AI Visibility Score · research
PL

The David Effect: How Small Websites Beat Corporations in AI Search

The David Effect is a phenomenon identified by AIVS in which small or single-operator websites achieve higher citation levels in generative AI model responses than large corporations — despite dramatic differences in scale, budgets, and brand recognition. AI models reward content citability: SSR rendering, FAQ, structured data. They do not reward company size or Google ranking.

Data from the AIVS™ audit — March 2026

A one-person coupon portal (sole proprietorship) cited by ChatGPT and Perplexity in 4 out of 10 queries about a fashion brand. The brand itself — in 3 out of 10, exclusively help pages. The portal achieved an AIVS™ Score of 44/100, the corporation with billions PLN in revenue — 24/100.

The portal serves 294 KB of clean HTML (SSR), has dedicated per-brand articles with FAQ, Article schema + BreadcrumbList. The corporation serves 3.5 MB of JavaScript (CSR), zero blog articles, zero JSON-LD/Schema.org blocks, llms.txt returns a 301 redirect to /no-route/.

Three reasons why small beats large in AI search

1. SSR vs CSR — AI does not read JavaScript. AI crawlers act like simplified browsers without a JavaScript engine. 294 KB of clean HTML vs 3.5 MB of JavaScript — a 12x difference.

2. An article with FAQ answers the question — a product page does not. AI cites whoever answered the question, not whoever is bigger. GEO-BENCH research (Aggarwal et al., 2023) confirms: proper content structure increases AI response visibility by ~40%.

3. Article schema + breadcrumbs > zero schema. Structured data is the language a website uses to speak to AI models directly.

Strategic implications

A corporation with domain history, content budget, and a technical team has an AIVS™ Score potential of 70–85/100. Switching rendering from CSR to SSR, implementing schema, llms.txt, launching a blog section with FAQ — that is 4–6 weeks of work.

Small websites have a window of opportunity — but a short one. This window will close when corporations start taking GEO seriously.

Research finding · The David Effect
Łukasz S.·March 2026

Why a one-person blog
beats a billion-dollar
corporation
in AI search

— and what it tells us about how AI models work

A one-person coupon portal (sole proprietorship) cited by ChatGPT and Perplexity in 4 out of 10 queries about a fashion brand. The brand itself — in 3 out of 10, exclusively help pages. Data from the AIVS™ audit, March 2026. We call this phenomenon the David Effect.

The David Effect

A phenomenon in which small or single-operator websites achieve higher citation levels in generative AI model responses than large corporations — despite dramatic differences in scale, budgets, and brand recognition. AI models reward content citability: SSR rendering, FAQ, structured data. They do not reward company size or Google ranking.

Term identified and named by AIVS based on AIVS™ audits, 2026.

01
Audit data · March 2026

Numbers that should concern every marketing department

In March 2026, we conducted an AIVS™ audit for a major Polish fashion brand from a group listed on the Warsaw Stock Exchange. Alongside it, we audited a small coupon portal that wrote about the brand — a sole proprietorship run for 12 years by one person.

The results were surprising even to us. The one-person portal achieved a higher AIVS™ Score and more AI citations than the brand whose promotions it aggregated. We ran 10 synthetic benchmark queries in ChatGPT and Perplexity — questions that real customers ask when looking for reviews and comparisons.

AIVS™ audit comparison — same category, same dateAudit: March 13, 2026
Coupon portal
Sole proprietorship · 1 person · 12 years in operation · ~300k visits/month
44AIVS™ Score
/ 100
Cited: 4/10 queries about the brand
  • SSR — 294 KB of clean HTML for AI crawlers
  • Dedicated per-brand articles with FAQ in content
  • Article schema + BreadcrumbList on articles
  • Title = user question (semantic match)
  • No llms.txt (missed opportunity)
  • No FAQPage schema (FAQ exists but is not marked up)
  • Authors use pseudonyms — weak E-E-A-T
Fashion brand — large corporation
Listed on WSE · Billions PLN in revenue · Thousands of employees
24AIVS™ Score
/ 100
Cited: 3/10 queries — exclusively help pages
  • CSR — 3.5 MB JavaScript, AI crawlers see an empty page
  • Zero blog articles, zero FAQ with content
  • Zero JSON-LD/Schema.org blocks across the entire site
  • llms.txt → 301 redirect to /no-route/
  • The /about page rendered via CSR — invisible to AI
  • ~40 test pages in the public sitemap
  • Competitive gap: Zalando ~52, H&M ~48 — Sinsay 24

A portal run by one person controls more of the brand narrative in AI search than the brand’s entire marketing department.

A corporation with billions in revenue, thousands of employees, and an extensive marketing department is losing the narrative about itself to a one-person coupon aggregator. When a customer asks AI about reviews, product quality, or the loyalty program — they get an answer from the coupon portal, not from the brand website.

