Outrings
AI agents and automation

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.

3 min read
Short answer

It is a weighted average of individual checks, each of which either passed or failed against a rule someone wrote. That makes it useful for tracking your own progress and for spotting a regression. It is not a measure of quality, it does not predict traffic, and it is not comparable across tools.

How the number is built

The construction is worth knowing, because it explains both the uses and the limits.

  1. Each check produces a verdict: pass, fail, warn, or undetermined.
  2. Each check carries a weight reflecting how much it matters within its category.
  3. Category scores are the weighted proportion of passed checks, with undetermined ones excluded entirely.
  4. The overall score is a weighted combination of the categories.

Step three is the part most people would not guess and the part that keeps the number honest: a check that could not be evaluated does not count as a pass or a fail. It leaves the calculation, and the fact that it did is reported separately.

What it can honestly tell you

  • Whether you are improving. The same instrument over time is a real signal, and the primary legitimate use.
  • Whether something regressed. A sudden drop means a specific check started failing, and the finding names it.
  • Roughly how much is outstanding. 60 means substantially more is failing than at 90.
  • Where to look first. The weakest category is usually where the cheapest wins are.

What it cannot tell you

  • Whether your website is good. A beautifully written, genuinely useful site with imperfect headers scores below an empty template with perfect ones. The score measures configuration, not worth.
  • How much traffic you will get. No audit score predicts rankings. The checks correlate with the technical preconditions for ranking; that is not the same thing.
  • How you compare to a competitor scored by another tool. Different scope and weights make cross-tool numbers meaningless.
  • That everything is fine at 100. It means everything checked passed. What was not covered is a separate list, and it is deliberately published.
The last one deserves emphasis. A perfect score on a narrow audit is a weaker statement than an imperfect score on a broad one. Read the coverage before you read the number — which is why every response carries what it did not examine.

Reading it well

RangeUsually meansSensible response
90–100Fundamentals are in placeTrack for regressions; work on content instead
75–89Solid, with real gapsFix the high-severity findings; the rest can wait
55–74Several genuine problemsWork the ranked list from the top
Below 55Something structuralLook for one root cause — often client-side rendering or a blanket robots rule

Below 55, resist the temptation to work through the list. Very low scores usually have a single cause producing dozens of symptoms, and fixing that one thing moves everything at once.

The better question

Rather than "what is my score", ask "what is currently failing, and what would each fix cost". That question has an actionable answer, cannot be gamed, and does not require you to trust anyone's weighting decisions — which is why findings ranked by points recovered per unit of effort are more useful than the number they sum to.

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 with its weight, so the score can be decomposed rather than taken on trust.
  • Category scores alongside the overall figure, since the average hides which area is weak.
  • Findings ranked by points recovered per unit of effort, computed from the same weights that produced the score.
  • An explicit coverage statement, so a high score is never mistaken for a complete one.

For agents and scripts, the same measurement is at /api/v1/summary?url=yoursite.com — see the API documentation.

Related questions

Is 100 achievable?

Yes, on the checks performed — it means everything measured passed. It is a statement about coverage as much as about quality, which is why the list of what was not covered is published alongside it.

Does a higher score mean more traffic?

No. The checks cover technical preconditions — being crawlable, being indexable, being fast enough, being trustworthy. Meeting them does not produce demand for what you publish.

Why did my score drop when I added content?

New pages bring new opportunities to fail checks: a missing alt attribute, a long title, a page without a description. The score is an average, so adding weaker pages lowers it even though the site got bigger.

Should I show my score publicly?

Only if it is genuinely good. A badge showing a low number faithfully advertises the low number, and it links to a live audit so a visitor can confirm it.

Read next

All 100 guides · How every check works · API for agents