Kihan Marketing
Confidential · Kihan Marketing · November 18, 2026

The invisible reasons
wikipedia.org isn’t winning
search.

SEO & AI Visibility Audit, prepared exclusively for Wikipedia (Wikipedia Foundation)
Prepared for
Wikipedia (Wikipedia Foundation)

Prepared by
Kihan Marketing
SEO · GEO (AI Search) · Growth

01 / 16

The Verdict
76 OUT OF 100

21FINDINGS

3 critical
2 high
8 medium
8 info

76/100, grade B.
Not broken. Just invisible.

Your site works fine for visitors. The invisible layer Google and AI assistants judge you on is incomplete. None of it is visible on the page. All of it is being read right now. The good news: these are the fastest, cheapest points in your marketing, and we can close most of them in weeks.

3Critical issues losing you enquiries right now
2High-impact structural gaps holding the whole site back
8Medium issues quietly compounding every month

The 3 critical issues explain most of the gap, and all three ship inside week 1 of working together. This isn’t a rebuild. It’s plumbing.

Kihan Marketing · Confidential, prepared for wikipedia.org02 / 17
Health Score & Coverage
ContentTechnicalOn-pageLinksSchemaGEO / AIImages

Where the
points live

Content quality / E-E-A-T
92
Technical SEO
75
On-page SEO
70
Link authority
72
Schema / structured data
55
AI search readiness (GEO)
73
Images
95

How to read this: the radar plots your 7 measurable categories out of 100. Bars show the same scores. Unmeasured checks (pagespeed: pagespeed) are excluded from your score, never guessed. 21 findings across 7 measured categories.

Kihan Marketing · Confidential, prepared for wikipedia.org03 / 17
The Leak Map

Where you leak, stage by stage

01
Crawl1 of 6
Google and AI bots reach your pages
Sitemap returns HTTP 403. Crawlers are locked out entirely.
BLOCKED
02
Index2 of 6
Pages actually enter the index
No canonical tags. 0.0 internal links per page. Dead ends.
WEAK
03
Rank3 of 6
Pages earn positions
On-page 70/100. Schema 55/100. Content strong, structure poor.
HELD BACK
04
Get cited4 of 6
AI names you in answers
No Organization entity. No llms.txt. Citability unstructured.
INVISIBLE
05
Get clicked5 of 6
Links render worth tapping
og:url and twitter:card missing. Shares show nothing.
BROKEN
06
Convert6 of 6
Visitors become enquiries
Core Web Vitals unmeasured (Google quota). Baseline blind spot.
UNKNOWN
Content 92, substance holds up
Images 95, one minor srcset gap
22 AI crawlers, none explicitly blocked

Kihan Marketing · Confidential, prepared for wikipedia.org04 / 17
Fix 01 · CRITICAL
AI can fix now

Your links look broken when people share them

Severity · CRITICAL
Fix time · Under 1 hour
AI can fix now

BEFORE · TODAY

A bare blue URL. No image, no title. Zero reason to tap.

Missing required: og:url

AFTER · WEEK 1

A rich preview card with your brand image, headline and description. Verified on every platform.

Verified before we call it done

Why it matters: When someone shares your page on WhatsApp, LinkedIn or Facebook, the preview shows a wrong or missing image, title or link. First impressions decide the click, and right now yours says nothing.

Kihan Marketing · Confidential, prepared for wikipedia.org05 / 17
Fix 02 · CRITICAL
We handle it together

Google doesn't know who you are

Severity · CRITICAL
Fix time · 1 week
We handle it together

BEFORE · TODAY

Invisible entity. AI and Google quote you with no brand, no logo, no trust signal.

No Organization/Person entity found in JSON-LD.

AFTER · WEEK 1

One clean entity graph: Google and AI resolve your brand to a verified, cited source.

Verified before we call it done

Why it matters: Structured data is the ID card Google reads to show your brand in search. Without an Organization or Person entity in your JSON-LD, Google can't connect your site to a real, trusted entity, so AI assistants cite someone else.

Kihan Marketing · Confidential, prepared for wikipedia.org06 / 17
Fix 03 · CRITICAL
We handle it together

Google's table of contents is locked out

Severity · CRITICAL
Fix time · 1–2 days
We handle it together

BEFORE · TODAY

Crawlers hit a locked door (403). Your new pages never get read.

Primary sitemap returned HTTP 403: https://en.wikipedia.org/w/rest.php/site/v1/sitemap/0

AFTER · WEEK 1

Every page discoverable on schedule. Re-indexing confirmed in Search Console.

Verified before we call it done

Why it matters: Your sitemap, the file that tells Google every page you want ranked, returned HTTP 403 (forbidden). New and updated pages can sit invisible for weeks. Nothing else matters until crawlers can read it.

