Every Hebrew page of the Israeli site — 660 pages — with its own full report: measured data + prioritized recommendations. Click any row to open the page's report (each has its own URL).
660 pages · 100% coverage2,932 recommendationsaudit date 2026-10-01read-only · no site changes
9
CRITICAL pages
417
HIGH priority pages
118
MEDIUM
72
indexable pages with ZERO internal links
60
money pages with no inbound links (orphans)
16–20s
mobile LCP (homepage, lab)
10.9MB
homepage weight (mobile)
100
product pages without Product schema
820
broken backlinks (10.6%)
#1
AI-recommendation position in Israel (mention rate 70.5%)
67%
of AI citations about the brand come from third parties, not the site
33.5K
Google clicks (90d, Search Console)
26.5K
monthly impressions leaking at <1% CTR on 3 pages
§All 660 pages — click to open a page's full report
Every public content type of www.caesarstone.co.il was enumerated and pulled in full — 666 URLs, of which 660 are the Hebrew pages covered here (English-only pages and the separately-managed studio subdomain are summarized below instead). For each page: title & meta description, indexing & canonical, content depth, headings, keyword density, outgoing/incoming internal links with anchor text, images & alt text, structured data, social preview data, duplication, freshness — and a prioritized recommendation list.
Studio subdomain (studio.caesarstone.co.il) — summary
The B2B designer portal carries 1,019 URLs, 100% noindexed (correct for a portal with cart/checkout/3D tooling). Two findings matter for the main site: 194 slugs exist on both properties (97 catalog + 85 product + 12 pages; 46 pairs byte-identical) and one content bug — studio's /catalog/508-isobellia/ serves lucillia content. While studio stays noindexed there is no Google conflict, but add studio→www canonicals as defense-in-depth before any future indexation.
02Key site-wide findings
A. Performance (mobile, measured 2026-10-01)
Largest Contentful Paint16.1–19.9 s (target ≤ 2.5 s)Total Blocking Time2.4–2.8 s (target ≤ 0.2 s)Time to Interactive34–39 sFirst Contentful Paint2.3–11.0 s (target ≤ 1.8 s)Cumulative Layout Shift0 ✓Server response (TTFB)20 ms ✓ — hosting/CDN are finePage weight10.9 MB per page (target < 2 MB)Unused JavaScript~2.2 MB removable (≈6–8.5 s)Real-user field dataAVERAGE — the lab numbers are worse than the field median, meaning weak devices/mobile networks suffer most
Diagnosis: the server responds instantly; the problem is front-end weight — page-builder + 3D-visualizer JavaScript shipping site-wide, unoptimized images, and heavy main-thread work. Fix order: (1) load the 3D/visualizer bundle only on pages that use it; (2) WebP/AVIF images + responsive sizes + lazy-loading; (3) purge unused CSS/JS and defer non-critical scripts; (4) long-cache static assets; (5) fix production console errors; (6) re-measure weekly until LCP ≤ 2.5 s in the field.
B. Titles & meta descriptions
Homepage title is "Home Page" (9 chars) — the single most important page on the domain carries a placeholder. Fix first: "אבן קיסר | משטחי קוורץ ופורצלן למטבח ולבית".
8 pages with no title tag at all (mostly cnm/* material pages — noindexed, but fix before any indexation change).
52 titles over 60 chars (truncated in search results — mostly FAQ questions up to 86 chars).
216 indexable titles under 20 chars — wasted ranking signal (e.g. /icon/ titled "ICON" despite 1,451 words of content; /new_porcelain/ titled "GAP").
13 indexable pages missing a meta description; 64 groups of duplicated descriptions (template copy).
278 duplicate-title groups — mostly harmless (studio twins/noindex), but one is live cannibalization: "שיש למטבחים" exists as both a blog post and a page → merge + 301.
Some descriptions contain raw HTML entities (") and one is 453 chars (competition terms page) — clean and shorten.
C. Internal linking
72 indexable content pages have ZERO internal links in their content (avg 1.36 links/page; healthy: 3–8).
60 money pages (catalog/material/core) receive no inbound internal link — they cannot rank well without equity flowing in.
Link equity concentrates on /warranty (116 inbound links — a utility page), while the catalog hub gets 36 and material pages far fewer.
Only 2 links connect the main site and the studio subdomain.
Generic/raw-URL anchors in use (e.g. "לחצו כאן", pasted URLs) — replace with keyword anchors.
Fix at template level: blog → auto "related products" module (3–5 links); catalog → collection + siblings + guide + visualizer links; FAQ → hub links. One implementation fixes hundreds of pages.
D. Structured data
All pages carry the generic set (WebPage + WebSite + Breadcrumb) — but 100 catalog/product pages have no Product schema and 53 FAQ pages have no FAQPage schema (site-wide there is exactly 1 FAQPage block, on a blog post).
