Understanding your SEO Score
The SEO Score is the headline number on your project overview — a single 0–100 figure summarizing your site’s technical SEO and AI-visibility health. It is a weighted blend of four independent components, each measuring a different surface. The breakdown line under the score (e.g. “Issues 83 · LH SEO 100 · Bot access 84 · Search perf 60”) shows each component’s own score, and the headline is their weighted average.
This page explains what each component measures and exactly what raises or lowers it, so a number is never a mystery.
How the headline is built
Section titled “How the headline is built”| Component | Weight | What it measures |
|---|---|---|
| Issues | 50% | The severity-weighted volume of audit findings from your latest crawl |
| LH SEO | 25% | Google Lighthouse’s own SEO category score for your site |
| Bot access | 15% | Whether search + AI crawlers can actually reach your pages |
| Search perf | 10% | How your pages rank and get clicked in Google Search (from GSC) |
The headline = Σ(component score × weight) ÷ Σ(weights). A component only counts when its data exists — if you haven’t connected Search Console, Search perf drops out and the remaining three re-weight among themselves. So connecting more data sources makes the score more complete, not automatically lower.
The score floors at 5 (never 0) and is always a whole number.
1. Issues (50% — the biggest lever)
Section titled “1. Issues (50% — the biggest lever)”This is your latest crawl’s findings, turned into a score. It is not a raw count — it’s severity-weighted and dampened so one template-wide issue can’t tank the whole score.
How it’s computed:
- Each finding carries a severity. Severities are weighted: critical = 15, high = 5, medium = 1.5, low = 0.3, notice = 0 (notices never affect the score).
- Findings are grouped by check, and each group contributes
weight × (1 + log₁₀(count))— the log dampening means 100 pages with the same issue cost far less than 100× a single page (a template defect is one problem, not a hundred). - The total penalty is normalized by site size (so a large site isn’t unfairly punished for having more pages) and mapped through a curve to a 5–100 score.
What raises it: fixing critical and high findings first (they dominate the penalty), and fixing template-wide issues at the template (which clears the whole group at once). A re-crawl then recomputes the score.
What lowers it: new critical/high findings — things that block indexation or break crawlability. Low/notice findings barely move it.
The honest part: “notice” and “low” findings — and our AI scoring-input “carriers” (per-page signals that feed the AI-citation score, like entity-clarity and answer-clarity) — are excluded from the actionable count shown on the crawl card and the Issue Breakdown. They’re score inputs, not defects, so they don’t inflate the number you see. The Issues component reflects real, fixable problems.
2. LH SEO (25%)
Section titled “2. LH SEO (25%)”This is Google Lighthouse’s own SEO category score (0–100), run against your site. We prefer the desktop result and fall back to mobile. It checks Google’s documented technical-SEO basics: a valid <title>, a meta description, crawlable links, a valid robots.txt, hreflang validity, legible font sizes, tap-target sizing, and image alt text.
What raises it: the standard Lighthouse SEO checklist — make sure every page has a unique title + meta description, links are real <a href> (not JS-only), robots.txt is valid, and images have alt text.
What lowers it: missing titles/descriptions, JS-only navigation, an invalid robots.txt, or pages blocked from indexing that shouldn’t be.
Note: this is a lab score from Lighthouse, separate from your real-user field data (CrUX / your first-party tag). A perfect 100 here means the technical-SEO fundamentals pass; it does not measure rankings or speed-in-the-field.
3. Bot access (15%)
Section titled “3. Bot access (15%)”This measures whether crawlers — both traditional search bots (Googlebot, Bingbot) and AI bots (GPTBot, ClaudeBot, PerplexityBot, OAI-SearchBot, Google-Extended) — can actually reach your content. It starts at 100 and deducts a penalty per blocked bot.
How the penalty works:
- A bot blocked from the whole site (
Disallow: /): −20 for a core search bot (Googlebot/Bingbot), −5 for others. This is the serious one — a total block. - A bot blocked from real content paths (a partial block that hits indexable pages): −8 core, −2 others.
- A bot blocked only from admin/system paths (
/admin,/api,/dashboard,/login,/cart…): no penalty. Blocking admin paths is correct, healthy hygiene — we treat it as “allowed (admin paths blocked)”, not a problem.
What raises it: allow AI bots in your robots.txt (most sites should — it’s how you become citable in ChatGPT/Claude/Perplexity), and make sure no core search bot is blocked from real content.
What lowers it: a Disallow: / for any important bot, a WAF/CDN returning 403/429 to a crawler, or disallowing real content directories. Blocking only admin/auth paths will not lower this — if you see a penalty, it’s because a bot is blocked from content, not housekeeping.
4. Search perf (10%)
Section titled “4. Search perf (10%)”This reflects how your site actually performs in Google Search, drawn from your connected Search Console data. It is two parts added together: a position score plus a CTR bonus.
Position score (from your average query ranking position):
| Avg position | Position score |
|---|---|
| Top 3 (≤ 3) | 95 |
| Page 1 (4–10) | 80–95 |
| Page 2 (11–20) | 60–80 |
| Pages 3–5 (21–50) | 40–60 |
| Beyond 50 | 20–40 |
CTR bonus (your overall click-through rate): +10 at ≥ 5% CTR, +5 at ≥ 2%, nothing below 2%.
So Search perf = position score + CTR bonus (capped at 100).
What raises it:
- Rankings (the big lever): move your highest-impression queries that sit at positions 4–20 onto page 1. Better internal linking to those pages (see the Internal Linking report) and stronger on-page content are the levers. Crossing from page 2 (avg ~15–19) to page 1 (avg ~8) can lift this component 20+ points.
- CTR: if you’re just under 5%, sharper title tags + meta descriptions on high-impression pages can push you over the 5% threshold for the full +10 — no ranking change needed.
What lowers it: a high average position (ranking deep on page 2+) or a low CTR (good rankings that don’t earn clicks — usually a title/description problem).
Note: this requires Google Search Console connected with data. No GSC connection → this component is simply absent (the other three re-weight). This component is a real performance signal — a low Search perf isn’t a bug, it’s an honest read that the site ranks well for a long tail but averages a deeper position.
Why the score changed between crawls
Section titled “Why the score changed between crawls”- You fixed issues and re-crawled → the Issues component recomputes; resolved critical/high findings raise it.
- A connector synced new data → Search perf or LH SEO can shift as fresh GSC/Lighthouse data lands.
- You connected a new data source → the score becomes more complete; a previously-absent component now contributes.
- We improved a check → occasionally an audit-engine fix removes false positives (e.g. a healthy page wrongly flagged), which can move the Issues component. We note these in the changelog.
The short version
Section titled “The short version”- Want the biggest improvement? Fix critical and high findings (50% weight), especially template-wide ones.
- Want AI/search reach? Keep Bot access clean — allow AI bots, block only admin paths.
- Want a rankings lift? Work the Search perf lever — move page-2 queries to page 1 and tighten titles for CTR.
- Connect all four data sources (crawl is automatic; add GSC, Lighthouse/PSI, and Bing) so the score reflects your whole picture.