Kihan Marketing · Confidential, prepared for wikipedia.org07 / 17
Fix 04 · HIGH
We handle it together

Your pages don't pass authority to each other

Severity · HIGH
Fix time · 2 weeks
We handle it together

BEFORE · TODAY

Orphan pages. Ranking power stuck on the homepage, nothing flowing deeper.

Average internal links per page is only 0.0 (target: 5-10).

AFTER · WEEK 1

Money pages inherit authority. Impressions turn into rankings across the whole site.

Verified before we call it done

Why it matters: Average internal links per page is 0.0. Your pages are islands. Internal links are how ranking power flows from strong pages to the ones that sell. Right now, every page fights alone.

Kihan Marketing · Confidential, prepared for wikipedia.org08 / 17
Fix 05 · HIGH
Fix in our working session

Google doesn't know which version of you is real

Severity · HIGH
Fix time · 1–2 days
Fix in our working session

BEFORE · TODAY

Duplicate URLs splitting your own ranking signals across copies.

No canonical URL in HTML or HTTP Link header.

AFTER · WEEK 1

One authoritative URL per page. All signals consolidated, nothing split.

Verified before we call it done

Why it matters: Without a canonical tag, every parameter or duplicate URL competes with the original. Ranking power splits instead of combining. Google picks its own favorite, often the wrong one.

Kihan Marketing · Confidential, prepared for wikipedia.org09 / 17
The Complete Record

All 21 findings, nothing hidden

3
2
8
8
3 critical2 high8 medium0 low8 info
Severity Finding Category
CRITICAL Missing required: og:url social
CRITICAL No Organization/Person entity found in JSON-LD. entity
CRITICAL Primary sitemap returned HTTP 403: https://en.wikipedia.org/w/rest.php/site/v1/sitemap/0 sitemap
HIGH Average internal links per page is only 0.0 (target: 5-10). link_profile
HIGH No canonical URL in HTML or HTTP Link header. canonical
MEDIUM 5 security header(s) missing: Content-Security-Policy (CSP), X-Frame-Options, X-Content-Type-Options, Referrer-P security
MEDIUM 1 page(s) have fewer than 3 internal links internal_links
MEDIUM Missing: twitter:card social
MEDIUM Entity schema sameAs is missing 4 profile link(s): Wikipedia (Primary), Wikidata (Primary), LinkedIn (Strong), T entity
MEDIUM Suspected firewall block: AI search crawler user agents were refused while a browser was served: meta-webindexer ai_bot_access
MEDIUM 1 page(s) with no outbound internal links (dead ends). link_profile
MEDIUM 6/7 images lack srcset. image_seo
MEDIUM 3 redirect chain(s) detected (>1 hop) broken_links
INFO 22 AI crawlers have no rule of their own (allowed unless * blocks them): OAI-SearchBot, Claude-SearchBot, Perple robots
INFO No clear author/byline signal detected content_quality
INFO No publication or updated date detected content_quality
INFO Suspected firewall block: AI user-fetch user agents were refused while a browser was served: meta-externalfetche ai_bot_access
INFO Link equity by section was not measured. architecture
INFO Content readability is moderate (Flesch: 55.2) — suitable for educated audience readability
INFO 1 image(s) have srcset but no sizes attribute. image_seo
INFO 7 image(s) served as JPEG/PNG instead of WebP. image_seo
Kihan Marketing · Confidential, prepared for wikipedia.org10 / 17
The New Front Door · GEO

Google ranks you. AI decides
whether to name you.

73 OUT OF 100

AI-VISIBILITY SCORE
01
AI crawler accessGPTBot, PerplexityBot, OAI-SearchBot and Claude-SearchBot allowed on your money pages, not just the homepage.
02
llms.txtA published map of your 10 highest-value pages for AI ingestion. You have none today.
03
Entity consistencyOne organization entity everywhere: JSON-LD sameAs, Wikidata, LinkedIn. AI answers about you resolve to one clean fact.
04
Citability passages134–167 word answer blocks, claims first. Engineered so quoting you beats quoting a competitor.

Kihan Marketing · Confidential, prepared for wikipedia.org11 / 17
The Cost Of Waiting

Every month you wait,
a competitor books the call
that should be yours.

Search demand100%
Crawled & indexed✕ blocked
Ranked & cited by AIweak
Clickedbroken
Enquiry?

01
Stay invisible

AI Overviews and ChatGPT cite whoever has schema, sitemaps and entity markup. Being the named answer is the new page 1.

02
Pay more later

Technical debt compounds. Duplicates multiply, chains grow, crawlers give up. 90 days of rot costs 5× fixing it now.

03
Ads work overtime

Without organic trust, paid traffic carries everything, and never gets cheaper. Organic and AI citations are the only channels that appreciate.

The fix is priced in days.The leak costs you every month. One call closes it.
Kihan Marketing · Confidential, prepared for wikipedia.org12 / 17
The Plan

90 days. Three moves.