No Article/BlogPosting on the 119 blog posts; no Organization/LocalBusiness entity block.
This is the highest-leverage on-page fix: template-level implementation unlocks rich results AND AI-answer extraction across the whole site at once.
E. Content & freshness
44 of 119 blog posts are under 500 words; two are near-empty (48–57 words) — including "אז מה ההבדל בין קוורץ לפורצלן?", a high-intent comparison query answered in 48 words.
79 posts last modified before 2024 — refresh, merge or prune.
53 FAQ answers average ~50 words — too thin for featured snippets/AI citations; expand top 20 to 150–300 words, answer-first.
Catalog pages show ~30–50 crawlable words — product data renders through custom fields; verify it is server-rendered (visible in page source), otherwise Google sees near-empty pages.
56 exact-duplicate content groups within the site (incl. a 3-page group where /catalog/508-isobellia/ shows lucillia content — a content bug).
F. Search visibility (google.co.il, Hebrew, 2026-10-01)
Query
Intent
caesarstone.co.il positions
Verdict
אבן קיסר
brand
#4 (homepage, live check) — but GSC avg position 1.1–1.3 with 49.9% CTR on the homepage
Live-SERP vs GSC gap: the homepage usually ranks #1, but at check time own pages (/project-checklist/, a gallery page) sat above it — brand SERP dilution is intermittent; consolidate with sitelinks-worthy structure
caesarstone
brand (EN)
#2 (homepage)
Global caesarstone.* domains hold the rest — expected
משטחי קוורץ
category
#1 (/caesarstone-quartz/)
✓ money keyword won
שיש למטבח
category
#1 (/catalog/) + #6 (blog)
✓ but the duplicate-title pair threatens this
משטח אבן קיסר מחיר
commercial
ABSENT from top 30
A third-party fabricator (la-casa.co.il, "Caesarstone Group B") captures this buying-intent brand query at #4. Build a pricing/where-to-buy page targeting it.
אבן קיסר קטלוג
navigational
#1 + #2
✓
AI Overviews appear on 4 of 6 queries. With no Product/FAQ schema, no llms.txt and thin FAQ answers, the brand has no structured footprint in the AI answer layer for its own market (see GEO below).
G. Backlinks & authority (measured 2026-10-01)
Total backlinks7,725Referring domains568 (512 main domains)Broken backlinks820 (10.6%) → 44 broken on-site targetsSpam scorelow ✓Link types81% image links · only 1,433 anchor-text links — thin contextual profileTop referrerscaesarstone.com / .us / .ca / .in / .co.za / .com.au … — overwhelmingly the corporate family; genuine third-party Israeli editorial links are a small fraction (.co.il: 505)
Actions: (1) broken-link reclamation — 301 the 44 dead targets, contact high-authority referrers with corrected URLs; (2) digital-PR for contextual anchor links from Israeli design/real-estate media; (3) competitor-gap analysis vs Silestone/Dekton IL.
H. AI visibility — MEASURED (Israel scope, 30-day window, data refreshed 2026-10-01)
#1
overall position in AI engine recommendations (IL)
70.5%
brand mention rate across AI answers (71.5% latest, 2026-10-01)
2.8
avg. rank when mentioned (improving → 2.5)
0.2
sentiment score (positive, -0.02 vs previous period)
The brand leads AI answers in Israel — but the lead is thin and eroding at the edges. Competitor standings over the same window:
Brand
Position
Mention rate
Avg. rank
Note
Caesarstone
#1
70.5%
2.8
own brand
Dekton
#2
65.2%
3.1
gap of 5.3 pts — closing fast
Laminam
#3
55.9%
4.1
—
Cosentino
#4
34.1%
3.9
—
Silestone
#5
30.0%
4.4
—
Dekton is 5.3 points behind and rising — and the tracked AI answers about Dekton comparisons repeatedly cite third-party comparison pages (14.9% of comparison answers cite one source alone). Every comparison question ("Caesarstone vs Dekton", "quartz vs porcelain") is currently answered by OTHER sites' content, not yours.
Who the AI engines cite when Caesarstone comes up (30d, IL)
Read this table carefully: the #1 source is caesarstone.co.il itself (27.9% — led by /caesarstone-quartz/, /faq/, /blog/marble-kitchen-tops/). But positions #2–#15 are distributors, fabricators and comparison sites (midrag.co.il, bonimbayit.co.il, jan-marble.co.il, marble-elchai.co.il, liveinstyle.co.il…). The AI answer layer about YOUR brand is 2/3 controlled by resellers and third parties. On the global (US) battery the same pattern holds: caesarstoneus.com pages get cited, caesarstone.co.il gets zero global citations — and the editorial domains that shape answers (thespruce.com, slabwise.com, consumerreports.org, lifestoryresearch.com) have no IL equivalent on your side.
