Best in Show · The Dog on the Table agency index · August 2026 edition

Best in Show: I judged 50 UK agencies against my published visibility framework. Nobody smashed it.

The mean score was 40.0/100. Not one agency met the Standard's Strong bar. I scored 50 UK marketing and SEO agencies on the public-page checks in the Modern SEO + AI Optimisation Standard.

0/100

Average score across all 50 agencies

0

Agencies that reached Strong (70+)

0/50

Blog posts with no credentialed expert quote

0/50

Now publish an llms.txt for AI crawlers

Scored 19th August 2026 against the Modern SEO + AI Optimisation Standard: the 13 of 27 checks observable from public pages, re-based to 100. Full methodology.

What I did

I scored 50 UK marketing and SEO agencies against the Modern SEO + AI Optimisation Standard. I assessed the 13 of 27 checks visible on public pages, covering 52 of 100 weight points, then re-based each result to 100. Process and off-site checks stayed out. On 19th August 2026, deterministic parsing and an AI reading covered each agency's homepage, about page and one recent blog post. I verified every blog post URL by hand.

The yardstick: my published Standard

I wrote and published the Modern SEO + AI Optimisation Standard on this site. It draws on named industry sources and 17+ years of hands-on search work. Every pillar cites its sources.

It isn't gospel. It is one practitioner's open codification of what should and shouldn't be done in SEO, GEO and AEO in 2026 and beyond. This ranking measures agencies against my published opinion.

The Standard has 10 pillars and 27 weighted checks, tiered as wins, shoulds and coulds. Scores run from 0 to 100 across five bands: Leading, Strong, Emerging, Exposed and Absent. The rules are public.

Read the full Standard. Score your own site against all 27 checks, free and with no sign-up.

Read the Standard → Score your own site free

Where 50 agencies landed

Not one agency reached Strong, and none reached Leading. Most landed in Exposed: 30 of 50. The mean was 40.0/100. The Standard describes Exposed plainly: "Visible to classic search at best, largely invisible to AI answers."

Leading 85+

0

Strong 70+

0

Emerging 50+

9

Exposed 30+

30

Absent 0+

11

Band thresholds and labels are the Standard's own. Nobody reached the top two bands.

All 50, ranked

70 · Strong

50 · Emerging

30 · Exposed

Top 10, named below Ranks 11 to 50, anonymised (city and band on hover)

View as a table

Rank	Agency	City	Band	Score
1	Opace	Birmingham	Emerging	67.3
2	Rapid SEO London	London	Emerging	64.4
3	SALT.agency	Leeds	Emerging	60.6
4	Vixen Digital	Brighton	Emerging	58.7
5	Get Found	Birmingham	Emerging	53.8
6	Dark Horse	Manchester	Emerging	52.9
7	StudioHawk UK	London	Emerging	50
8	Hallam	Nottingham	Emerging	50
9	Blue Array	Reading	Emerging	50
10	Sierra Six Media	London	Exposed	49
11	Agency #11	Manchester	Exposed	48.1
12	Agency #12	Sussex	Exposed	48.1
13	Agency #13	London	Exposed	47.1
14	Agency #14	Nottingham	Exposed	47.1
15	Agency #15	London	Exposed	47.1
16	Agency #16	Leeds	Exposed	46.2
17	Agency #17	London	Exposed	45.2
18	Agency #18	Sheffield	Exposed	45.2
19	Agency #19	London	Exposed	45.2
20	Agency #20	London	Exposed	45.2
21	Agency #21	London	Exposed	44.2
22	Agency #22	London	Exposed	43.3
23	Agency #23	London	Exposed	43.3
24	Agency #24	Birmingham	Exposed	42.3
25	Agency #25	Leeds	Exposed	42.3
26	Agency #26	UK	Exposed	42.3
27	Agency #27	Derby	Exposed	41.3
28	Agency #28	Manchester	Exposed	40.4
29	Agency #29	London	Exposed	39.4
30	Agency #30	Brighton	Exposed	39.4
31	Agency #31	London	Exposed	37.5
32	Agency #32	Manchester	Exposed	37.5
33	Agency #33	Brighton	Exposed	36.5
34	Agency #34	London	Exposed	36.5
35	Agency #35	Manchester	Exposed	36.5
36	Agency #36	Manchester	Exposed	35.6
37	Agency #37	Brighton	Exposed	34.6
38	Agency #38	Warrington	Exposed	33.7
39	Agency #39	Henley-on-Thames	Exposed	32.7
40	Agency #40	London	Absent	27.9
41	Agency #41	Leeds	Absent	26.9
42	Agency #42	Salisbury	Absent	26
43	Agency #43	Manchester	Absent	24
44	Agency #44	Birmingham	Absent	24
45	Agency #45	Manchester	Absent	24
46	Agency #46	Sheffield	Absent	22.1
47	Agency #47	Manchester	Absent	21.2
48	Agency #48	London	Absent	19.2
49	Agency #49	London	Absent	14.4
50	Agency #50	London	Absent	11.5

The four patterns that decide the scores

Nobody quotes an expert

44 of 50 blog posts include no credentialed expert quote. That gap is hard to miss.

Agencies tell clients to show expertise, then publish posts that quote nobody. The advice is sound. The example needs work.

44/50

posts with no credentialed expert quote

3 yes3 partly44 no

Claims without numbers

43 of 50 posts have no attributed statistics. The supporting source is missing.

A number alone isn't enough. It needs a named source that a reader or system can check. For 43 posts, it isn't there.

