Home/Blog/GEO & SEO
GEO & SEO

Rank 1 of 10 brands, and still not a single JSON-LD block

61.7 out of 100 points, 16.3 points ahead of the competitive average. And not a single JSON-LD block on the category pages. Checked three times independently, three times the same result.

Felix Zeh·August 17, 2026·6 min read·AI-assisted
LEADING BRAND61.7COMPETITOR AVG.45.4100 POINTSSTRUCTURED DATA (JSON-LD) PER BRAND0FILLED = JSON-LD CONFIRMED (5) · DASHED = LEADING BRAND, NO JSON-LD · EMPTY = NOT DISCLOSED
What you'll take from this article
  • Why technical structure and content authority are two separate levers
  • What a GEO competitive scan across ten brands actually measures
  • Why a lead with no technical foundation stays exposed
  • Which structured data actually matters first
61.7of 100 points, rank 1 of 10 brands
+16.3points ahead of the average (45.4)
0JSON-LD blocks on category pages
3,305pages checked in the full crawl

Alongside the conversation analysis, we ran a technical scan of ten brands in the same category: crawled each one's German-language category page, checked for structured data, and scored them along four axes. The result contradicts the standard SEO reflex.

How does that add up?

The brand leading the field scores 61.7 out of 100 points. The competitive average sits at 45.4. A 16.3-point lead, which is anything but close.

That same brand runs no JSON-LD on its category pages. All that's present is breadcrumb navigation using the older Microdata format. Nothing else.

Checked three timesA full crawl across 3,305 pages plus two later spot checks over three weeks all arrive at the same result. A measurement error is largely ruled out.

Where does the lead come from, then?

From two sources, neither of which has anything to do with technology.

Externally confirmed authority. A very strong score from an established comparison test, trade press coverage of the professional models, mentions across advice portals. Evidence that doesn't come from the company itself.

Content differentiation. The pages visibly distinguish between different use situations instead of just listing product features. That makes them fit a wider range of different questions.

Both together carry the top position. The technical markup doesn't.

What are competitors doing?

Exactly what the market leader isn't. Several competitors in the same field, including strong runners-up, run JSON-LD on their category pages on top of everything else, and are systematically building out their machine-readability.

That's the real story in this finding. The leader's advantage rests on authority and content, factors others can catch up on over time. Meanwhile the technical foundation, which would be comparatively cheap to retrofit, sits idle. If a runner-up closes the content gap while already ahead technically, the position tips.

What separates GEO from classic SEO?

A pure SEO audit would likely have flagged this case as a red alert. Missing JSON-LD has counted as baseline hygiene there for years.

In the GEO score, content and reputational strength are enough to stay ahead anyway. A language model doesn't primarily judge a page by its markup, it judges whether the page reliably answers a question and whether independent sources back that up.

That shouldn't lead you to conclude structured data is pointless, though. The right reading is: these are two independent levers. Stand on only one, and you're standing on shaky ground.

Where to start technically

If you're not using structured data yet, or only partially, here's a sensible order.

  1. Check what's actually there first

    Don't ask whether it was built, go look. Google's Rich Results Test or a crawler will show you what's genuinely delivered per page type. It's often quite different from what people expect.

  2. Start with the Organization schema

    Company name, founding year, logo and references to Wikipedia or Wikidata via the sameAs field. That clarifies who you even are, and it's the foundation for placement in the knowledge graph. Effort: low.

  3. Mark up product pages with Product and Offer

    Including price or price range, availability and warranty length. Exactly the details missing at the purchase phase, as the conversation analysis showed.

  4. Mark FAQ sections as FAQPage

    Wherever you're already answering questions, the markup makes the question-answer pairs unambiguously recognizable. Low effort on content that already exists.

  5. Clean up duplicate page titles

    In the case we analyzed, two different pages carried the exact same title. That's an hour's fix and removes an unnecessary ambiguity.

  6. Don't forget authority at the end

    Test results, certificates and awards belong on your own domain, citable, with a source and date. That's the lever that carried the top position in the case we analyzed.

Frequently asked questions

Do I really need JSON-LD if the market leader gets by without it?

The market leader gets by because it built up authority over years that compensates for it. Very few brands start from that position. For everyone else, technical markup is the cheaper, faster lever, because it ships in days, while authority takes years.

Is Microdata good enough instead of JSON-LD?

Technically, Google understands both formats. But JSON-LD is Google's recommended format, can be maintained independently of the visible markup, and is considerably less error-prone when things change. For new implementations, there's hardly a reason to choose Microdata.

What exactly does a GEO composite score measure?

In our case, four axes: clarity as an entity and knowledge-panel readiness, content substance against E-E-A-T criteria, fitness for conversational answers, and presence across different platforms. Part of it is scored by a language model, which is why such scores are only comparable within the same evaluation model.

How often should you repeat a competitive scan like this?

Once or twice a year is usually enough. The deterministic parts, like structured data and crawlability, rarely change abruptly. For the model-scored axes, staying on the same evaluation model matters more than measuring frequently.

How this was measured

Technical GEO crawl of ten brands, each brand's German-language category page, scored along four axes by a language model. Structured data checked deterministically: full crawl across 3,305 pages on July 13, 2026, plus spot checks on July 26 and August 5, 2026, all with an identical result (Microdata BreadcrumbList present, no JSON-LD). Title, description and word count of the target page were unchanged across all three surveys.

Transparency note · AI-generated content

A note on this article: research, analysis and writing were produced with the support of AI systems (the Ex Tenebris agent team) and reviewed editorially by Felix Zeh. Every figure cited comes from a real LUX/GEO analysis; the analyzed brand is anonymized to protect the client relationship.

FZ
Felix Zeh
Founder, Ex Tenebris · Stuttgart, Germany

Has spent over a decade at the intersection of marketing, data and analytics, and uses LUX to measure how brands show up in the answers of ChatGPT, Claude, Gemini and Perplexity. Questions or pushback on this article? kontakt@extenebris.de

Get in touch

How visible is your brand in the AI answer today?

The methodology behind this article is the same one we use for every assessment we run, tailored to your category, your competitors and your need states.