Barking Cat, Commerce Advisory
The Commercial Experience Index Edition Two  /  Premier League Club Shops  /  August 2026

Good for humans. Poor for machines.

All twenty Premier League club online stores, scored out of 100 on the experience each club chose to publish. Bournemouth score 73 and top the table. The establishment is behind a promoted-era coastal club on the measure that is about to matter most.

DOWNLOAD THE FULL REPORT (PDF) 26 pages · 213 KB · No form, no email, no gate

The predictions are in. A supercomputer has run ten thousand simulations, a features desk has walked all twenty clubs in order, and neither can agree on where most of the league will finish. That is the annual ritual, and it is good fun. It is also a guess about a season nobody has played yet.

There is another table that does not need a simulation. It can be read today, from the front of each club's online shop, and it measures something already true: whether the experience a club has built helps a supporter buy, and whether the machine now standing between the club and that supporter can read the store at all.

This is that table. The Commercial Experience Index scores all twenty Premier League club stores out of 100, across eight observable measures, read entirely from the live shopfront. It does not see the profit and loss, the traffic, the conversion rate, or the margin. It sees what a customer meets, and increasingly what a machine meets, and it scores that.

01

The table is compressed, and the compression is the finding.

Bournemouth score 73. The next nineteen sit between 57 and 72, and fourteen of them fall inside a six-point band from 60 to 66. On the measures a customer actually feels, most of these stores are competent and unremarkable, and the separation, where it exists, does not follow the size of the club.

02

The famous clubs build well for people and badly for machines.

Arsenal run one of the best customer experiences in the league and score 3.5 on AI Readiness. Chelsea pair a category-leading store with almost no machine-readable data. The pattern holds across all five enterprise stacks belonging to the biggest clubs.

03

One agent standard reached the league through one platform.

An llms.txt file was found on exactly four clubs, and all four run Shopify: Bournemouth, Ipswich, Newcastle and Forest. Every one of the other sixteen returned a 404. The newest agent-readiness standard arrived at the smaller clubs first, and it arrived by accident of platform.

All twenty clubs, ranked out of 100. Each lens is scored 1 to 10 from the live shopfront in August 2026. Five lenses are operator judgement, three read objective site behaviour. Scores will move as clubs update their stores.
#ClubPlatform Prop Exp Pers Chk Copy Perf SEO AI Total
1BournemouthShopify686867.98.69.073
2ChelseaFanatics787786.77.54.572
3Leeds UnitedFanatics776788.27.14.570
4ArsenalSAP Hybris787685.48.13.569
4LiverpoolSalesforce / Centra786676.38.06.069
6Aston VillaFanatics676778.17.17.068
6Nottingham ForestShopify776775.77.28.068
6SunderlandFanatics676798.07.14.568
9EvertonFanatics676777.27.17.067
10BrightonJonas676767.27.85.066
10FulhamJonas676767.28.05.066
12Manchester CitySalesforce776764.17.66.065
12Manchester UnitedSCAYLE776665.48.16.065
14Ipswich TownShopify666766.08.17.864
15Crystal PalaceJonas676765.48.02.563
16Tottenham HotspurSalesforce676855.85.74.262
17Hull CityJonas566756.57.55.060
17Newcastle UnitedShopify566756.07.26.860
19Coventry CityJonas664567.37.65.058
20BrentfordJonas665566.07.22.557

The lens scores do not sum to the total. The eight combine on a weighted basis, and that weighting is the Barking Cat method rather than published arithmetic. A single point is not the story either. Two clubs can share a total and leak in entirely different places, and the more useful question is where each one is strong and where it is losing the supporter.

Five measures read what a customer meets. Three read what a machine meets. Each is scored 1 to 10 from the live shopfront, bracket-first, with 5 as the resting place of competent-but-unremarkable.

01

Proposition & Brand Clarity

With the sound off, in three seconds, can a stranger tell what this brand is, who it is for, and why it over the alternative?

