Search "free seo audit" on Google in Canada and, as of this writing, every one of the nine organic results on the first page is a tool asking for your website address. They compete on volume: one snippet promises 172 checks, another more than 310. That is the market's answer to the question: a form, a score, green and red dots. In SEO, that score has its uses, as long as you know what it measures.
An automated tool is excellent at finding errors that have one right answer. A page has a title or it doesn't. It blocks crawlers or it lets them through. The rest, meaning why a site has stalled, needs data the tool doesn't have and judgment it can't exercise.
This article draws the honest line: what a free tool catches reliably, what its score hides, and what only a person with access to your Search Console can see. If "indexing" or "query" feel fuzzy, our SEO explained without jargon covers the basics.
The short answer, for the busy
A free SEO audit tool reads your site the way a crawler does and checks what can be checked without context: whether a title and meta description exist, a forgotten noindex tag, the canonical URL, the HTTPS certificate, broken links, and speed as measured by Lighthouse in a simulated test. Those checks are reliable but binary: Lighthouse flags a missing description, never a badly written one. The overall score hides what matters most. A red score can point at a detail your real visitors never notice, and a green score says nothing about demand, because a flawless site nobody searches for stays invisible. The rest lives in Google Search Console, open only to the site's owner and the people they invite: the queries already showing your pages, the pages sitting just below the top ten, two of your own pages competing for one query, and the pages Google crawled and then left out of its index.
What the tool catches, and you should fix
The errors these tools find are often silent and serious. A noindex tag left in place after a new site goes live can pull every page out of Google's results, and nobody in the company notices until the phone rings less. An expired HTTPS certificate puts a warning in front of every visitor. A page deleted without a redirect becomes a dead page your customers find before you do.
This is where automation beats a person: it doesn't get tired on page 40, and it runs the same check the same way every time.
It still pays to read what each check actually verifies. The documentation for Lighthouse, the Google audit tool behind PageSpeed Insights, is blunt about it. Its meta description audit fails if the element is missing or empty, and the page states that Lighthouse doesn't evaluate the quality of your description. Its indexing audit only looks for tags and headers that block all search engines at once. A green dot means "the element exists", not "the element does its job". A three-word description passes as easily as one that makes people want to click.
What the score hides
The number out of 100 is the most visible part of an audit, and the most misleading. Take performance, often the lowest score.
Lighthouse's Performance score is a weighted average of five metrics. In version 10, Total Blocking Time counts for 30%, Largest Contentful Paint and Cumulative Layout Shift for 25% each, First Contentful Paint and Speed Index for 10% each. Each metric becomes a score based on where it falls against real sites tracked by the HTTP Archive. And the test itself is a simulation: a mid-tier device on a mobile network, run from a Google data centre. The PageSpeed Insights FAQ even has an entry for the person who watches the score move between runs without having touched the page.
So your site can score orange or red in the simulated test and still give real visitors a good experience. The same FAQ admits it: good lab data does not necessarily mean real-user experiences will also be good, and the two sometimes contradict each other. What counts is the timing your actual visitors get, when your site has enough traffic for Google to publish it. Our piece on site speed and Core Web Vitals explains how to read both.
The opposite mistake costs more. A site can score 100 across the board, fast and technically clean, and get no visits from Google because nobody searches for what it says, in the words it uses. No tool that only reads your pages can know that. It measures supply, never demand.
What only Search Console shows
Google does publish demand, but only to the site's owner. Search Console requires you to prove you own the site before it shows anything, and only an owner can then grant access to others, as a full or restricted user. That is the clean line between the free tool and an audit done by a person: the tool sees your site from outside, Search Console shows how Google sees it from inside.
The Performance report shows clicks, impressions, click-through rate and average position, by default for the last three months, grouped by query or by page. Average position is the topmost position your site held for each query, averaged across all the queries where it appeared. Three readings beat any score.
First, the queries already showing your pages. They often reveal that customers search with different words from the ones on your site, or that a page you considered minor collects more impressions than your home page. Second, the pages sitting just below the top ten. They get impressions and almost no clicks, and a rewritten title or an added section can be enough to lift them. It is often the best-paying work in SEO, and it is invisible from outside. Third, pages with plenty of impressions and few clicks despite a good position: either the title doesn't pull, or Google's AI Overview answers in your page's place. The two causes call for different fixes, and you have to look at the results page itself to tell which.
Two of your pages competing for one query
The clearest case we have comes from our own site. In peich.xyz's Search Console, two of our French articles, one on the privacy incident register and one on the incident response procedure, alternated around position 19 for the query "incident de confidentialité". Each article was fine on its own. An audit tool would have given them the same respectable score. The problem only existed between them: Google kept switching from one to the other, and neither one won. We have since changed the title of one of them.
The report helps you spot this if you know where to look. In the chart, two results from the same site for one query count as a single impression. Filter on the query, open the Pages tab, and you see two URLs splitting the impressions. The fix isn't technical: give each page its own query in its title, and link the two so Google understands which answers what. It is an editorial decision, and it requires knowing what each page is supposed to sell.
The pages Google read and set aside
The Page indexing report lists the URLs Google knows about but hasn't indexed, with a reason for each. Two statuses deserve a careful read. "Crawled - currently not indexed": Google read the page and didn't keep it, and it may or may not be indexed in the future. "Duplicate, Google chose different canonical than user": you marked a page as the reference version, and Google preferred another.
Google doesn't say why a crawled page stays out. That is where judgment comes in: is the page too thin, nearly identical to another, or just too new to worry about? Google also plays down the report itself: if your site has fewer than 500 pages, its documentation says, you probably don't need it, and you are better off checking key pages one by one with the URL Inspection tool. The same page warns against expecting every URL to be indexed, only the canonical ones. For most small business sites, the useful question isn't "how many pages are excluded" but "are my five important pages indexed, under the right address".
Where to start
Do both, in order. First the tool, to clear out the binary errors. Our free audit does that job: in about a minute, it hands you a report on your site's speed, search, security and compliance. It reads the page you give it, using Lighthouse and, when Google has enough of it, real-visitor data. Fix what is red and has one right answer, and ignore the simulated speed score without guilt if your real visitors are in the green. Website security deserves the same reading: HTTPS and updates first, the rest after.
Then open your Search Console with someone who knows how to read it. An hour on queries, pages near the top ten and the indexing of your key pages teaches you more about your SEO than ten automated audits. If you don't have Search Console set up yet, that is the first job.
Written byXavier Peich