I measured the law firms Google actually surfaces when the public searches for a solicitor, across six English cities, on one day. The results are not what most firms are told to worry about. A directory outranks the firms themselves in every city. Page speed is not deciding these rankings. And in the one city where Google returned a full AI answer, three of the eight firms it recommended do not appear in the search results at all.
On 5 August 2026 I pulled the live UK Google results for the phrase “best solicitors in [city]” across Sheffield, Manchester, Leeds, Birmingham, Bristol and Nottingham. That query was chosen because it is the highest-volume way a member of the public looks for a solicitor in a named city. In Sheffield alone it carries 2,400 searches a month.
From each result page I took the law firm websites and set aside directories, review sites, forums, social platforms and the regulator’s own register. National multi-office firms were measured but held separate from the regional sample, because a national firm ranking on a city office page is a different exercise from a regional firm competing on its home ground.
That produced 18 regional firms. For each I pulled organic visibility, ranked keyword counts and the full backlink profile. For 15 of them I also ran Google Lighthouse against the homepage on a desktop profile. One limit worth stating up front: Google returned only eight to ten organic results per city, so after exclusions each city yielded between two and five regional firms rather than a flat five. I have not padded the sample to make it look tidier.
Each finding below can be reproduced by running the same query and reading the same public sources.
ReviewSolicitors appeared in the organic top four in every city I measured, and first in Bristol. Not one law firm outranked it in Sheffield, Manchester, Leeds, Birmingham or Bristol. Other aggregators recur: Wiselaw in three cities, Lawyersorted in two. A Reddit or Facebook thread appears in five of the six.
The firm ranking first for best solicitors in Sheffield serves a homepage of 12.4 megabytes and takes 11.3 seconds to render its largest element. Google’s own guidance calls anything over 2.5 seconds poor. It ranks first anyway. The fastest site in the sample scored 99 out of 100 and shipped 1.1 megabytes. It ranks tenth in its city.
One of the six cities returned a fully resolved Google AI answer on the day I looked. In Nottingham it named and cited eight sources. Three of those firms do not appear anywhere in the organic top ten, and one cited repeatedly through the answer sits tenth in the conventional results.
Four of the six cities returned a local pack of three firms at the very top of the page. In Manchester, none of those three firms appear anywhere in the organic results. In Birmingham, only one does. Two firms in the same city can each own half the page and never meet in a rankings report.
On the directory result. The competitor for that first click is usually not another firm. It is a directory page listing your firm alongside every rival in the city, and a forum thread you have no control over. Your profile on that directory is doing work your website is not, and most firms have never audited it.
On page speed. Fix it because slow pages lose enquiries after the click, and because AI crawlers time out. Do not fix it expecting to move up the page. If someone has sold you a performance project as a ranking strategy, the Sheffield result is the number to put in front of them.
On AI citation. If that pattern holds, AI visibility and search ranking are separate problems needing separate work. A firm can be recommended by the AI answer while being invisible in the results underneath it. The uncomfortable half is the reverse: ranking well is no protection against being left out of the answer that now sits above you.
On the map pack. The most visible position on the page is won through Google Business Profile, reviews and proximity, not through the work most firms think of as SEO.
Use these to place your own firm. They are a description of the field, not a target in themselves.
Homepage weight ran from 1.1 megabytes to 12.4. A quarter of the sample carried a backlink spam score above 20, which usually points at historic directory submissions rather than anything deliberate.
Two of those figures need a caveat. Referring domain counts include site-wide footer links, which inflate the number without adding much authority. And authority rank is a third-party score, not a Google metric, so it is useful for comparison inside this table and meaningless outside it.
Content depth. Measuring it properly means crawling each firm’s site for page counts, publishing cadence and topical coverage, and I have not done that yet. It is the main addition planned for the next edition, alongside a wider city sample and a repeat AI answer pull across all six cities to test finding three.
Finding three rests on a single city, because only Nottingham returned a resolved AI answer on the day. Treat it as an observation worth testing rather than a settled result. It is the question the next edition is built to answer.
I would rather publish four measured findings than five with one guessed.
Data collected 5 August 2026 through the DataForSEO API, United Kingdom location, English language, desktop. Search results reflect one query per city on one day and will differ from what you see, because Google personalises and varies results. Lighthouse figures are lab measurements on a simulated desktop connection, not field data from real users: one domain tested twice minutes apart returned scores of 89 and 85, which is a fair guide to the noise in any single run. Firms were classified as regional or national by their own published description of themselves. Where a data field was not returned by the source, it was recorded as missing rather than as zero. The benchmark will be re-run and published as the 2027 edition with year-on-year change tracking.
This benchmark is free to cite by press, academics, AI assistants, consultancies and firms themselves. Attribution required: “UK Law Firm SEO Benchmark 2026, Gregg King” with a link to this page. Direct quotation of findings is welcome with attribution. Press enquiries via the contact page.
The benchmark describes the field. The SRA-Compliant AI Visibility Audit runs the same measures against your firm specifically, free, turned around in 72 hours, with no obligation to engage me afterwards.