Source momentum (90 days)
Rising citers
Jan Marble (jan-marble.co.il) — 0.1% → 1.8% (+1638.9%)
Even Marble (even-marble.co.il) — 1.0% → 0.0% (-100.0%)
Top Best Stone (topbeststone.com) — 0.5% → 0.0% (-100.0%)
Wikipedia (wikipedia.org) — 0.5% → 0.0% (-95.3%)
Fixr (fixr.com) — 0.7% → 0.0% (-95.1%)
Darkstone Natural Stone (darkstone.co.il) — 0.5% → 0.1% (-89.5%)
i-Marble (i-marble.co.il) — 1.1% → 0.1% (-87.1%)
Chef Center (sheff.co.il) — 0.5% → 0.1% (-77.8%)
Ben-Eli Law Office (ben-eli.co.il) — 0.6% → 0.1% (-77.0%)
Notable declines: wikipedia.org (-95%), reddit.com (-72%), realsimple.com (-76%) — the trust-layer sources that AI engines weight heavily are citing Caesarstone LESS than 90 days ago. New entrant: la-casa.co.il (the same reseller that captures your price query in Google — it is now entering the AI answer layer too).
The measured action list (17 open items)
Action type
Count
Impact/Effort
Triggering source
** comparison_page
4
High/Medium
** Marble Elchai
** create_how_to
4
High/Medium
** Infinity Surfaces
** create_blog_post
3
High/Medium
** Laminam Israel
** create_what_is
2
High/Medium
** Laminam Israel
** create_listicle
2
Medium/Medium
** Marble Elchai
** update_robots_txt
1
High/Low
** Caesarstone Israel
** faq_missing
1
Medium/Low
** Caesarstone Israel
Two items are flagged High-impact / Low-effort and corroborate the crawl findings independently: robots.txt blocks parts of the site from AI systems (see §I) and there is no consolidated FAQ surface for the core money queries (שיש למטבח, שיש קוורץ, שיש פורצלן, משטח למטבח) — the 53 FAQ fragments are not recognized as an FAQ entity because they lack FAQPage schema (see §D). The comparison-page items name the exact third-party pages winning the "vs Dekton / vs Cosentino / vs Laminam" answers.
What a tracked global AI answer actually says (sample, 2026-10-01)
"What are the best-reviewed countertop brands for home renovations?" — the engine shortlisted Caesarstone #1 ("strong consumer trust, extensive designs, established brand"), citing a 2026 consumer-trust study. All 5 Caesarstone citations in that answer were caesarstoneus.com pages; none from the .co.il site. The engines also volunteered the caveat that "the fabricator/installer can matter as much as the brand" — which is precisely the message your Israeli distributors' pages are positioned to shape.
H2. Google Search Console — the site's own numbers (connected property)
33,540
clicks (90 days)
529,493
impressions (90 days)
7,418
clicks (28 days, all devices)
9.8
avg. position (28d, by-page)
Top queries (28d)
Query
Clicks
Impr.
CTR
Pos.
אבן קיסר
906
1,624
55.8%
1.3
שיש קיסר
349
930
37.5%
2.1
אבן קיסר קטלוג
294
537
54.7%
4.7
שיש למטבח
115
2,086
5.5%
2.5
שיש קיסר קטלוג
94
321
29.3%
8
שיש
85
1,797
4.7%
4.9
שיש אבן קיסר
81
393
20.6%
5.2
שיש קיסר קטלוג מחירים
73
336
21.7%
7.6
אבן קיסר שירות לקוחות
34
100
34%
17.5
קיסר
29
1,301
2.2%
6.6
tci ehxr
28
35
80%
1
אבן קיסר בני ברק
27
162
16.7%
2.6
Device split (28d)
Device
Clicks
Impr.
CTR
Pos.
MOBILE
4,746
65,800
7.2%
6.9
DESKTOP
2,624
67,296
3.9%
14.5
TABLET
48
580
8.3%
8.1
Mobile earns 64% of clicks with position 6.9; desktop sits at 14.5. The performance disaster in §A is hitting exactly the device that converts — every second shaved off LCP protects this split.
Rich-result presence
Only appearance type registered: VIDEO (216 impressions, position 44.8). Zero FAQ, Product, Breadcrumb or Review rich results in 28 days — direct confirmation of §D: the schema simply is not there to be shown.
Query → page mapping (28d):אבן קיסר lands on the homepage at 49.9% CTR ✓ · שיש למטבח (1,931 impressions, the biggest non-brand term) lands on /catalog/ at 5.7% CTR and on /project/שיש-למטבחים/ at 1% CTR from position 11.2 with 8,597 impressions — the single largest click-leakage page on the site. /blog/kitchen-island-ideas/ (9,603 impr., 0.6% CTR) and /blog/marble-kitchen-tops/ (8,286 impr., 0.6% CTR) leak the same way. These three pages = ~26,500 impressions/month converting at <1%: title/snippet rewrites here are worth more than any other on-page change.
