Caesarstone Israel — Full-Site SEO / GEO / AI Audit
Scope: www.caesarstone.co.il + studio.caesarstone.co.il · 1,168 URLs captured live on 2026-10-01 · Prepared by the MaximoSEO team
1 · Executive Summary
URLs audited
Empty pages (200, no content)
"Not found" pages in sitemap
Studio pages noindexed
Missing canonical
Duplicate titles
Missing meta description
Images missing alt
Critical — the entire studio subdomain is blocked from indexing.
All content pages on studio.caesarstone.co.il carry
<meta name="robots" content="noindex, nofollow">, and Google returns zero indexed pages for
site:studio.caesarstone.co.il. The A&D professional platform (catalog, projects, kits, guides) is
invisible to Google and AI engines until this is removed.
Critical — 445 sitemap URLs serve empty pages.
The studio “3D files” section (DWG / DXF / GSM download endpoints) returns HTTP 200 with an empty document
(106 bytes). They are listed in the sitemap, cannot render (Lighthouse: “The page did not paint any content”),
and waste crawl budget.
Performance on mobile templates is catastrophic.
Sampled key pages score 4–42 / 100 (Performance, mobile, simulated throttling): homepage LCP 19.6 s,
catalog LCP 36.1 s, collections LCP 34.7 s, TBT up to 44 s. Studio pages score 32–35 with SEO 50–58.
Backlink profile is healthy but leaking.
DR 46 · 16,859 live backlinks · 1,122 live referring domains. Legacy .aspx URLs still attract links
(203 refdomains to one), two legacy pages now 404 with 30–37 refdomains each, and a WordPress
wp-signup.php URL collected 749 referring domains of spam. High-authority media links
(WSJ, CNN, FT, ynet, mako) mostly point at the homepage.
Freshness & content signals lag.
The blog's newest lastmod is 2025-09-16; inspiration galleries date to 2024; several studio sitemaps date to
early 2025. Content update cadence is a headwind for both classic rankings and AI-engine citations.
Full per-page detail (every URL with its findings and recommendation codes) is in section 11 and the
downloadable CSV / XLSX files.
2 · Methodology & Data Caveats
Capture: every URL in both XML sitemaps (1,168 URLs + 1 discovered) fetched live on 2026-10-01 via Firecrawl
with a GB egress location to guarantee the Israeli version of each page. (US visitors are geo-redirected to
caesarstoneus.com — verified; capture from a US egress returns the US site, so all data here is the IL site.)
Rendering & parsing: raw HTML of each page parsed for titles, meta robots, canonicals, hreflang,
headings, structured data, images/alt, and internal links (full link graph built from 719 real pages).
Performance: Lighthouse 12.8.2, mobile default with simulated throttling, run locally on 15 key templates
(www + studio). One empty 3D page intentionally tested and confirmed unrenderable.
Backlinks: Ahrefs Site Explorer (DR, backlink stats, top referring domains, top linked pages, broken
backlink targets) pulled 2026-10-01.
Cross-checks: Firecrawl US vs GB egress comparison, Google Translate proxy fetch, Wayback Machine snapshot,
Google SERP (site: queries). Direct server access from datacenter IPs is blocked by the WAF (403),
documented for reference.
Caveats: “Empty” pages return HTTP 200 with a 106-byte document — they behave as blank pages for every
non-interactive client; if they trigger a file download for browsers, they still must not be in the sitemap.
Word counts include template text (nav/footer). “Not found” pages are detected by the site's own 404 template title.
3 · Inventory & Site Architecture
Two properties were audited in full. The main site is the consumer/brand site (Hebrew); the studio subdomain is the
professional (A&D — architects & designers) platform with its own catalog, project gallery, kits and downloadable
3D assets. A third brand subdomain exists (mos.caesarstone.co.il — Master of Stone, for fabricators) and is
outside this audit's sitemap scope.
www URLs
studio URLs
Real content pages
Internal edges crawled
Sitemap health
Every sitemap entry carries a lastmod. Freshness varies dramatically: page/catalog sitemaps were updated
days before capture, while galleries date to 2024 and the blog's newest item is 2025-09-16. Two junk sitemaps
(project-tags, project-date) contain a single URL.
4 · Technical SEO Findings
4.1 Indexability
Studio subdomain: sitewide noindex, nofollow — pages.