This is not an exception or an anecdote. It is a pattern that repeats in every industry we have audited: Polish corporations systematically lose the AI search narrative to smaller but better-structured websites.

12×
Less code for an AI crawler to process — portal (294 KB) vs corporation (3.5 MB)
AIVS™ technical audit · March 2026
0
JSON-LD/Schema.org blocks on the corporation's site — homepage, products, help
AIVS™ audit · Manual verification
+20pts
AIVS™ advantage of the coupon portal over the corporation in the same category
44 vs 24 · AIVS™ Score · March 2026
02
Mechanism · Three reasons

Why small beats large
in AI search

The David Effect is not random. It has concrete, technical causes — and each one is measurable. After analyzing dozens of Polish websites in AIVS™ audits, we see the same pattern: small sites win not because they are globally better, but because corporations make three fundamental mistakes.

1
Technical reason

SSR vs CSR — AI does not read JavaScript

AI model crawlers (GPTBot, ClaudeBot, PerplexityBot) act like simplified browsers without a JavaScript engine. When a server returns 3.5 MB of JS code, the crawler sees an empty page. When a server returns 294 KB of clean HTML with full content — the crawler processes everything. SSR (Server-Side Rendering) is not a design decision — it is a decision about AI visibility. Most Polish corporations chose React/CSR for beautiful UX and unknowingly zeroed out their AI search visibility.

2
Content reason

An article with FAQ answers the question — a product page does not

When a user asks AI „reviews of brand X — clothing quality, is it worth it?”, the model looks for content that answers that question. The coupon portal has an article titled literally „Reviews of brand X — clothing quality, prices, and customer complaints” with 4+ FAQ sections in the content. The corporation has a homepage with sliders and an /about page rendered via CSR. AI cites whoever answered the question. Not whoever is bigger. GEO-BENCH research (Aggarwal et al., 2023) confirms: proper content structure increases AI response visibility by ~40%.

3
Structural reason

Article schema + breadcrumbs > zero schema

The coupon portal has Article schema with headline, author, and interactionStatistic, plus BreadcrumbList on every article. The corporation has not a single JSON-LD block on any page. This is not a minor technicality — structured data is the language a website uses to speak to AI models directly: „I am an article, I have an author, I cover this topic.” Without schema, the model has to guess the context — and often gives up in favor of a site that communicates clearly.

AI citability factor
Sole-proprietor portal (44/100)
Corporation (24/100)
Rendering (SSR/CSR)
✓ SSR — 294 KB HTML
✗ CSR — 3.5 MB JS
Dedicated per-topic articles
✓ Articles + FAQ
✗ No content
Title = user question
✓ Semantic match
✗ Product names
Article / FAQ schema (JSON-LD)
✓ Article + Breadcrumb
✗ Zero schema
llms.txt
✗ Missing
✗ 301 → /no-route/
E-E-A-T signals
⚠ Weak (pseudonyms)
✗ No content
AI citations (10 queries)
✓ 4/10
✗ 3/10 (help only)
03
Strategic implications

What the David Effect means
for your company

The David Effect is not a permanent law of nature — it is a description of the current state of the market. Corporations are not doomed to lose. They are doomed to lose if they do nothing. The coupon portal wins not because it is hard to beat — it wins because it is the only player in this category that got to work.

Implication 1 · Corporations with budget

Weeks of work can reverse this relationship

A corporation with domain history, a content budget, and a technical team has an AIVS™ Score potential of 70–85/100. Switching rendering from CSR to SSR, implementing schema, llms.txt, and launching a blog section with FAQ — that is 4–6 weeks of work. The coupon portal does not stand a chance in the long run if the corporation takes GEO seriously.

Implication 2 · Small websites and agencies

The David Effect gives you a window of opportunity — but a short one

If your website is well-built technically and creates content that answers user questions, you have an advantage today over corporations in your category. This window will close when corporations start taking GEO seriously. Lock in your advantage now: expand your content, add FAQPage schema, create llms.txt, build a citation history before the big competitor enters this channel.

Implication 3 · For everyone

You cannot manage a narrative you cannot see

The corporation in our audit did not know that a coupon portal was controlling the narrative about its brand in AI. It had no data. Traditional SEO tools (Ahrefs, Semrush, Senuto) measure Google rankings — they do not tell you what ChatGPT says about your brand. The first step is measurement. Only by knowing where you stand can you decide where to go.

In AI search, there is no „too small to win.” There is „too little content and too much JavaScript.”