43/50

posts with no attributed statistics

2 yes5 partly43 no

No ground truth to quote

A genuine canonical fact base gives an AI crisp, fact-dense pages it can quote. Only two of 50 agencies maintain one. The list stops there.

For the other 48, an AI has no solid source for basic facts about the agency itself. The agency exists. The quotable version doesn't.

2/50

maintain a genuine canonical fact base about themselves

2 yes36 partly12 no

The open door with nothing behind it

23 of 50 agencies publish an llms.txt file, while only two block any AI crawler in robots.txt. The intent is clear: this industry wants AI systems to read its work. The door is open.

Some files are plugin-generated defaults, including ones created automatically by Rank Math. Their content mostly fails the citability checks. The door opens onto very little an AI can quote.

23/50

publish an llms.txt, while only 2 of 50 block any AI crawler

Industry fame and machine legibility measure different things. Rise at Seven ranks 46th, Screaming Frog 39th, Croud 34th, Brainlabs 31st and Builtvisible 21st. Their reputations aren't in question. This study measures whether three public pages give search engines and AI answers material they can read and cite.

Every check, every agency

The 13 checks scored, ordered by how often they fail. Hover a segment for the counts.

Credentialed expert quotes weight 4

Attributed statistics weight 5

Authoritative citations weight 3

Question headings weight 5

High-effort original content weight 4

First-hand experience weight 2

Canonical fact base weight 6

Scannable structure + markup weight 4

Authors as entities weight 3

Disambiguation factoids weight 4

Organisation schema + sameAs weight 4

Bylines to credentialed bios weight 4

Answer-first chunks weight 4

Yes Partly No

The top 10

Congratulations to the top 10. Opace took the top rosette at 67.3. Two-person Rapid SEO London placed second at 64.4 and sells SEO from £495 a month.

#	Agency	City	Score	Band	3-run range
1	Opace	Birmingham	67.3	Emerging	62.5 to 76
2	Rapid SEO London	London	64.4	Emerging	52.9 to 68.3
3	SALT.agency	Leeds	60.6	Emerging	55.8 to 65.4
4	Vixen Digital	Brighton	58.7	Emerging	51 to 62.5
5	Get Found	Birmingham	53.8	Emerging	53.8 to 55.8
6	Dark Horse	Manchester	52.9	Emerging	44.2 to 52.9
7	StudioHawk UK	London	50	Emerging	50 to 54.8
8	Hallam	Nottingham	50	Emerging	50 to 53.8
9	Blue Array	Reading	50	Emerging	50 to 58.7
10	Sierra Six Media	London	49	Exposed	46.2 to 60.6

Vixen Digital placed fourth nationally in this study, with 58.7. That's a fine result. Dog on the Table is Brighton-based too, so this is local pride, not rivalry.

Methodology, honestly

The method has clear limits. I measured only the 13 checks observable on public pages and didn't guess process or off-site work. A full 27-check score could be higher or lower. I judged the named top 10 and selected Sussex agencies three times, then used the modal result. I judged ranks 11 to 50 once and kept them anonymous.

What was scored, and what deliberately wasn't

Public evidence has limits. Only 13 of 27 checks are observable from public pages. They carry 52 of 100 weight points, which I re-based to 100.

I didn't fill the gaps. I excluded process, strategy and off-site checks rather than guessing. The full 27-check score could be higher or lower.

How each agency was assessed

On 19th August 2026, I fetched each agency's homepage, about page and one recent blog post. I used deterministic parsing for schema, headings, structure, bylines, robots.txt and llms.txt, plus an AI reading against the Standard's criteria. Both methods had defined jobs.

I verified every blog post URL by hand. Automated pickers sometimes grab service pages. Where the newest post was years old, I scored that post because assessing what exists is the honest method.

Named scores were judged three times

AI judging varies between runs. I judged every named agency three times independently: the top 10 and the Sussex agencies in the cohort. For each check, I used the modal status.

Each named table row shows the three-run range. I judged anonymised agencies once, which is why they appear only as rank, city and band.

The cohort, and who was excluded

I built the cohort from live UK search results for agency terms covering London, Manchester, Birmingham, Leeds, Brighton and the UK overall. I added well-known industry names. Dog on the Table clients and partners stayed out.

Before scoring, I struck 12 candidates: overseas parent networks, non-agency outliers and very small shops. Naming a two-person studio in a ranked list punches down. That wasn't the study I wanted.

Edge cases worth knowing about

One agency's site blocks ordinary crawlers. I fetched its pages through a rendering service instead. The same agency serves its llms.txt only to AI crawlers, the most deliberate AI-crawler engineering in the cohort, so I counted it as present.

I verified every llms.txt result by hand. I counted HTML error pages as absent and genuine files as present, including plugin-generated files.

This first edition sets the baseline. I'll run Best in Show quarterly with the same method. That makes movement measured, not guessed.

Questions agencies will ask

How were agencies scored?

On 19th August 2026, I assessed each agency's homepage, about page and one verified recent blog post. The 13 observable checks represented 52 of the Standard's 100 weight points, which I re-based to 100. I used public evidence only.

Why is my agency anonymous?

I named the top 10. Ranks 11 to 50 appear only by rank, city and band, unless an agency chooses otherwise.

How do I get my score?

Tell me which agency is yours. I'll send its private score if it's in the study or score your site if it isn't.

Want your score?

If your agency is in the study, I can share its private score. Anyone can have their site scored too. Make your brand the answer.

I already work behind the scenes with several agency brands, white-labelling senior GEO and SEO work. Your agency can be one of them.

Let's chat. How I work with agencies