Effect: zero Google visibility (confirmed via site: query returning no results at all).
www: utility pages noindexed — see list below.
P0 Remove the noindex from studio (or scope it to only the pages that must stay out).
This is the single highest-impact change in this audit once the platform is meant to be public.
4.2 Canonicals
pages have no canonical tag at all (studio set + www utilities).
pages point their canonical to a genuinely different URL (list below);
a further differ only by percent-encoding case (cosmetic, not harmful).
4.3 Titles
Duplicate-title groups above — incl. studio catalog pages showing another product's title
(e.g. the 6041-Nordic-Loft page titled “6046 Moorland Fog”).
4.4 Meta descriptions
pages without a meta description; AI engines and SERPs
will synthesize snippets from raw text instead.
4.5 Structure & templates
P1 pages use more than one H1 — standardize a single
page H1 per template. page has no H1 ( /icon/ ).
Structured data present everywhere (WebSite / BreadcrumbList / WebPage; FAQPage on 9 pages). Missing types:
Product on catalog pages, Organization/LocalBusiness sitewide, and Product/Offer on studio catalog items.
Image alt coverage: of
image tags lack alt text (~84%) — hurts accessibility, image search and AI understanding.
Hreflang: present on pages only (mostly studio he-IL
self-references). Not a current problem (single-language IL site), but keeps the option open.
4.6 Sitemap junk & soft-404s
P0 empty 200-pages (3D download endpoints) and
“Page not found” URLs are listed in sitemaps — remove from sitemaps; noindex or delete.
Example "not found" URLs in sitemap: /favorites/, /author/אבן-קיסר/,
/care-and-maintenance/, /catalog/3141-egg-shell/ (a real product page returning not-found).
4.7 robots.txt & llms.txt
robots.txt (www): User-agent: * / Disallow: /wp-admin/ / Allow: /wp-admin/admin-ajax.php + sitemap line. Studio adds a few checkout paths. Neither blocks AI crawlers — good.
No llms.txt exists on either domain: requesting /llms.txt returns an HTML page (soft-200 homepage on www, "Page not found" on studio). See §9 for the recommendation.
5 · Performance (Lighthouse, Mobile)
Lighthouse 12.8.2, mobile emulation with simulated throttling — the profile Google uses for mobile-first indexing.
Result: brand templates are severely under-optimized. Production data (CrUX) would need Search Console access to
cross-check, but these lab numbers already show unmissable regressions on the money templates.
Interpretation guide: Performance < 50 = failing (Core Web Vitals risk).
LCP > 4 s is "poor"; the catalog/collections templates exceed 30 s under throttling. Studio's SEO score of 50–58 is
driven primarily by the sitewide noindex (§4.1) and non-crawlable JS links. The empty 3D page hard-fails the run
(“page did not paint any content”) — consistent with §4.6.
6 · Backlink Profile (Ahrefs)
Domain rating
Live backlinks
Live referring domains
All-time ref. domains
6.1 Top referring domains
6.2 Top linked pages
Spam & legacy findings
wp-signup.php?new=caesarstone.co.il — 749 referring domains (§6.2). Classic WordPress
registration-spam attack target. Keep noindexed/blocked; disavow only after review (it 302s away, so equity is not at risk —
the risk is anchor pollution).
Spam-flagged referring domains include za.com (100 links), pages.dev, itxoft.comlivejournal.com (26).
Spam-free authority: WSJ (6 dofollow), CNN, FT, Barrons, CBS, BBC-tier news, ynet (13 dofollow), mako (34 dofollow),
walla, technion.ac.il (19) — these mostly point at the homepage or blog posts.
Image hotlinking: ~15 top-linked “pages” are /wp-content/uploads/*.jpg images with 20–84 referring
domains each — external sites embed your images directly. Consider brand-watermarking/CDN policy, and note the lost
referral/branding opportunity (no attribution links).
Legacy link equity (recoverable): the old site's /The-Catalog/Pages/*.aspx and
/design-inspiration/... URLs still hold links from serious Israeli forums (tapuz, DR 70, dozens of threads).
Most now 301 correctly, but these legacy patterns should be audited for 404s and kept mapped
(kitchen-design-new.aspx = 404 with 37 refdomains; bathroom-design.aspx
= 404 with 30 refdomains — see table below).
