutal audit I did for a SaaS client offering both API services and educational courses. They had `SoftwareApplication` and `Course` nested inside a single