How scoring works

How ready a website is to be found by search engines, answer engines, and AI, from 0.00 to 100.00. The average of five Pillar Scores. Every site is read the same way, once a day at most, and scored on the same five Pillars.

How a site is read

IndexBot fetches the homepage, the robots.txt file and the sitemap, then a sample of pages found through the sitemap and the site’s own links. It reads the raw HTML the way a search engine’s crawler does: no scripts run, no images load. Alongside the pages it gathers public facts about the domain itself, such as how many other sites link to it, whether real people visit it, and how long it has existed. Scores reflect what could be read from the pages available that day; an Analysis either completes or it does not, and there are no partial scores.

The five Pillars

The Index Quotient is the plain average of five Pillar Scores. Each Pillar is made of Themes, and a Scorecard reports every Theme that applies to a site at one of four Levels. Here is what each Theme looks at, and why it matters for being found.

Technical Health

Whether search engines can fetch, trust, and quickly load the site.

Secure connection
Browsers and search engines treat an encrypted site as the default and an unencrypted one as suspect.
Safe to visit
A site flagged for malware or phishing sits behind a warning in every browser and is pushed down in results.
Speed
Slow pages lose visitors before they lose rankings, and search engines measure both.
Crawlability
A search engine can only rank what it can reach; a sitemap and working links are how it finds the rest.
URL and canonical hygiene
When the same page answers at several addresses, its standing is split between them.

Content Quality

Whether pages say clearly what they are about, in a form machines can read.

Titles and descriptions
The title and description are what a search result shows, and they decide whether anyone clicks.
Headings
Headings are how a machine, and a hurried reader, work out what a page is about.
Depth and readability
A page has to say enough to be worth ranking, and say it plainly enough to be understood.
Images and links
Described images and links that say where they go let a crawler see more than the words.
Freshness
Dated, recently updated content tells search engines the site is alive.
Language
A declared language tells search engines whom a page is for and which version to show.

Answer Readiness

Whether content is shaped and marked up so an engine can lift a direct answer.

Structured data
Structured data tells an engine what a page is, so it can show it as more than a blue link.
Question-and-answer shape
Answer engines lift a short answer from under a question; content shaped that way gets chosen.
Scannable formatting
Lists, tables and a summary opening are what an engine can lift without reading everything.
Navigation aids
A breadcrumb trail shows an engine where a page sits in the site, and shows the reader the same.

AI Visibility

Whether AI systems can reach, read, and confidently identify the site.

AI access
An AI assistant can only cite a site its crawlers are allowed to read and quote.
Readable without JavaScript
Most AI crawlers read the raw page; content that only appears once scripts run is invisible to them.
Identity
An AI answers confidently about a site it can identify: who runs it, and whether the public record agrees.
Authorship and sourcing
A byline, a date and links to sources are what an AI weighs when deciding whether to trust a page.
Feeds
A feed and dated sitemap entries tell machines when something new appears, without re-reading the site.

Authority

Whether the wider web vouches for the site.

Backlinks
Links from other sites are the web's vote of confidence, and search engines still count them first.
Traffic
How many people already visit a site is the plainest measure of whether the web has found it.
Domain history
A domain that has been around for years has a track record; a new one has none yet.
Social presence
Profiles linked from the site connect it to the places people already are, and confirm they are the same organization.

Levels

A Theme’s Level is one of four words. A Scorecard lists the strong Themes under a Pillar as strengths and the rest as opportunities, a few of each, the ones with the most room first.

Strong
In good shape. Listed as a strength.
Partial
Much of it is there; some is not.
Weak
A little of it is there.
Missing
None of it, as far as IndexBot could read.

A Theme that does not apply to a site, such as freshness on a site with nothing dated, is left out rather than marked down.

What is not published

How much each Theme counts, the individual checks inside it, and where the line between one Level and the next sits are not published. A Leaderboard that published its checklist would measure who had read the checklist. What is published is enough to act on: which Themes are working and which have the most room.

Scores are estimates of readiness, not of results. A high Index Quotient means a site is easy for search engines, answer engines and AI to find, read and trust; it does not promise where the site ranks today.

IndexBot

IndexBot is the crawler behind SEO Leaderboard. It visits a website only when someone submits that domain, reads it the way a search engine would, and leaves.

What it does

  • Fetches the homepage, robots.txt, the sitemap, and up to 24 more pages, at most four at a time.
  • Reads raw HTML only. It runs no JavaScript and loads no images, fonts, or scripts.
  • Visits a domain at most once every 24 hours, however many people submit it.
  • Stores what it measured, never full copies of pages.

How it identifies itself

Mozilla/5.0 (compatible; IndexBot/1.0; +https://seoleaderboard.com/bot)

How to let it in

Bot protection often turns IndexBot away before it reads anything, and the domain then cannot be scored. If you run the site, allow the User-Agent IndexBot. In Cloudflare that is a WAF skip rule; most other tools have the same idea under a different name.

(http.user_agent contains "IndexBot")

IndexBot has no fixed IP range to allow instead, so the User-Agent is the only thing to match on. It never tries to disguise itself as a browser, and the limits above are the whole of what it asks for.

How to keep it out

IndexBot obeys robots.txt. Add this and it will not read your site, and your domain cannot be scored.

User-agent: IndexBot
Disallow: /

Questions

Write to bot@seoleaderboard.com.