IndexReady's Scoring Logic Explained: How We Grade 26 Items Across SEO and GEO
Why I'm publishing the scoring logic
IndexReady is a free tool I built to automatically grade websites on both traditional SEO and Generative Engine Optimization (GEO). Enter any URL and you get two independent 100-point scores within seconds.
I'm Josef — a Kanto-based web engineer, and IndexReady is a personal side project I've been running and improving continuously.
One question people ask me more than any other is: "What exactly are you grading, and why should I trust the score?" It's a completely fair question. Blindly acting on a black-box number is how teams end up wasting weeks on the wrong fixes.
So in this post, I'm laying out the complete scoring logic IndexReady uses today. I'm not publishing the implementation code line by line, but I'm sharing every grading item, every point allocation, and the reasoning behind each design choice.
The key design decision: SEO and GEO are scored separately
The biggest call I made early on was to keep SEO and GEO as independent 100-point categories rather than rolling them into a single combined score.
Most SEO tools hand you one aggregate number. That's easy to read, but it hides something important: a perfectly optimized Google-SEO site can still be completely invisible to AI search.
A site with a flawless title tag, a clean sitemap, and healthy Core Web Vitals can still have zero chance of being cited by ChatGPT or Perplexity if it's missing an llms.txt file and blocking GPTBot in its robots.txt. Cramming that signal into a single composite score buries it.
By presenting two parallel scores, IndexReady lets you see at a glance which front you're losing on — Google or AI search.
SEO scoring: 15 items, 100 points
The SEO category distributes 100 points across 15 items. Weights are assigned based on a blend of ranking impact and implementation effort.
| Item | Points | What's checked |
|---|---|---|
| Title tag | 10 | Presence and length (30–60 chars is optimal) |
| Meta description | 10 | Presence and length (70–160 chars is optimal) |
| PageSpeed | 8 | PSI score tiers |
| Core Web Vitals | 8 | LCP, INP, CLS combined |
| Meta robots noindex | 8 | Is noindex accidentally set? |
| Heading structure | 6 | H1 presence/count, H2+ hierarchy |
| OGP tags | 6 | og:title / description / image |
| HTTPS | 6 | Is the URL served over HTTPS? |
| Canonical tag | 6 | Presence and validity |
| Image alt attributes | 6 | Ratio of images missing alt |
| robots.txt | 6 | Presence and base configuration |
| sitemap.xml | 6 | Presence and validity |
| Content length | 6 | Body word count |
| HTML lang attribute | 4 | <html lang> presence and validity |
| Viewport meta | 4 | Viewport configuration |
Why title and description top the chart
Title tag and meta description get the maximum 10 points because they directly drive click-through rate (CTR). These are the first things a searcher sees in Google's results — weak copy here means even a top-ranking page leaks traffic. And they cost almost nothing to fix, which makes any site with poor titles genuinely leaving money on the table.
Why noindex earns 8 points
Accidental noindex is one of the most common catastrophic SEO mistakes I see. Someone pushes a staging template to production, and suddenly every page is telling Google to forget it exists. The impact is enormous but the fix takes seconds — which is exactly why I weight it heavily as a safety check.
Why lang and viewport are only 4 points each
Missing these won't tank your rankings immediately. But they matter for mobile-first indexing and for international sites, so they stay on the list as baseline hygiene.
GEO scoring: 11 items, 100 points
GEO has fewer items, so each one carries more weight than its SEO counterpart.
| Item | Points | What's checked |
|---|---|---|
| llms.txt | 12 | Presence, llms-full.txt, content quality |
| AI crawler permissions | 12 | GPTBot, ClaudeBot, PerplexityBot, etc. |
| Structured data | 10 | JSON-LD presence, schema types |
| Citation link quality | 10 | Outbound links to authoritative sources |
| Clear answer paragraphs | 8 | Concise, quotable text structure |
| FAQ / list / definition patterns | 8 | FAQ structure, lists, definition blocks |
| Question-format headings | 8 | H2/H3 phrased as questions |
| Statistics & numeric data | 8 | Specific numbers and data points |
| Content freshness | 8 | Publish/update metadata |
| E-E-A-T signals | 8 | Author info, citations, dates |
| Google-recommended structured data | 8 | Recommended schema types |
Why llms.txt and AI crawler permissions are 12 points each
These two top the chart at 12 points each because they're binary blockers. If either is missing or misconfigured, your AI search exposure drops close to zero.
- Without an
llms.txt, AI systems have no efficient way to understand your site structure. - If GPTBot and ClaudeBot are blocked in
robots.txt, you're literally invisible to the crawlers building AI training data and AI search indexes.
These aren't "nice to have." They're the door being locked.
The "quotability" cluster
A big chunk of GEO scoring rewards content that's structured to be easily quoted by an LLM:
- Clear answer paragraphs: LLMs prefer short, definitional paragraphs over meandering prose.
- Question-format headings: "What is X?" headings map cleanly to how users phrase AI queries.
- Statistics and numeric data: Specific numbers act as credibility signals when an AI decides what to cite.
- E-E-A-T: Experience, Expertise, Authoritativeness, Trustworthiness — the same signals Google documented for human quality raters also matter for generative citation selection.
These elements were supporting cast in classic SEO. In GEO, they're the stars.
Three principles that shape every point allocation
Across all 26 items, I consistently apply three rules:
- High-risk misconfigurations get heavy weight. Things that can silently kill your traffic (noindex, canonical, robots.txt) carry more points than their complexity would suggest.
- Low-effort, high-impact items get heavy weight. Title and meta description are the extreme examples.
- Blocker items get the maximum weight. If getting one thing wrong invalidates everything downstream, it deserves the maximum weight.
Conversely, fine-grained optimizations get lighter weights. IndexReady is designed to help you avoid catastrophic mistakes and nail the fundamentals — not to reward obsessive micro-tuning.
How to read your results
The recommended reading order:
- Start with errors. Red items are critical — fix them first.
- Then look at warnings. Yellow items are already dragging you down.
- Compare category scores. Is your SEO or your GEO the weaker side?
- Focus on high-weight items. Limited time goes furthest on items with 8+ points.
Aiming for a perfect score is rarely the right call. Zero errors plus a passing grade on your top five weighted items is a much better target.
FAQ
Will a high IndexReady score guarantee better Google rankings?
No, and I'm explicit about that throughout the product. The score measures whether your technical foundation is in place. Sites without a solid foundation can't rank no matter how good their content is — but having a great score doesn't automatically push you to page one either. Content quality, backlinks, and domain authority still matter.
Why does GEO matter right now?
ChatGPT, Perplexity, and Google AI Overview are changing how people find information. Being cited in an AI-generated answer requires a different optimization axis than ranking in a traditional search result. Teams that invest in GEO early are picking up first-mover advantages while most competitors haven't even noticed the shift.
Will the scoring items change over time?
Yes. The items evolve as AI search evolves and as Google rolls out new guidance. The current 26 items reflect the state of SEO and GEO as of early 2026.
Try it yourself
Head over to IndexReady and drop in any URL. You'll see the complete 26-item breakdown in a few seconds — free, no signup required.
For the per-item criteria, there's also a dedicated scoring page that documents every check in reference form.