All twenty pass the first two questions. The crest is present and the store is unmistakably official. Every point of separation sits in the third question, and specifically in whether the club authors an identity concept or drops the season's kit into a competent template. Seven clubs reached above average. Arsenal run Arsenal State of Mind off the back of the 2025/26 title, Chelsea build the store around 75 Years of the Rampant Lion, and Leeds carry 400 Years of Leeds, tied to the 400th anniversary of the city's town charter. Nottingham Forest is the quiet standout because it is the smallest name in that group: the tree crest, the FOREST wordmark and a hero that is an actual forest earn the score on craft. The floor is Newcastle and Hull, both at 5. Newcastle's hero names the kit and leans on the adidas partnership. Hull carry a problem the others do not, because the store trades as TigerLeisure.com rather than Hull City, so a stranger has to read the crest to work out whose shop they are in.

Two rules governed the scoring, and both were applied to every club. Supplier co-branding is not club identity work: a Nike, adidas, Puma, hummel or Macron lockup in the hero tells the stranger who makes the shirt, not who the club is. And an identity concept has to be sustained beyond a single banner to reach the upper band. One striking hero is a touch. A concept carried through the hero, the copy and the store structure is craft.

02

Experience, Clarity & Merchandising

Can the customer move without being made to think, and is product presented and sequenced with commercial logic?

The measure where the platform shows most clearly, and where one governing rule did the heavy lifting: platform-default furniture counts toward the average band and no further. The standard filters, the quick-view and the mega-menu that every Shopify, Jonas, Fanatics and Salesforce store ships out of the box are competent and nothing more. The Fanatics stores carry the richest faceted filtering in the study and it would be easy to reward that as craft. It is not craft. It is Fanatics, shared identically across five clubs.

Four clubs reached the top at 8. Bournemouth run a branded Premium Cherries lifestyle collection and a YETI collaboration. Chelsea layer story-led collections, including Summer of 64 from the club's Caribbean tour, on top of Shop By Player. Arsenal stack an adidas Collections carousel, a Shop By Category block and audience quick-filters. Liverpool run a genuinely bespoke store with tabbed kit navigation and the LFC Label own-brand line. The badge-versus-operator pattern is at its clearest here: on the same Jonas platform, Brighton, Palace and Fulham reach 7 while Brentford, Hull and Coventry sit at 6. The platform is identical. The difference is what the operator chose to build.

03

Persuasion & Buyer Journey

Are influence triggers present where the customer hesitates, and does the journey serve the different ways people buy?

The flattest of the judgement measures. Most clubs deploy the same light-touch persuasion: a free-shipping progress bar, transparent personalisation, a clean call to action. That is a genuine trigger or two and it earns the competent band. Arsenal and Chelsea are the only two to clear it. Arsenal is the strongest and the only club in the study to surface genuine customer-reviews social proof on the homepage, 4.9 out of 5 from more than 8,000 reviews, alongside a members-save-10% incentive. Coventry is the lowest at 4: an interstitial fires over the homepage and the checkout buries guest beneath an expanded sign-in form. Brentford sits just above at 5, firing an interstitial of its own, though that one carries a prize draw for a signed shirt. The gap between them is the difference between an interruption that offers the customer nothing and one that offers something real.

No club reached the exceptional band. Not one of the twenty builds a persuasion journey deliberate enough, across enough buyer modalities, to be called a competitive weapon. That is the honest state of persuasion in Premier League retail: competent triggers almost universally, and architecture almost nowhere.

04

Checkout & Friction

The commit path, read as far as the front end honestly allows: the guest gate, express pay, step signposting, trust at payment.

All twenty were walked independently, with no inference from a shared platform, and the read stops at the point of personal-data entry. No transaction was completed. The single most useful finding is that guest checkout exists at all twenty clubs. Not one forces account creation to buy, so the measure does not separate on availability. It separates on how prominently guest checkout is offered and on the breadth of express payment.

Tottenham run the best-signposted guest gate in the study, listed first and above login, with the fullest express-pay array seen anywhere: Apple Pay, PayPal, Google Pay, Amazon Pay and Onepay. Bournemouth match it at 8 on the cleanest Shopify flow, with no account wall, Shop Pay and Google Pay, United Kingdom preset, tax shown transparently and marketing consent switched off by default. The floor is Coventry and Brentford, both at 5, burying the guest option beneath an account default and restricting express pay to PayPal only. Brighton and Palace, on the identical Jonas platform, both score 7. The weaker checkout is an operator choice the same platform's better users have declined to make.

