What a page "footprint" is
A footprint is a set of telltale signals that let an ML model link several of your pages into a single cluster. If Facebook finds a cluster of 5 pages and one of them is tied to a previously banned account, the other 4 get banned too. Preemptively, with no warning.
Top 10 footprint markers
- Identical CSS class names (
.hero,.btn-primary) across your 10 pages - Identical DOM structure — the same sequence of <section> with the same classes
- Identical fonts from Google Fonts (Roboto + Inter are popular)
- Identical stock photos (especially from free libraries)
- Identical text patterns ("Welcome to our company", "About us", "Contact us")
- Identical domains at a single registrar with identical WHOIS
- Identical hosting IP
- Identical Schema.org JSON-LD
- Identical favicons
- Identical <meta> tags and
generatorstring
What to do about each marker
CSS classes
Prefix every class with a unique random suffix per page: .xz9hb-hero, .q5tn4-hero. AI generators like WhiteLands do this automatically.
DOM structure
Use different section orders. Instead of always "hero → features → testimonials → CTA", vary it:
- Page 1: hero → testimonials → features → CTA
- Page 2: hero → features → CTA → testimonials
- Page 3: hero → features → about → CTA
Fonts
A different pairing (heading + body) per page. 30+ pairs are enough for 1000 pages with no repeats. Avoid Roboto/Inter — too popular.
Stock photos
A different photo set per page. Pexels + Unsplash + Picsum give you plenty for 100+ pages with no overlap.
Text patterns
AI generation helps — every page gets different hero, section, and CTA copy. If you write them manually, vary the wording.
Domains
Use different registrars (Namecheap + Porkbun + Cloudflare Registrar). Don't register 20 domains on one card in a single day.
Hosting
Spread them across different providers:
- Cloudflare Pages (different accounts)
- Netlify (different accounts)
- Vercel
- Your own VPS on different providers
Schema.org JSON-LD
It should be on every page, but with different data: different addresses, phones, opening hours, social links.
Favicons
Generate a unique one for each page. You can do it automatically via AI or an emoji favicon maker.
What you should NOT change (vs. what to make unique)
A common mistake is over-uniquifying. Some things should look "ordinary":
- <head> structure — title, meta, link in the normal order
- The same charset (UTF-8) on every page — this isn't a tell
- Normal HTML5 doctype — nothing custom needed
How to check your page isn't footprint-clustered
- Open 2-3 of your pages in dev tools
- Look at the CSS source — are the classes unique
- Pull the HTML structure via
document.documentElement.outerHTML - Diff for similarity via diffchecker.com — if <30% match, you're good
- If >50% — your pages are in a cluster and need more uniquifying
Tools for anti-footprint
- WhiteLands (ours) — automatically randomizes CSS, fonts, structure, images
- Anti-detect browsers (Multilogin, Octo) — for managing accounts, not pages
- WP plugins (WP-Rocket + custom CSS minify) — only marginally useful
- Manual editor — the best control, but slow
Bottom line
Anti-footprint isn't an "option" in 2026 — it's a necessity. Without it, every ban drags 2-5 other pages down with it. If you're generating 20+ pages a month, pick tools that automate it.
Learn more about AI page generation with built-in anti-footprint.