Rozwiązanie · Pierwszy w Polsce

Raport AIVS™ -
AI Visibility Score

AIVS™ (AI Visibility Score) to autorski wskaźnik AIVS mierzący widoczność polskich firm w odpowiedziach generatywnych modeli AI. Skala 0–100. 6 ważonych kategorii. 14 crawlerów. Ponad 400 przeanalizowanych polskich firm w bazie benchmarkowej.

Raport AIVS™ dostarcza to, czego żadne narzędzie SEO nie pokaże: precyzyjną diagnozę, gdzie Twoja marka jest niewidoczna w AI - i konkretną mapę działań, żeby to zmienić. Raport w języku biznesu, nie technicznym.

  • Twój wynik AIVS™ i benchmark branżowy - jak wypadasz na tle 400+ polskich firm w Twojej kategorii
  • Archetyp widoczności AI - Invisible Giant, Organic Leader, AI-Native czy Closed Fortress?
  • Audyt 6 kategorii GEO - od struktury treści, przez dane techniczne, po sygnały autorytetu
  • Konkretne rekomendacje priorytetowe - co poprawić najpierw dla szybkiego wzrostu AIVS™
  • Widoczność w ChatGPT, Gemini, Perplexity i Google AI Overviews - cztery platformy w jednym raporcie
70
Przykładowy wynik AIVS™
80–100AI-Native
60–79Organic Leader
35–59Closed Fortress
0–34Invisible Giant
FAQ - pytania kluczowe

Najczęściej zadawane pytania o GEO i AIVS™

What is the David Effect in AI search?
The David Effect is a phenomenon identified by AIVS in which small websites achieve higher citation levels in AI responses than large corporations. AI models reward content citability — SSR, FAQ, structured data — not company size or Google ranking.
Why does SSR matter so much for AI search?
AI crawlers (GPTBot, ClaudeBot, PerplexityBot) do not execute JavaScript. A site with CSR returns only JS code to the crawler — content is invisible. A site with SSR returns full HTML with content. 294 KB of HTML is 12× less than 3.5 MB of JS and 100% accessible to AI models.
How is the David Effect measured?
Through an AIVS™ audit: 10–20 synthetic benchmark queries in ChatGPT, Gemini, Perplexity, and Google AI Overviews, analysis of 6 GEO categories and 14 crawlers, a 0–100 score. Comparing a company’s score with niche sites in the same category — that is the David Effect in numbers.
Can large companies overcome the David Effect?
Yes — and quickly. A corporation with domain history has an AIVS™ potential of 70–85/100. Migration to SSR, schema markup, llms.txt, and a blog section with FAQ takes 4–6 weeks of work. The coupon portal does not stand a chance in the long run if the corporation takes GEO seriously.
Does traditional SEO protect against the David Effect?
No. Ahrefs, Semrush, Senuto measure Google rankings — not AI citations. A company can rank #1 on Google and be completely invisible in ChatGPT. These are two separate channels requiring separate optimization.
What are the fastest quick wins for companies affected by the David Effect?
According to AIVS™ audits: (1) llms.txt — 20 minutes of work, an immediate signal for models, (2) FAQPage schema on existing FAQ pages, (3) one expert article per key customer query with FAQ in the content, (4) Organization schema with sameAs on the homepage. Potential AIVS™ Score increase: +10–15 pts in a week.
Read next
GEO is today what SEO was in 2010 — a window of opportunity that will close fast
Read article →
Podsumowanie · Czas na decyzję

Nowy wyścig już trwa.
Pytanie brzmi: czy Twoja marka
biegnie - czy stoi na trybunach?

Świat nie wróci do czasów, gdy każda decyzja zakupowa zaczynała się od listy linków w Google. Coraz częściej zaczyna się od pytania do asystenta AI - a ten wskazuje 2–3 opcje, które „ma w głowie". Dobra wiadomość: wielu polskich graczy wciąż popełnia podstawowe błędy. To rzadkie okno szansy: firmy, które zaczną teraz, mogą w ciągu kilku lat „przeskoczyć kolejkę" w rekomendacjach asystentów AI.

Pierwszy krok nic nie kosztuje: darmowy AIVS™ Score — automatyczny audyt widoczności Twojej firmy w ChatGPT, Gemini i Perplexity. W 10–15 minut zobaczysz wynik 0–100 i pozycję na tle 284 przebadanych firm. Pełny raport — z listą barier i planem naprawy — odblokujesz za 149 zł netto.

Wolisz poczytać? Pobierz raport PDF →
Wynik w 10–15 minut · Pełny raport 149 zł netto · Benchmark 284 polskich firm
Łukasz S.
Founder, AIVS

Creator of the AIVS™ framework and methodology for measuring brand visibility in AI search engines. Audited over 400 Polish companies for AI-search readiness.

AIVSAIVS™ · AI Visibility Score · Poland · 2026March 2026www.aixellence.pl