Why do two SEO tools give my site different scores?
Different checks, different weights, different definitions of failure — and why comparing scores across tools is meaningless.
Because a score is a summary of one tool's opinion about which checks matter and how much. Two tools measuring the same site disagree because they check different things, weight them differently, and draw the pass line in different places. Compare a tool against itself over time; comparing scores between tools tells you nothing.
A score is a compression, and compressions differ
Every audit reduces dozens or hundreds of observations to one number. That reduction requires three decisions, none of which is standardised:
- What to check. One tool checks email authentication; another considers that out of scope. Neither is wrong, and their scores are now measuring different things.
- How much each check is worth. Is a missing CSP worth ten per cent of the security score or two? Entirely a judgement call.
- Where failure begins. Is a 165-character meta description a fail, a warning, or fine? Different tools draw that line differently, and each will tell you their line is the right one.
Given three unstandardised decisions, identical inputs producing different outputs is the expected result, not a malfunction.
The specific disagreements you will see
| Disagreement | Typical cause |
|---|---|
| One tool says 90, another says 62 | Different scope — one is checking categories the other ignores entirely |
| Same finding, different severity | Different weighting philosophy; both defensible |
| One reports a problem the other does not | Either a scope gap, or one measured what the other could not reach |
| Scores differ on re-run of the same tool | Non-determinism in that tool. This one is a genuine defect. |
| A tool scores a blocked site well | Undetermined checks being counted as passes |
The first three are normal. The last two are the ones to be suspicious of.
What to compare instead
Scores are for tracking one instrument over time. Findings are for deciding what to do.
- Compare findings, not numbers. "No CSP header" is a fact both tools can be checked against. "Security: 68" is an opinion.
- Ask what each tool did not check. A high score from a narrow tool and a low score from a broad one are not in conflict.
- Check whether undetermined is separated out. A tool that folds unknowns into passes will always look more flattering.
- Prefer the tool that shows evidence. If it names the header value it read, you can verify it in ten seconds and the score becomes almost irrelevant.
When a disagreement is a real signal
One case is worth investigating rather than shrugging at: when two tools disagree about a fact rather than a judgement. If one says your HSTS header is present and another says it is absent, one of them is looking at something different — commonly the www host versus the apex, the redirect rather than the destination, or a CDN serving different responses to different clients. That disagreement is telling you something real about your infrastructure, and it is worth ten minutes.
What our audit reports about this
Every item below is measured directly, not inferred. Run it against your own site and the result names the exact rule or header responsible.
- Every check documented individually, with its weight, so the score can be taken apart rather than trusted whole.
- The observed value attached to each verdict, so a disagreement between tools can be settled by looking rather than arguing.
- An explicit list of what is not covered, so a high score is not mistaken for a broad one.
- Undetermined checks excluded from scoring, so obstruction cannot inflate the number.
For agents and scripts, the same measurement is at
/api/v1/seo?url=yoursite.com —
see the API documentation.
Related questions
Which tool is right?
Usually neither and both — they are answering slightly different questions. Judge them on whether each individual finding is true of your site, which is checkable, rather than on which number you prefer.
Should I use several tools?
For findings, yes: different scopes surface different real problems. For scores, no — pick one and track it over time, because that is the only comparison a score supports.
Why does my score differ from Lighthouse?
Lighthouse measures rendering performance and in-page quality in a real browser. A server-side audit measures configuration — headers, transport, DNS, crawler rules. They overlap barely and are not substitutes for one another.
Is a higher score always better?
Not across tools. A narrow tool giving you 95 and a broad one giving you 70 are consistent: the second one looked at more. Always read what was covered before reading the number.
Read next
What does a website score actually mean?
How a score out of 100 is built, what it can honestly tell you, and the four things it cannot.
ReadWhat is a deterministic website audit?
Why running the same audit twice should give the same answer, what breaks that property, and why it matters more than accuracy.
ReadWhy does my audit score disagree with Lighthouse?
Two tools measuring genuinely different things — and why expecting them to agree is the actual mistake.
ReadWhy did my website score change when I did not change anything?
The four causes of an unexplained score movement, and how to tell which one you are looking at in under a minute.
Read