Per-page GSC data is merged into the page reports below: 49 pages carry their own clicks/impressions/CTR/position card (§10 of each report) + performance-driven recommendations.
I. Technical & security notes
3 real canonical mismatches on indexable URLs: /catalog-ar/→/catalog/ (Arabic catalog folded into Hebrew — verify intent), plus two correct legacy consolidations (/porcelain_2022/, /inspiration-gallery/).
156 URLs use percent-encoded Hebrew slugs; 34 slugs exceed 70 chars — keep existing URLs (renaming costs 301s), enforce short slugs on new content.
No hreflang anywhere — acceptable if the site is Hebrew-only for Israel; declare lang="he" consistently and verify no language variants target the same users.
Production JavaScript console errors on the homepage; deprecated APIs flagged; 2 third-party cookies not partitioned (Chrome readiness).
Mobile viewport misconfiguration on the main site (user-scaling/width flags) + missing frame titles, empty link names, contrast failures — accessibility items that double as mobile-usability signals.
The public content API exposes the entire site structure unauthenticated — rate-limit or restrict it if the front-end does not depend on it.
Automated server-side crawling is blocked by the site's bot protection (all non-browser traffic gets 403) — verify legitimate search-engine and AI crawlers are allow-listed at the protection layer; one measurement indicates a render-path block on the studio subdomain.
03Prioritized action plan
P0 — this week
Homepage title "Home Page" → branded Hebrew keyword title + meta description.
Fix the mobile viewport flag + production console errors.
Add meta descriptions to the 13 indexable pages missing them.
Resolve the "שיש למטבחים" blog-vs-page duplicate (merge + 301).
Fix the isobellia/lucillia content bug (studio twin) and the 3 canonical mismatches (verify /catalog-ar/ intent).
Build a pricing/"where to buy" page targeting "משטח אבן קיסר מחיר" — currently lost to a reseller.
CTR-leak triage (GSC-measured): rewrite titles/snippets on the 3 biggest leaking pages — /project/שיש-למטבחים/ (8,597 impr. @ 1% CTR), /blog/kitchen-island-ideas/ (9,603 @ 0.6%), /blog/marble-kitchen-tops/ (8,286 @ 0.6%). ~26.5K monthly impressions converting below 1% — largest single click-gain available.
Unblock AI crawlers: the measured action list flags robots.txt blocking key sections from AI systems — audit the robots directives and allow the money paths (High impact / Low effort).
P1 — this month
Schema program: Product (catalog), FAQPage (53 FAQs), Article (119 posts), Organization site-wide — template-level. GSC confirms zero rich results except VIDEO in 28 days.
Consolidated FAQ surface for the money queries (שיש למטבח, שיש קוורץ, שיש פורצלן, משטח למטבח) — the 53 fragments are not recognized as an FAQ entity.
Internal-linking program: related-content modules in blog/catalog/FAQ templates; de-orphan the 60 money pages; rebalance equity from /warranty to catalog hubs.
Performance sprint 1: conditional-load the 3D/visualizer bundle; WebP/AVIF + lazy images; purge unused CSS/JS → target LCP < 4 s. Mobile = 64% of clicks (position 6.9 mobile vs 14.5 desktop) — this protects the majority of search revenue.
Rewrite the 52 over-length and 216 under-length titles (indexable first), prioritizing the high-impression/low-CTR pages flagged in each page's §10 card.
Expand top 20 FAQs to 150–300 words, answer-first format.
Decide cnm/* indexation; verify page-builder content is server-rendered on catalog pages.
Own the comparison answers: publish first-party "Caesarstone vs Dekton / vs Laminam / vs Cosentino" and "quartz vs porcelain" comparison pages — third-party comparison pages currently appear in ~15% of the AI answers on these questions, and Dekton's mention rate (65.2%) is 5.3 points behind and rising.
P2 — this quarter
Performance to LCP ≤ 2.5 s field: code-splitting, caching, legacy-JS purge; weekly regression measurements.
GEO program: llms.txt, AI-crawler policy, AI-answer monitoring, citation-share target — cut the third-party citation share (currently ~2/3 of AI citations about the brand) by growing owned-page citations beyond the current 27.9%.
Trust-layer reclamation: wikipedia/reddit/realsimple citations fell 72–95% over 90 days — targeted digital-PR to restore them; monitor la-casa.co.il-style resellers entering the answer layer.
Blog refresh: 79 pre-2024 posts — update/merge/prune; quarterly cadence.
Broken-backlink reclamation (820 links / 44 targets) + digital-PR for contextual Israeli media links.
Studio strategy: cross-subdomain canonicals; decide if B2B landing pages should index.