Law firm SEO agency: what evidence to ask for in the UK

Last updated: 5 August 2026

What is an agency actually claiming?

Three different things, usually in the same paragraph, and each one is verified by a different document.

The third row is the one to slow down on. An agency cannot evidence instructions won, because instructions are counted inside the firm and not inside any tool an agency has access to. A supplier who states a matter count is either passing on a figure a client gave them, which they should say, or producing a number nobody can check.

What evidence should a firm ask for, in writing?

Six items. A request for all six takes about ten minutes to write and is the cheapest filter available.

1. A Search Console export for one named client property, covering at least twelve months, showing the date range on the face of it.
2. The brand and non-brand split of that traffic. A firm's own name is the largest single search term for most established practices, and growth in it usually reflects something other than the agency's work.
3. The starting position, not only the current one. Movement from 40 to 12 and movement from 4 to 3 are both true statements and are not the same achievement.
4. The comparable client. A firm of similar size, in an area of work with similar competition, in this jurisdiction. A conveyancing result in a market town does not predict an immigration result in London.
5. The names of the people who will do the work, and whether the writing is done by the same people who take the briefing.
6. A written statement of what the agency will not do. A supplier who cannot complete that sentence has not defined the engagement.

Item two removes more inflated claims than the other five together, and it costs one line in an email.

What does a refusal to provide it mean?

It depends entirely on which refusal it is, and the difference is easy to test.

A legitimate refusal sounds like: we cannot name that client without asking them, and here is a redacted export with the property name removed but the dates, the trend and the brand split intact. Or: we will ask them, and if they agree you can speak to them directly. Client confidentiality is real, and an agency that respects it is showing you how it will treat your data.

An evasive refusal sounds like: we do not share client data. Full stop, for every client, in every form, redacted or not, with no offer to ask anybody. That position cannot be about confidentiality, because a redacted export identifies nobody. It is a statement that no such export exists in a form that would help their case.

The test that separates the two: ask whether they will request permission from one client. A supplier with a result they are proud of will make that call. A supplier without one will explain why the request is unreasonable.

The same logic applies to a case study with no dates on it. A study that says traffic rose by a proportion, with no period stated and no baseline, is not a redacted document. It is an unfalsifiable one, and the missing dates are the point rather than an oversight.

Which claims cannot be evidenced at all?

The guarantee, and every version of it.

Eight of the 40 suppliers we read on 3 August 2026 claimed a guarantee of some kind. Nobody controls Google's results, so the guarantee is either about something else — a volume of work delivered, a refund — or it is a promise about an outcome the supplier does not own. Both are worth reading closely, because the wording usually reveals which.

It is worth being precise about why this matters more here than in most industries. This audience cross-examines for a living. A supplier who offers certainty about a ranking has either not understood their own service or has decided the buyer will not check, and both conclusions are damaging to everything else in the same document.

The related figure is that 15 of the 62 questions buyers ask on these searches are price questions, and 12 of the 40 suppliers publish a price at all. A market that will promise a position it cannot control while declining to state a fee it sets itself is telling a firm something about where the certainty actually lies.

Who appears when you search for a law firm SEO agency?

A mixture, and only some of it is agencies.

Two of these three are agencies a firm could instruct and both publish a price, which puts them in the minority of the 40 we read. The third is a legal publisher that ranks on this search because it publishes substantial material about running a firm, not because it sells this service.

⚠ We record whether a price appeared on the page, not what it was. The figures our tooling collected mixed genuine fees with case values and fragments, so attributing a number to a named business would publish something unverified about a competitor.

That mixture is worth noticing before a shortlist is built from a search result. A page-one listing is a statement about who publishes well, and publishing well is a service some of these organisations sell and others simply demonstrate.

What can we evidence, and what can we not?

We would fail item four of our own list for most readers of this page, and it is better to say so here.

We do not publish client names, case studies or before-and-after figures. No firm on our books has agreed to have its search data published, and we will not anonymise a result so thoroughly that it becomes the unfalsifiable document described above. That means a firm applying the six-item test to us gets a straight answer on items one to three only after a client has been asked and has agreed, which takes a few days rather than a click.

What we do publish instead, and what can be checked without our permission:

• The benchmark itself. 40 named suppliers, read in full on 3 August 2026, with the counts stated so anybody can repeat the exercise on the same pages.
• Our prices, as fixed figures rather than ranges, on the page rather than after a call.
• Our own demand data, city by city, with the method attached: Google Ads volume multiplied by cost per click, which is what the traffic costs to buy at Google's own prices.

That is a weaker form of proof than a named client result and we are not going to pretend otherwise. It has one advantage: it is repeatable by the reader, which no case study is.

Where is another supplier the better choice?

Three situations, said plainly because a comparison page that always concludes "us" is discarded by this audience within a paragraph.

When the firm needs instructions this quarter. Nothing in search works that fast. We do not manage ad spend, so the correct answer is a paid specialist and we will say so rather than sell a slower thing.

When the firm wants a named, referenceable client result before it signs anything. An agency that publishes case studies with client consent already in place can satisfy that on the first call. We cannot, and a buyer for whom that is the deciding test should use it.

When the constraint is fee-earner capacity rather than enquiries. Better visibility produces work the firm then declines. No supplier fixes that, and one who takes the engagement anyway is selling something the firm does not need.

What did we find when we read 40 legal-marketing suppliers?

We read 40 legal-marketing suppliers in full in August 2026 — including several on this page. Twelve published a price. One offered work with no minimum term.

The free report gives you the full count, named, so you can repeat it rather than take ours. It also shows how many people search for your practice areas in your town, and what those clicks cost to buy at Google's own prices.

Get the report — five questions, about two minutes →

No call required. Our own prices are on that page.

Which other law firm marketing pages should I read?

If the search is really for a broader supplier: marketing agency for law firm and law firm marketing company.

If one person would suit better than a firm: law firm seo consultant and law firm seo experts.

On a channel where the claims are checked differently: law firm marketing social media.

Frequently asked questions