05

Copy & Voice

Do the words do a job, and does the brand sound like someone or like a theme's placeholder?

The widest genuine range of any judgement lens, because copy is the one thing entirely within a club's control regardless of platform. A club on the strictest shared backend can still write like itself, or fail to. Sunderland top the measure at 9, the only exceptional score on any judgement lens in the study. The 2026/27 home shirt copy tells the story of the club's 1937 FA Cup win: the jacquard S woven in the shape of the cup handles, the crest reproduced from the 1937 design, and the names of that year's first eleven printed inside the neck. Chelsea, Leeds and Arsenal follow at 8, each with a sustained narrative rather than a slogan. The floor is Tottenham, Newcastle and Hull at 5, functional and correct and carrying no distinct club voice.

06

Performance

Objective

How quickly and cleanly the store loads and runs on a mobile device, which is the right form factor: Google indexes mobile-first and most football retail traffic is mobile.

This measure discriminates hard, and it delivers the edition's most striking inversion. The promoted and smaller clubs run the fastest, best-built stores and the giants run the slowest. Leeds top it at 8.2, with Aston Villa (8.1), Sunderland (8.0) and Bournemouth (7.9) close behind. At the other end sit Manchester City at 4.1, then Arsenal and Manchester United at 5.4, all three heavier and slower to load than clubs a fraction of their size. Manchester City is the clearest case: the store is not catastrophically slow, but its build hygiene is the weakest in the study and the measure catches it.

Two clubs deserve a specific watch-out the headline softens. Crystal Palace and Brighton both post fast paint times and fail badly on layout stability, Palace on a Cumulative Layout Shift of 0.44 and Brighton on a brutal 0.78, both deep in the poor band. A page jumping under the customer's thumb as it loads is a real conversion risk that a fast paint time hides.

07

SEO Readiness

Objective

The structural conditions to be found, read from the front end only: title and meta structure, heading hierarchy, schema presence, crawlable navigation, clean URLs, canonical hygiene. Not rankings, not traffic.

The least dramatic measure in the study, because front-end SEO structure is largely a platform inheritance and every current platform does the basics well. The band is tight and high. Bournemouth top it at 8.6, with Manchester United, Ipswich and Arsenal on 8.1, and Fulham and Crystal Palace on 8.0.

One club breaks the pattern downward, and it is a name that should not. Tottenham score 5.7, the lowest in the study, on a shopfront whose structural SEO is measurably weaker than a field where everyone else is clean. For a club of that scale and resource it is a genuine and fixable underperformance, and the most notable single result on this measure.

08

AI Readiness

Objective

Can a machine find, read, quote and route to the store? Readiness, not output-presence: whether an agent can reach and parse the store, not whether assistants currently recommend it.

Weighted below organic search today, because AI-assisted discovery is a real but still smaller commercial channel, and built to matter more in later editions as the channel takes share. It is the measure that sorts the field most decisively by operator, and where the real story of the big clubs sits. Bournemouth top it at 9.0: the owner-run Shopify store exposes a complete, clean-rendering product schema and serves a full agent-commerce llms.txt file, verified live. Nottingham Forest (8.0) and Ipswich (7.8) follow on the same Shopify strength.

The floor is stark. Crystal Palace and Brentford sit at 2.5, both returning nothing an agent can transact against and no owned structured data. The famous names sit far lower than their customer experience would suggest: Arsenal 3.5, Spurs 4.2, and Chelsea, Leeds and Sunderland all on 4.5. These are clubs that build good stores for humans and expose almost nothing an AI agent can transact against from data the club controls.

Two findings carry weight beyond the scores. The llms.txt agent-instruction file was found on exactly four clubs, and all four are the Shopify stores: Bournemouth, Ipswich, Newcastle and Forest. It is a byte-for-byte identical, platform-generated file. Every one of the sixteen clubs on any other platform, including all five enterprise stacks belonging to the biggest clubs, returned a 404. Second, Liverpool carry a confirmed defect: the store's structured data serves a price of 150 US dollars to the crawler while the shopper sees 120 pounds sterling. A machine comparing Liverpool against a rival reads the wrong price in the wrong currency.