Day 1
Day 14
Day 30
Day 60
Day 90

Fix · Week 1–2

  • Sitemap unblocked at CDN, resubmitted. Crawlers read everything by day 3
  • og:url and share tags live, verified on WhatsApp, LinkedIn, X
  • Organization + Person JSON-LD with sameAs entity graph
  • Canonicals everywhere, duplicates consolidated
  • robots.txt tuned: search and AI crawlers flowing

Build · Day 30

  • Internal linking rebuilt around money pages: 5–10 contextual links each
  • Author, dates, credentials on key content. E-E-A-T repaired
  • llms.txt published. Citability passages on top pages
  • Schema complete: Organization, LocalBusiness, FAQ, Breadcrumb

Compound · Day 60–90

  • AI-citation sprint: engineer the answer so AI names you
  • Drift monitoring: alerts the day anything breaks
  • Day-90 re-audit vs this baseline, in plain English

Kihan Marketing · Confidential, prepared for wikipedia.org13 / 17
Our 20-Minute Call

Four questions I’ll ask
on our first call

1
“Which search makes you money right now?”Most owners can name traffic, almost none can name the query that produced a sale. Until we know that, every tactic is decoration.
2
“When a buyer asks ChatGPT about your category, who gets named?”We’ll run it live, together. If it isn’t you, we’ll watch it flip in 30 days, and keep it that way.
3
“What would one extra qualified lead a day be worth?”Your score is 76. The gap to visible-and-cited is mostly plumbing, not luck. If the math works, we work.
4
“What can we fix this week without touching your team?”The AI-can-fix-now items ship in days with zero disruption. I’ll bring the exact list to the call.
Kihan Marketing · Confidential, prepared for wikipedia.org14 / 17
The Receipts · Evidence Appendix

The actual links that
waste your visitors’ clicks

370 links checked1 page(s) crawled9 documented receipts
# URL Issue Where
01
www.wikisource.org/
link text: “WikisourceFree content library”
2-hop redirect

→ https://wikisource.org/wiki/Wikisource:Main_Page (2 hops)
found on
www.wikipedia.org
02
itunes.apple.com/app/apple-store/id324715238?pt=208305&ct=portal&mt=8
link text: “Apple App Store”
2-hop redirect

→ https://apps.apple.com/us/app/wikipedia/id324715238 (2 hops)
found on
www.wikipedia.org
03
to.wikipedia.org/
link text: “faka Tonga”
Refused / 429

HTTP 429
found on
www.wikipedia.org
04
ty.wikipedia.org/
link text: “Reo tahiti”
Refused / 429

HTTP 429
found on
www.wikipedia.org
05
ts.wikipedia.org/
link text: “Xitsonga”
Refused / 429

HTTP 429
found on
www.wikipedia.org
06
tk.wikipedia.org/
link text: “Türkmençe”
Refused / 429

HTTP 429
found on
www.wikipedia.org
07
tw.wikipedia.org/
link text: “Twi”
Refused / 429

HTTP 429
found on
www.wikipedia.org
08
tyv.wikipedia.org/
link text: “Тыва дыл”
Refused / 429

HTTP 429
found on
www.wikipedia.org
09
udm.wikipedia.org/
link text: “Удмурт”
Refused / 429

HTTP 429
found on
www.wikipedia.org

Why it matters: every dead or redirected link is a paid-for or earned visitor ending at nothing. Visitors blame you, not the link. Google counts every one against you.

Kihan Marketing · Confidential, prepared for wikipedia.org15 / 17
Appendix · Methodology

Honest limits, stated plainly

What was measured, 18 check categories

Meta and social tags · JSON-LD schema · entity consistency · robots.txt and AI-crawler access · sitemap reachability · canonicalization · internal link graph and dead ends · broken links and redirect chains · content quality and E-E-A-T · readability (Flesch 55.2) · images · security headers · hreflang · llms.txt · hidden-instruction scan · GEO citability structure.

What was not measured, and why

Core Web Vitals (PageSpeed): Google’s free quota was exhausted this pass. With a dedicated key it re-runs in minutes. Local signals: not applicable to this property. Link equity by section: crawl depth limited this run; the Fix phase includes a full crawl.

What this is not

Not a ranking guarantee. No ethical agency sells those. This is a measured technical and AI-readiness baseline. Every claim in this deck is reproducible with the same toolchain. The day-90 re-audit shows the delta, in numbers.

Data sources

Read-only inspection of public pages and files of wikipedia.org on November 18, 2026. No private data, no accounts, no credentials. Every finding falsifiable. Ask for the evidence file behind any slide.

Kihan Marketing · Confidential, prepared for wikipedia.org16 / 17
Kihan Marketing

Quiet site.
Loud competitors.
Let’s fix it.

One 20-minute call. I’ll show you your top flippable page live, and the exact 30-day sequence we’d run. No deck theatre. No fluff.

Book your 20-min callReply to this email

KIHAN MARKETING — SEO · GEO · GROWTH

Kihan Marketing · Confidential, prepared for wikipedia.org17 / 17