AUDIT

Why a Website Score Isn't an Audit

Ben Foord, author
Ben Foord
· 6 min read

An SEO or website score is a weighted average of the checks a tool managed to run, rolled into one number. It tells you roughly how worried to be, but not what to fix first, why it matters, who should do it or whether the fix worked. An audit is the list that answers those four questions.

Paste your web address into almost any free checker and you get a number back: 62, 78, 91 out of 100. It feels like an answer. It is closer to a thermometer reading. Useful for spotting that something is off, no help at all with what to do about it.

This post explains what sits behind a website score, walks through a real page where the number and the reality pointed in different directions, and sets out what an audit should give you instead.

What Is an SEO Score?

An SEO score, or website score, is a single number, usually out of 100, that a tool such as HubSpot's Website Grader, SEOptimer or a site-health report gives a page or a site. It is not something Google uses or shows. Almost every score is built the same way. A tool runs a set of checks, decides how much each one matters, and rolls the results into one figure. Three choices hide inside that figure:

  • Which checks it runs. One tool leans on page speed, another on titles and descriptions, another on links. The same page can score very differently from one tool to the next, and both numbers can be "right".
  • How much each check weighs. A missing image description and a page Google cannot index might move the number by similar amounts, even though one matters far more to your business.
  • What happens to checks it could not run. When a measurement times out or needs data the tool does not have, some tools leave it out of the average rather than counting it against you. The score can rise because less was measured. That is how our own engine worked out the number in the example below.

Google's Lighthouse is open about this. Its performance score is a weighted blend of several speed measurements, and the weights have changed between versions. That is fine for a developer comparing two builds of the same page. It is not a priority list for a business.

A Real Example: 84 Out of 100, and Email at Risk

Our public sample report is a real scan of a real UK business site (the owner asked us to hide the domain). When our engine scanned it on 8 July 2026, it still produced an overall score, and the page got 84 out of 100. The verdict line read "A solid page".

Look underneath and three things stand out:

  • The speed section scored 100, but most of it was not measured. Four of the six speed checks, phone speed among them, came back "not checked". The two that ran passed, so the section averaged out at a perfect score.
  • The pass rate and the score disagreed. 18 of the 23 checks that ran passed, which is 78%. The score said 84, because the checks that passed happened to carry more weight.
  • The most important problem was invisible in the number. The domain had no SPF record and no DMARC policy. In plain terms, email sent from that domain, including enquiry notifications, quotes and invoices, was more likely to land in spam. That is the kind of problem that quietly costs a small business work, and an 84 gives no hint of it.

AuditCrow's free report no longer shows a score. It leads with the fix instead.

What Is a Good SEO Score?

There is no universal answer, because there is no universal scale. A 75 from one tool and a 75 from another are measuring different checks with different weights, so the number only means something against the same tool's own bands. Google publishes bands for Lighthouse speed scores (90 and above is good, 50 to 89 needs improvement, below 50 is poor), and our PageSpeed guide covers what moves them. Most SEO graders publish nothing comparable.

A more useful question than "is my score good?" is "is anything on this page costing me enquiries, sales or visibility?" A page can score 92 and still have a broken contact form. It can score 60 because of a dozen cosmetic warnings and be doing its job perfectly well.

What a Score Is Good For

Scores are not useless. They are good at three things:

  • A first glance. A 30 on a page you thought was fine tells you to look closer.
  • Tracking one tool over time. The same tool, the same page, before and after a change. The direction matters more than the figure.
  • Developer benchmarks. A Lighthouse score helps a developer check that a change did not make a page slower.

What a score cannot do is tell you which change will matter to the business.

What an Audit Should Give You Instead

A useful audit answers four questions for every problem it finds:

  1. What to fix first. Ranked by the difference it makes to what the site is for, whether that is enquiries, sales or being found, weighed against the effort. Not by how many points it adds.
  2. Why it matters now. In plain English, tied to a result you care about. "Your enquiry emails are likely landing in spam" gets acted on. "D1: fail" does not.
  3. Who should do it. The owner, the marketer, the web developer or whoever looks after the domain. A fix with no owner tends to sit in a report for months.
  4. How you will know it worked. A re-check once the change is made, so "I've sorted it" becomes "it's live".

Take the email problem from the sample. As a score, it was a few points lost somewhere inside an 84. As an audit finding, it reads: the domain has no SPF record, so quotes and invoices are more likely to land in spam; it is an easy fix in the domain's DNS settings, usually for whoever manages your domain or email; and a re-check confirms the record is live. The SPF, DKIM and DMARC guide covers the fix itself.

Why Chasing the Number Backfires

When the score becomes the goal, effort drifts to whatever moves it fastest, not whatever matters most. Watch for these habits:

  • Fixing quick warnings first because they are quick. Ten cosmetic fixes can lift the number while the problem that costs enquiries waits.
  • Treating 100 as finished. A perfect score on one page says nothing about the rest of the site, or about whether visitors can actually book, buy or get in touch.
  • Comparing numbers from different tools. Your 78 and a competitor's 91 are not measuring the same things.

Questions to Ask of Any Report

Whatever tool or agency you use, ask:

  • What is the single most valuable thing to fix first, and why?
  • Who needs to do it?
  • How will we confirm it is live?
  • What did you not measure? Every report has gaps, and a good one says so.

If a report cannot answer those, the number on the front is not telling you much.

How AuditCrow Handles It

The free page check gives your page no score. It tells you whether a visitor can do what the page exists for, and names the fix we trust most and why it matters. The full report, with the steps for every fix, is one email away. The full platform turns each fix into a task with an owner, re-checks it once it is done and marks it Verified live.

For more on working through a report, read how to read a website audit report and what to do after your website audit.

See where your site stands
One free page scan, results in under a minute.
Scan My Site
Filed under: Audit · Seo Basics
Like reading AuditCrow?
Add us as a preferred source so more of our posts show up in your Google results.
Free first audit

Ready to see what's holding your site back?

Run one page through AuditCrow and get a clear, prioritised report on what is standing between your site and the job you need it to do.

One page free · Plain-language fixes · Your first taste of the full platform