7 · Internal Linking
Internal edges
Orphan pages
Link targets outside sitemap
Of which file/asset links
The internal graph is well connected overall — only real pages receive no internal
links. The larger hygiene issue: distinct internal link targets point to URLs that
are not in the crawl set (no-sitemap utility URLs, assets, query URLs) — checked, most are harmless (cart/checkout/search
paths that are deliberately non-indexable), but the list should be reviewed so no valuable page silently disappears from
the click graph.
7.1 Orphan pages
7.2 Most internally linked pages
7.3 Non-sitemap link targets (top)
8 · Duplicate Pages & Cannibalization
Duplicate titles: groups (§4.3) — including studio product pages
carrying another SKU's title. Fix at the data/template level.
Thematic overlap to review (blog):/blog/kitchens-design/, /blog/country-kitchen/,
/blog/marble-kitchen-tops/, /blog/marble-for-kitchens/ and /blog/kitchen-island-ideas/
all target kitchen-surface intent. They currently interlink poorly and compete for the same Hebrew queries
(“שיש למטבח”, “עיצוב מטבחים”). Recommend a hub-and-spoke restructure with one primary per intent.
Legacy duplicates handled: the old /design-inspiration/articles/.../kitchens-design/ URL 301s to the
blog version — correct pattern; audit remaining design-inspiration URLs for the same treatment.
Near-duplicate galleries: both “inspiration-gallery” (150 URLs) and “collections-gallery” (30 URLs) exist on www
with heavy image overlap — consider consolidation to one gallery hub.
Raw duplicate-title groups, canonical mismatches and their per-page details are in the appendix table.
9 · GEO / AI-Engine Readiness
P0Unblock the studio site — noindex also hides it from AI crawlers
(GPTBot, ClaudeBot, PerplexityBot respect robots meta). Today the whole A&D platform is uncitable.
llms.txt missing. Add /llms.txt (and optionally /llms-full.txt) per the llmstxt.org
convention: one-paragraph brand positioning, key sections (catalog, collections, projects, guides, showrooms), canonical
links, contact. Low effort, improves machine readability for AI agents.
Structured data for AI answers: add Organization (with logo, sameAs to the brand's Wikipedia/
social profiles) sitewide; Product + Offer/Material on all catalog items;
FAQPage beyond the current 9 pages (the FAQ and care pages are ideal).
Freshness: blog latest update 2025-09-16, galleries 2024 — AI engines favor recently-updated sources for
“best/top/2026” queries. Establish a monthly refresh cadence for the top-20 pages by backlinks.
Content patterns AI engines quote: add concise Q&A blocks (“How do you clean Caesarstone?”, “What is the
warranty?”) on product/FAQ pages; keep answers 40–60 words and self-contained.
Measurement: the connected AI-visibility tracker (ZeroRank) has no active tracking for this brand. Add
Caesarstone IL (prompts like “recommended quartz countertops in Israel”) to measure AI citations month-over-month.
10 · Recommendations Summary
P0 — do first
Remove noindex,nofollow from the studio site; re-check site:studio.caesarstone.co.il in 2–4 weeks; submit sitemap in GSC.
Exclude the 445 empty 3D download URLs and the 4 “not-found” URLs from sitemaps; noindex the download endpoints.
Fix / restore or 301 the 4 not-found sitemap URLs, incl. the real product page /catalog/3141-egg-shell/.
P1 — high impact
Performance pass on money templates: visualizer, catalog, collections, home (mobile LCP 20–36 s → target < 2.5 s;
TBT → < 200 ms). Start with image delivery (WebP/AVIF, correct sizes), JS deferral/splitting and caching.
Add self-referencing canonicals across studio; fix the genuinely-wrong canonicals.
Unique titles for the 17 duplicate groups (incl. studio product pages showing wrong SKU names); write meta descriptions for the 32 pages missing them.
Add descriptions/alt text: prioritize the ~top 200 landing pages; image alt coverage is ~16%.
Recover legacy link equity: review the legacy .aspx 404s (37 + 30 refdomains) and map them to current equivalents.
P2 — optimization
Single-H1 templates; Product/Organization/FAQ schema; llms.txt; internal-linking hub for the kitchen-content cluster;
monthly freshness cadence; AI-visibility tracking onboarding.