Free Website Audit Tools: What Each One Actually Measures

Why running five tools gives you one answer

Search for a website audit and you will be handed a list of ten free tools. Run all ten and a strange thing happens: they mostly agree, they mostly tell you about images and scripts, and none of them explain why your enquiries stopped.

There is a structural reason. Most free audit tools are built on the same engine — Google’s Lighthouse — or are measuring the same page-load waterfall from a slightly different location. They are not ten opinions. They are one opinion in ten wrappers, and it is an opinion about how the page is built, not about whether it works.

Ten tools measuring the same thing is not a second opinion. It is the same opinion, ten times.

Google PageSpeed Insights & Lighthouse

What it genuinely measures: performance, accessibility, best practices and a basic technical SEO category, on one URL at a time. It combines a lab test run on Google’s infrastructure with field data from real Chrome users where enough of them exist for your site.

What it is genuinely excellent at: Core Web Vitals. If you want to know your Largest Contentful Paint, layout stability and main-thread responsiveness, this is the source, and everything else is a copy of it. The diagnostics are specific enough to hand to a developer.

Where people misread it: two places. First, the SEO score is a hygiene check — does the page have a title, a description, is it crawlable — not an assessment of your rankings, content or competitors. A perfect 100 there says almost nothing about whether you will rank. Second, the whole report is about page construction. It will never tell you that a visitor had no way to contact you.

The trap

Because it produces a big colourful number, PageSpeed gets treated as a verdict on the site. It is a verdict on the page’s engineering. Those are different things, and confusing them is how businesses spend three months shaving 400 milliseconds off a page that generates no leads at either speed.

GTmetrix and the speed testers

What they measure: page load, from a chosen test location, usually presented as a waterfall chart showing every request in order. Several are built on Lighthouse underneath.

Genuinely useful for: the waterfall. When a page is slow and you do not know why, seeing every request in sequence — which one blocked, which one took two seconds, which third-party script arrived late — is more diagnostic than a score. Choosing a test location near your actual customers also matters more than people realise.

The overlap: if you have already run PageSpeed Insights, most of the summary numbers here will tell you the same story. Use these when you need to debug a specific slowness, not as a second opinion on whether you are slow.

Google Search Console

What it measures: how Google actually sees, crawls, indexes and ranks your site — using Google’s own data rather than an outside guess.

Why it is in a different category: every other tool on this page is inferring things from the outside. Search Console is the record. Which pages are indexed and which are not, which queries you appear for, where you rank, which pages have errors, whether your sitemap was read. If you only ever set up one thing, set up this.

The limitations: it only works for sites you can verify ownership of, so it is useless for looking at a competitor or a prospect. It reports on the past rather than diagnosing the present, and it says nothing about the page as an experience. It also has nothing to offer on day one — it needs weeks of data before the interesting parts fill in.

Site crawlers

What they measure: your whole site rather than one page. Broken links, redirect chains, duplicate titles, orphan pages, missing descriptions, thin content — at scale.

When they earn their place: the moment you have more than about thirty pages. The problems a crawler finds are ones you literally cannot see from a homepage check, because they live in the relationships between pages. On a five-page brochure site, they are overkill.

The catch: the free tiers are limited, the output is long, and the reports are written for specialists. A first-time user typically gets a list of six hundred issues with no sense of which three matter. That is not a flaw in the tool, but it does mean it is the wrong first tool.

Accessibility checkers

What they measure: contrast, labels, alt text, keyboard navigation, heading structure — whether the site can be used by people using assistive technology.

Why bother: the ethical reason is sufficient on its own. The commercial one is that accessibility failures and conversion failures overlap heavily. Unlabelled form fields, poor contrast on your primary button and a heading structure nobody can follow cost you customers who have no disability at all.

The gap all of them share

Line the whole category up and something becomes obvious. They measure how well the website is built. Not one of them measures whether it does its job.

None of the mainstream free tools will tell you:

• Whether there is a visible phone number a visitor can tap.
• Whether there is a WhatsApp or instant messaging path — the default for a large share of Indian small business enquiries.
• Whether there is a call to action above the fold, before anyone scrolls.
• Whether there is a lead capture form at all.
• Whether your offer is stated in plain words rather than implied by a photograph.
• Whether you have testimonials, an About section, a location or policy pages — the signals that tell a stranger you are a real business.

Those six things decide whether a visitor who already wants what you sell becomes an enquiry. A site can pass every technical audit in existence and fail all six. In my experience it usually does, because nobody has ever measured them, and what nobody measures nobody fixes.

Check the half nobody else measures →

The three-tool stack I would actually run

1. Google Search Console — set it up once, on every site you own. It is the only source of truth for how Google sees you, and it takes weeks to become useful, so the cost of delay is real.
2. PageSpeed Insights — for performance, accessibility and technical hygiene. Run it on your homepage and your two most important pages. Ignore anyone who tells you the score must be 100.
3. A conversion and trust check — the missing half. Whether a visitor can act, and whether they believe you are real.

Add a crawler when you pass thirty pages. Add a speed tester when something is slow and you need the waterfall to find out why. Everything beyond that is a fourth number about a thing you have already measured twice.

Where our scanner sits in this

The AI Growth Score is the third item on that list, with the first two folded in where they fit. It runs 41 checks across five categories — website experience, performance, SEO fundamentals, conversion readiness and trust — each worth 20 points, and returns a score out of 100 in about thirty seconds. Google’s own performance measurements are layered in as part of the performance category once your score is already on screen.

Two design decisions worth knowing about. Anything the scan cannot verify is excluded from the score entirely rather than counted as a failure — so you are never marked down for something the scanner simply could not see. And you do not need to enter an email to get your score; details are only asked for if you want the full report with the recommended fix for every opportunity found.

It does not replace Search Console, and it is not trying to. It covers the two categories the rest of the free tooling has quietly agreed to ignore.

FAQ