SEO for law firm websites in Australia: the technical layer
SEO for a law firm website is mostly mechanical: whether a page can be fetched, rendered, understood and linked to. Google generated an AI answer on 39 of the 54 highest-value legal searches we measured on 3 August 2026, and a page that does not render cannot be quoted inside one.
Last updated: 5 August 2026
Both of our scopes include this layer: A$1,900 for the sprint, A$4,200 for the build. Fixed, published, no minimum term.
Can the page be fetched at all?
Often not, and it is the failure that wastes the most money: everything downstream depends on it.
Four things block a page from being retrieved, in rough order of how often we find them on a law firm site:
- A staging site never switched back. A firm's site is rebuilt, the instruction telling search engines to stay away travels with it to the live domain, and nobody looks again. Every page is invisible and everything else on this list is irrelevant.
- A sitemap listing pages that no longer exist, or omitting the ones that matter. The file is a statement of what the firm wants found, and on most firm sites it is generated once and never read.
- A canonical tag pointing somewhere else. Two versions of the same practice-area page, one pointing at the other, and the wrong one wins.
- Deliberate exclusion, used correctly. Our own conversion page is marked not to index, on purpose — it is a form and a price list. A firm's enquiry-confirmation pages belong in the same category.
Does the page exist before JavaScript runs?
This is the check most worth running, and it takes about a minute.
A site whose heading, body text and internal links are assembled in the visitor's browser can look perfect to a partner and be close to empty to anything reading it mechanically. The check is to look at what the server sends before any script runs.
Two ways, same test. In a browser, open the page source rather than the inspector — the inspector shows the assembled page, the source shows what arrived — and search it for your own heading. If it is not there, it was not sent. From a command line, the same test is one request piped to a search for the heading tag, which is the gate we run against our own pages before they ship.
This matters more than it did: the answer layer sits above everything. Across the whole keyword set we measured on 3 August 2026, 863 of 1,235 keywords carried an AI answer. A page that is assembled after the fact is not in the pool those answers are drawn from.
What actually makes a law firm site slow?
Four things, and none of them are the ones usually blamed.
Photography first: partner portraits delivered at print resolution and placed on a page at a tenth of the size. Then embedded third-party blocks — chat widgets, review carousels, booking scripts — each fetching its own code from elsewhere before the page settles. Then fonts, loaded in several weights when the design uses two. Then a layout that shifts under the reader as those pieces arrive, which is the part a visitor notices.
No threshold figures appear here: Google's own measurement tools report against them, and any number we quoted would not be traceable to our source pack. All four causes are fixable without rebuilding anything, which is why they sit in the sprint rather than the build.
What should the structured data say, and what must it never say?
Four types, and one absolute prohibition.
| Markup | Use it for | The failure to avoid |
|---|---|---|
Organization | The firm itself, once, site-wide | Several conflicting versions across templates |
Service or equivalent | What the firm does, per practice area | A description that does not appear in the visible text |
FAQPage | A question block that is genuinely visible on the page | Questions in the markup that no reader can see |
BreadcrumbList | Where the page sits in the site | A path that does not match the real one |
The prohibition: never mark up a rating, a review count or an aggregate score the site does not display, and never claim a physical presence that does not exist. We apply it to ourselves — no rating markup, because we hold no ratings, and no local-business markup, because we serve three countries remotely. Invented markup is ignored at best and a manual penalty at worst.
Everything in the markup has to be in the visible text. That single rule prevents most of the ways this goes wrong.
How should the pages point at each other?
Upward, from the specific to the general, with anchor text that says where it goes.
A practice-area page is the hub. Every page answering a question inside that area links up to it, in a sentence rather than in a footer list. The hub links back down to the questions worth reading before somebody rings. Pages about named fee earners link to the areas they work in. Nothing important should be more than a couple of clicks from the front page, and nothing at all should be reachable only from a sitemap.
Anchor text carries the meaning. A link reading "conveyancing in Canberra" tells a machine what is at the other end. A link reading "read more" tells it nothing, and there is no second signal to fall back on.
Which of these can a firm check without a supplier?
Most of them, in about twenty minutes.
| Check | How to run it | What a failure looks like |
|---|---|---|
| Is the page fetchable | Search Google for a distinctive sentence from the page, in quotes | No result, or a result with no description |
| Does it render | View source, search for your own heading text | The heading is absent from the source |
| Is the sitemap honest | Open the sitemap file and read the list | Addresses that no longer resolve, or missing practice areas |
| Is the markup honest | Compare each claim in the markup against the visible page | A rating, a review count or an address the page never shows |
| Do the links point up | Open any question page and look for a link to its practice area | The only route back is the navigation bar |
None of it needs a paid tool, and a firm that runs these before taking a proposal can tell whether a supplier's findings match what it already knows.
What does this layer not do?
It removes obstacles. It does not create demand.
A fast, correctly marked-up, perfectly rendered page about nothing anybody searched for will rank for nothing, and no amount of technical work changes that. The demand is fixed by the market: in Canberra, 860 people a month search for a conveyancer at around A$13.63 a click, and in Hobart 480 a month search for a family lawyer at around A$7.67, both measured 3 August 2026. Fixing the mechanics decides whether a firm is eligible for that traffic. It does not decide whether the traffic exists.
We also do not manage ad spend, do not fix what happens after an enquiry arrives, and give no advice about professional obligations. Australia has no rule requiring a firm to publish anything on its site, so what goes on it is a commercial decision for the firm and its own advisers.
What this is worth where you are
In Brisbane, 8,800 people a month search for family law, and Google charges around A$24.29 for one of those clicks. We hold the same figures for 52 cities.
The free report gives you yours, plus which of the checks above your site currently fails, and what 40 suppliers charge.
Where should I read next?
If the prior question is whether any of this is worth doing: does SEO work for law firms? and does a law firm need a website?.
On what to put on the pages once they render: what is the best way to advertise a law firm?.
On the two ranking questions people confuse with this one: what are the ranks in a law firm? covers internal seniority, and how to get ranked in Legal 500? covers the directory.
Frequently asked questions
What is the technical layer of law firm website SEO?
Whether a page can be fetched, whether it exists before JavaScript runs, how fast it settles, what its structured data asserts, and how the pages point at each other. It is the part that decides eligibility rather than the part that decides demand.
How can I tell whether my pages actually render?
View the page source rather than the browser inspector, and search it for the wording of your own heading. The inspector shows the assembled page; the source shows what the server sent. If the heading is not in the source, it was not sent.
Why does rendering matter more now?
Because the answer layer sits above the ordinary results. Across the keyword set we measured on 3 August 2026, 863 of 1,235 keywords carried an AI answer, and 39 of the 54 highest-value legal searches did.
What makes a law firm website slow?
Print-resolution partner photography, embedded third-party blocks such as chat widgets and review carousels, fonts loaded in more weights than the design uses, and a layout that shifts as those pieces arrive.
What structured data should a law firm site carry?
Markup for the firm itself, for what it does, for a question block that is genuinely visible, and for the page's position in the site. Everything asserted in the markup has to appear in the visible text.
What structured data must never appear?
A rating, a review count or an aggregate score the page does not display, and any claim of a physical presence that does not exist. Our own pages carry neither, because we hold no ratings and serve three countries remotely.
Does fixing the mechanics make a firm rank?
No. It removes obstacles rather than creating demand, and a fast, correctly marked-up page about a subject nobody searches for will rank for nothing.