How your SEO score is calculated.
Five weighted pillars, severity-weighted checks, and an A–F grade. The part everyone misses: a pillar we can't measure is removed from the maths, not scored zero — which is why connecting Search Console genuinely moves your number.
Your SEO score is one number out of 100 with a letter grade beside it. It is not a vibe — every point is traceable to a specific check on a specific page, and every fix card tells you exactly how many points it is worth. This page explains the arithmetic, because a score you can't interrogate is a score you can't trust.
The five pillars
Every check belongs to one of five pillars. Each pillar is scored 0–100 on its own, then blended by weight into your overall score.
| Pillar | What it's asking | Weight |
|---|---|---|
| Crawlability & indexing | Can search engines reach your pages, and are you accidentally telling them not to? | 25% |
| On-page essentials | Titles, descriptions, headings, canonicals — the basics on every page. | 25% |
| Technical & performance | HTTPS, mobile viewport, server response time, Core Web Vitals. | 20% |
| Structured data & richness | Schema, social tags, alt text, internal links, content depth. | 15% |
| Search visibility | Are your tracked queries actually ranking? This one comes from Search Console. | 15% |
The two 25% pillars are first for a reason. If Google can't crawl you, nothing else you do matters; and if your titles are missing, no amount of schema will save you.
Severity weighting
Within a pillar, checks are not equal. Each one carries a severity, and severity is a multiplier on how much that check moves its pillar:
| Severity | Weight | Shown in the report as |
|---|---|---|
| High | ×3 | Error |
| Medium | ×2 | Warning |
| Low | ×1 | Notice |
So a pillar's score is the severity-weighted average of its checks. Blocking crawlers (High) hurts three times as much as a missing lang attribute (Low), which is roughly the ratio by which they hurt in real life. Every check and its severity →
Passing isn't always pass/fail
Most checks are binary — you either have a canonical tag or you don't. But some are graded, scoring between 0 and 1 on each page, so that a partial fix moves your score instead of counting for nothing:
- Title length — 45–60 characters scores full marks; 30–70 scores partial; anything else scores badly.
- Description length — 70–155 characters is full marks, 50–180 partial.
- Alt text — you score the proportion of images that have it. Half your images covered scores a half.
- Internal links — three links is full marks; one link scores a third.
- Content depth and server response time — both banded the same way.
A check is then scored across every page it applies to, and shown as passing only when it averages 90% or better. That's a deliberately high bar: a title-length check sitting at 70% across your site is not a pass, it's a warning with a real number of points behind it.
It also means a check can be a fix card while most of your pages are fine. The card says so — “12 of 40 pages affected” — and lists every URL.
The A–F grade
The letter in the corner of the gauge is a straight mapping from the score:
| Grade | The verdict we show | Score |
|---|---|---|
| A | Excellent — you're ahead of the pack | 90–100 |
| B | Good — you're on the right track | 80–89 |
| C | Good, shading into needs-work | 65–79 |
| D | Needs work — you're close | 50–64 |
| F | Critical — needs attention | Below 50 |
A good score is 90+. That's the line the report itself uses, and it's the number the “points to gain” line is measuring you against. Most sites that have never been audited land somewhere in the C–D range on their first run, and the first three fix cards usually account for most of the gap.
An unavailable pillar is reweighted out, not zeroed
This is the most important paragraph on this page, and the one people get wrong.
Two things in the score can be genuinely unmeasurable rather than merely bad:
- Search visibility — we can't know your rankings until you connect Google Search Console.
- Core Web Vitals — a low-traffic site has no real-user performance data for Google to give us. And an Essential-depth audit skips the step by design.
When we can't measure something, we remove it from the calculation entirely. We do not score you zero for it. The remaining pillars are rescaled so their weights still add up to 100%.
// the Search-visibility pillar is dropped, not failed Crawlability 25 // → 29.4% of the rescaled total On-page 25 // → 29.4% Technical 20 // → 23.5% Structured data 15 // → 17.7% Search visibility — unavailable, removed // weights actually used: 85 → rescaled to 100
Not because we hand you free points — because it switches a pillar on. Your score stops being “the best we could measure” and becomes the real thing, with 15% of it riding on whether your queries actually rank. It's free, it takes a minute, and it's the single most honest change you can make to this number. Connect it →
The same logic runs in reverse. Connecting Search Console can send your score down — if your rankings are poor, you've just switched on a pillar you were failing invisibly. That's the point. A score that only counts the things you're good at isn't measuring anything.
Where “+N points to gain” comes from
Every fix card carries a green +N pts chip, and the score hero adds them all up: “A good score is 90+. You have +23 pts to gain from the fixes below.”
That number is not a guess or a flat “+4 for everything”. It is exactly what your overall score would rise by if you fixed that one check completely and changed nothing else. It falls out of three things:
- Which pillar it's in — a crawlability fix is worth more than a structured-data fix, because crawlability carries more weight.
- How severe it is — a High-severity check is worth three times a Low one inside the same pillar.
- How badly you're failing it — a check failing on 30 pages is worth far more than the same check failing on one. Fix half of them and the points on the card go down, because you've already banked half of them.
Two consequences worth internalising:
- The fix list is already in the right order. Cards are sorted by points, highest first. Working top-down is the optimal strategy — there's no cleverer sequence to find.
- The points are honest, not motivational. If a fix is worth 1 point, the card says 1 point. We'd rather tell you a job isn't worth doing than pad the number to make the list look exciting.
One presentation detail: Core Web Vitals has no fix card. Its points are real and they're inside your Technical pillar and the total — but they're shown as a chip on the Core Web Vitals section instead of a card on the board, so the same points aren't listed twice. Core Web Vitals →