The winner, the split, the craft and the floor.

Four results that carry the shape of the whole table.

73
AFC Bournemouth
The clear winner

The only club in the study strong on both what a customer meets and what a machine meets. Bournemouth top three of the eight measures, Experience (8), SEO Readiness (8.6) and AI Readiness (9.0), and finish no lower than competent on any of them. The owner-run Shopify store pairs the cleanest checkout in the Shopify group with a genuine merchandising idea and a fully realised agent-commerce setup.

Room to improve: Proposition and Copy both sit at 6. Lifting the storytelling to match the build is the difference between 73 and the high 70s.

69
Arsenal
The split, at its clearest

The best all-round customer experience in the study alongside Chelsea, and the only club to surface genuine reviews-based social proof on the homepage. Arsenal reach 8 on Experience and Copy and 7 on Proposition and Persuasion, then score 3.5 on AI Readiness, the lowest of the enterprise clubs, with nothing an agent can transact against and no llms.txt.

Arsenal control their own platform, which means the machine-facing gap is theirs to fix and theirs to own. There is no operator to blame for it.

68
Nottingham Forest
Craft, not reputation

The quiet standout of the top group, because it is the smallest name in it. Forest reach 7 on Proposition on genuine craft, the tree crest, the FOREST wordmark and a hero that is an actual forest, and 8.0 on AI Readiness with a live llms.txt. A smaller club authoring a concept and building for the machine, sitting level with clubs several times its size.

Room to improve: Performance at 5.7 is the softest of the top eight.

57
Brentford
The floor, and it is fixable

Bottom of the table on a combination of a buried checkout and the weakest AI readiness, 5 and 2.5. Brentford finish last not because the store is broadly bad, it is competent on the customer-facing measures, but because it is weak on the two specific things this instrument weights and the market is moving toward.

Both are fixable, and the same Jonas platform proves it: Brighton and Palace score 7 on Checkout on the identical stack.

All twenty in rank order, what works and what leaks.

73/ 100

1. AFC Bournemouth

Shopify, owner-run

Working: the objective backbone. Fast, clean, machine-readable and frictionless to buy from, with the Premium Cherries lifestyle collection and a YETI collaboration doing real merchandising work. Not working: Proposition and Copy both at 6. The store sells the season's kit tidily without authoring a club-identity concept the way Forest or Arsenal do. The win is built on execution, and execution is copyable.

72/ 100

2. Chelsea

Fanatics

Working: the best-realised customer experience in the study. Proposition, merchandising, persuasion and copy all above convention on a shared platform, carried by 75 Years of the Rampant Lion and story-led collections. Not working: AI Readiness at 4.5 and no llms.txt. Built well for humans and poorly for machines, and the Fanatics stack caps the owned-data measures structurally.

70/ 100

3. Leeds United

Fanatics

Working: identity and performance together. The 400 Years of Leeds town-charter story is authored properly into the store rather than left on the shirt, and the build tops the Performance measure at 8.2. Not working: AI Readiness at 4.5 and no llms.txt, the Fanatics owned-data ceiling again, and only the operator can change it.

69/ 100

4. Arsenal

SAP Hybris, self-run

Working: persuasion specifically. The 4.9 out of 5 from 8,000-plus reviews and the members-save-10% incentive make this the strongest persuasion result in the study, and the only genuine multi-trigger architecture found anywhere. Not working: the objective backbone lets a strong front end down. AI Readiness 3.5 and Performance 5.4, on a platform Arsenal control themselves.

69/ 100

4. Liverpool

Salesforce / Centra

Working: design and merchandising. A non-templated build with tabbed kit navigation, the LFC Label own-brand fashion line and a footwear range. This looks authored, not inherited. Not working: two specific, fixable defects. Free delivery is nudged behind account login, and the structured data serves 150 US dollars to the crawler while the shopper sees 120 pounds. That is the most concrete single technical fault in the study, and it is invisible to the human shopper.

68/ 100

6. Aston Villa

Fanatics

Working: a rare club strong on both sides of the split. The Lions heritage voice is authored, and the objective scores are among the best of the established names, Performance 8.1 and AI Readiness 7.0. Not working: Proposition holds at 6. Villa also show what Fanatics can do in the right operator's hands, since the same stack that caps Chelsea and Leeds on AI readiness scores 7.0 here.

68/ 100

6. Nottingham Forest

Shopify

Working: proposition earned on craft rather than reputation, and a strong machine-facing backbone at 8.0 with a live llms.txt. Forest author an identity concept most bigger clubs do not attempt. Not working: Performance at 5.7, the softest of the top eight, and Persuasion holds at the competent 6.

68/ 100

6. Sunderland

Fanatics

Working: the single best copy score in the study at 9, earned on object-level storytelling most clubs do not attempt. The 1937 FA Cup shirt story is a piece of history the customer carries, written in a voice that sounds like no one else. Not working: AI Readiness at 4.5 and Proposition at 6. The words are exceptional; the machine-readability is not.

67/ 100

9. Everton

Fanatics

Working: heritage voice and a respectable machine-facing score of 7.0. The writing leans on dockside and marine heritage and the club's Latin motto, and sounds recognisably like Everton. Not working: Proposition and Persuasion both at 6. A competent all-rounder with no weak lens and no standout, and the route up is one authored discovery idea.

66/ 100

10. Brighton & Hove Albion

Jonas

Working: the strongest of the Jonas clubs, lifted to 7 on Experience by a personalised Recommended For You row the platform does not give for free, with clean SEO readiness at 7.8. Not working: a Cumulative Layout Shift of 0.78, deep in the poor band. The clearest case in the study of a blended objective score masking one catastrophic metric.

66/ 100

10. Fulham

Jonas

Working: two operator-authored discovery routes on a shared platform, themed gifting collections and Shop By Player, plus the strongest SEO readiness of the Jonas group at 8.0. Not working: Proposition, Persuasion and Copy all at 6. Well organised, and it does not yet sound distinctly like Fulham.

65/ 100

12. Manchester City

Salesforce Commerce Cloud

Working: the customer-facing build is genuinely strong. Layered merchandising, the sustained City State of Mind identity and confident copy on a bespoke enterprise stack. Not working: Performance at 4.1, the weakest build hygiene in the study. The store is not catastrophically slow, but the engineering is heavy and third-party-laden, and for a club of City's resource that is conspicuous.

65/ 100

12. Manchester United

SCAYLE

Working: a confident, well-merchandised store with a clear identity and among the best structural SEO in the study at 8.1. Not working: AI Readiness 6.0 and Performance 5.4. The machine-facing measures sit below the customer-facing ones, the pattern that runs through every big club, and even a modern stack like SCAYLE has not closed it.

64/ 100

14. Ipswich Town

Shopify

Working: the machine-facing backbone, AI Readiness 7.8 and SEO 8.1 with a live llms.txt. Ipswich are more ready for the agent channel than most clubs several divisions of fame above them. Not working: every judgement lens holds at the competent 6. The store works and embarrasses no one, and authors no discovery idea and no distinct voice.

63/ 100

15. Crystal Palace

Jonas

Working: fit-based navigation, letting a customer shop by body fit rather than only kit type. It is the most useful discovery idea in the whole study, and the platform does not provide it. Not working: AI Readiness at 2.5, joint-lowest, with nothing an agent can transact against, and a Cumulative Layout Shift of 0.44. The best single merchandising idea in the study sits on the worst machine-readiness.

62/ 100

16. Tottenham Hotspur

Salesforce Commerce Cloud

Working: the checkout, at 8. Guest checkout first and above login, with Apple Pay, PayPal, Google Pay, Amazon Pay and Onepay all offered. This is how the commit path should be built. Not working: SEO Readiness at 5.7, the lowest in the study, and Copy at 5 with no distinct club voice. For a club of Tottenham's scale, two conspicuous underperformances.

60/ 100

17. Hull City

Jonas

Working: value and speed. The cheapest standard shirt in the league at 64 pounds, a loyalty-points scheme, an Auctions section, a Performance score of 6.5 and a genuine promotion-story hook in We Are Premier League. Not working: Proposition and Copy both at 5. The store trades as TigerLeisure.com rather than Hull City, which is the clearest self-inflicted proposition weakness in the study, and it sits right at the top of the funnel.

60/ 100

17. Newcastle United

Shopify

Working: the machine-facing backbone inherited from Shopify, AI Readiness 6.8 with a live llms.txt, and a polished premium visual treatment on the category pages. Not working: Proposition and Copy both at 5. The hero names the kit and leans on the adidas partnership, and the copy carries no distinct voice. The platform gives them a strong base; the club adds little on top.

58/ 100

19. Coventry City

Jonas

Working: the category page is genuinely competent, with sub-category pills, a full filter set, quick-view and strong editorial photography, and Performance is a respectable 7.3. Not working: the commit path and the persuasion around it, 5 and 4, the joint-lowest on both. An interstitial fires over the homepage and the checkout buries guest beneath an expanded sign-in form with PayPal as the only express option. The fix is cheap and proven on the same platform.

57/ 100

20. Brentford

Jonas

Working: the Leave Your Mark campaign hero has genuine attitude and voice, enough to lift Copy to 6, and the category page is competently built. Not working: guest checkout is buried, express pay is PayPal only, and the store exposes almost nothing an AI agent can use, scoring 2.5. Bottom of the table on the two specific things the market is moving toward, and both are fixable.

The famous clubs do not run bad stores.

Arsenal, Chelsea and Liverpool build some of the best customer experiences in the league. What they run are stores that are good for humans and poor for machines.

Strong on proposition, experience, persuasion and copy. Weak, sometimes badly weak, on AI readiness and, for several, on performance. Arsenal build one of the best customer experiences in the league and expose almost nothing an agent can read. Chelsea pair a category-leading store with no llms.txt and little machine-readable data. The pattern holds across all five enterprise stores belonging to the biggest clubs.

The one club strong on both sides is Bournemouth, an owner-run Shopify store from a mid-table coastal club. The separation at the top of Premier League retail is no longer about how well you sell to a person, because most of these clubs do that competently. It is about whether you have also built for the machine that is increasingly standing between the club and the customer, and on that measure the establishment is behind a promoted-era minnow.

None of this predicts a league position. It measures something already true, on the day it was read, about the experience each club chose to publish. Two of the three tables printed alongside this one in August are guesses about a season nobody has played. This one is a measurement.

Edition Two, in full. Twenty-six pages.

Every lens explained, all twenty clubs written up in rank order with what each does well, where it leaks and the single thing worth watching, plus the full methodology. Free, and there is no form in front of it.

PDF · 26 pages · 213 KB
DOWNLOAD THE PDF

What the Index measures, and what it does not

The Commercial Experience Index scores the digital experience a brand chose to publish, read entirely from the front end of a live site, out of 100. Edition Two scores eight lenses: Proposition and Brand Clarity, Experience and Merchandising, Persuasion and Buyer Journey, Checkout and Friction, Performance, SEO Readiness, AI Readiness, and Copy and Voice. Five are operator judgement, three read objective site behaviour. Every score traces to something observed on the store or in the objective data.

It does not include commercial diagnosis, organic search performance, paid media, smart retail, or AI output-presence, meaning whether assistants actually recommend the club when a real buyer prompts them. Those need data or live prompting beyond the URL and belong to the full Barking Cat Review. Checkout is scored on the observable path only, with no transaction placed.

Each lens is scored 1 to 10, bracket-first. The centre of the scale is 5, the resting place of competent-but-unremarkable. A professionally run, unremarkable store scores in the 50s, not the 70s.

All twenty stores were read live in August 2026. Scores will move as clubs update their stores.

Edition One scored ten UK retail sites on seven measures. Read Edition One, or download the Edition One PDF.

Back to Thinking
The uncomfortable read
“The establishment is behind a minnow.”

Sixteen of twenty clubs return a 404 where the agent looks first, and all five enterprise stacks belonging to the biggest names are among them.

Barking Cat · Diagnosis before solutions