Clifton Website Audit

The short answer

Two problems, one plan.

The problem

Buying is harder than it should be, and a redesign could quietly switch off the things that get Clifton found.

1. Buyers have to work too hard. Someone ordering custom clothing has to work out what the garment is, what it costs and what happens after they pay. Pages describe the order process in different ways, the phone basket leads with money before the garment, and some controls can't be used without a mouse. Part 1 · Tandem Part 3 · No Logo
2. A new design can break what search relies on. Google, and AI tools like ChatGPT, find and recommend Clifton because of a handful of hidden details on every page. Hassan, the search specialist, has seven rules to protect them. The current site follows most. A new design might not, and nobody would notice until the enquiries dipped. Part 2 · SEO

The fix

Make the journey plain, protect what search relies on, and fix what's broken underneath.

Make the journey plain. Show the garment first, then the specification, then the price, then honest delivery dates. First job: confirm how an order really works with Clifton, because nobody can write it clearly until the facts are agreed. That's Tandem's part.
Make the search rules pass or fail. Every new page type is tested against Hassan's seven rules before launch. If a page fails, it doesn't ship.
Fix what's broken underneath. A size-guide button a keyboard can't reach, slow loading, a few defaults and numbers that need checking. That list belongs to No Logo.
Keep what already works. The evidence doesn't call for a rebrand, a new platform or a rebuilt designer. It calls for clearer presentation, consistent facts, usable controls and some engineering.

Who reads what

Three parts. Each reader only needs their own.

Simon only has to read Tandem's part. The other two go to the people who own them. Pick one to open it, or use the switch in the bar above.

Part 1 · For Simon and the Clifton team

Tandem

How we bring character back to Clifton Clothing, and make buying easier.

This is our part of the work: the look, the words, the images, and the way someone moves from "I like that" to "I've ordered it". Emma and Teagan design and write it. Clifton checks the facts, and No Logo builds it.

Owner: TandemClifton checks the factsNo Logo builds

What we're responsible for

Show the garment. Explain the choice. Make approval easy.

Clifton makes clothing from scratch. The site has to make that quality easier to see, the choices easier to understand and the order easier to approve. The redesign has four jobs.

01 · Reassure

Make the quality visible

Real garment detail, useful fabric information and relevant customer proof.

02 · Explain

Help people choose

Comparable products, and a clear link between style, fabric, price and delivery.

03 · Approve

Make the order complete

The design, sizes, delivered cost and next commitment in one understandable place.

04 · Mobile

Make it work on a phone

Readable information, visible choices and buying controls people can actually use.

How character comes back

Strong individual images don't yet make a consistent family, and several moving strips and large sections compete with the garment. Simon wants the made-from-scratch difference to be easier to recognise.

So we'll develop a few coherent visual directions from the same product and order facts. They'll show fabric, construction and group identity, with a shorter hero and a calmer page. Names such as Hanbury and Invicti are in the copy already but used differently in titles, cards and the designer, so we'll use name plus descriptor, the same way every time.

The test: can a buyer say what Clifton makes, who it's for and why it isn't a logo printed on a stock shirt?

What stays, and what we're not doing

✓Clifton's identity and expertise. The existing images are a starting reference, redone into one consistent, accurate system.
✓The existing platform and the designer, improved rather than replaced. How far the designer can change is still to be confirmed with No Logo.
×A rebrand
×Replacing the platform
×A photo shoot

Character can't live in the page's main headline. Search needs that line to name the subject, so personality goes in the banner, the images and the body copy.

The buying journey

Seven moments where a little more clarity helps.

Right now the buyer does a lot of joining the dots. What makes this fabric right? Why did the price change? What needs approving? Will it arrive in time? That uncertainty matters when someone is ordering for a whole group. These steps combine a review of the live site with Tandem's own walkthrough on 1 November. Pick a step to see what we found, why it matters and what we'll change.

Before we design

Five decisions, and what we recommend.

Walking the live site ourselves on 1 November turned up five calls that change what we design. None of them is made yet. Each one has an owner and our recommendation.

Clifton and Tandem

Who the front door is for

A short list of four or five audiences, with one kind of split. Not audiences mixed with body parts. We agree the exact list once the cut list is in.

Clifton

Which products to cut

Sixteen or more tiles is too many. We need the list of what goes, and sales by product, so the cut is based on what sells.

Hassan decides, No Logo builds

What happens to the five rugby pages

Fold them into one page with style and fabric swatches, and redirect the old addresses with their fabric pre-selected. Not before Hassan agrees which address and headline the merged page keeps.

Clifton, Tandem and No Logo

Where gender is chosen

Inside the style step, as a fit choice, so mixed club orders aren't forced to pick a side on the home page. Separate pages only where the range genuinely differs, as the women's cut seems to.

Tandem, then Clifton

Where the logo row goes

After the claims, not before them. It's a judgement call, so we would test the position instead of assuming.

M-01 · Top priority

Nobody can say what happens after you order, because the pages disagree.

Pages describe the digital proof, payment, order-pack approval and delivery timing in conflicting ways. The audits couldn't confirm the real order of events, so we're not guessing. The order below is deliberately unconfirmed.

Design?Proof?Approval?Payment?Production?Delivery

Read on 1 November, three pages already tell three different stories:

Rugby category

Describes a proof review before checkout.

Artwork FAQ

Describes an order pack after payment.

Shipping and returns

Describes approval as optional.

Six places that have to tell the same story:

Product pageHow it worksFAQBasketCheckoutFollow-up emails

Why it matters, and what we're doing

A buyer who can't tell what they're committing to will hesitate, and so will whoever has to approve the order. The fix starts with one real walkthrough of an online order and an assisted order, then one timeline for each, written once and used everywhere. It also unlocks the "lead time" line Hassan wants on product pages. That can't be filled in until this is settled.

Images and words

Make the quality easier to see.

People can't touch a fabric through a screen. The gallery and the copy have to do more of that explaining, and do it the same way across the whole range. Tap a view to see its job.

Illustration of the gallery slots, not real photography. Source accuracy and image rights still need checking before anything is used.

One family, not identical pictures

Keep a consistent scale, background and image order. Adapt the details to a blazer, rugby shirt or technical jersey so each product still looks like itself.

Explain the material where people choose

Pair a swatch with verified composition, feel, weight where available and useful buying guidance. Put the facts in readable text, not only in a picture.

Make help part of the offer

A confident buyer can start designing. Someone who needs advice should find a clear alternative beside that action, with an honest explanation of the service.

Use proof people can trust

Use accurate garment references and permissioned customer material. A review badge needs a verified profile and agreed timing. Generated images never stand in for real customer or factory evidence.

What stays within the proposal

Tandem's quoted imagery work redoes or regenerates the imagery, using the existing library as a starting reference. Phase 2 includes up to 300 images, with a repeatable workflow and accuracy checks. That is a ceiling, not a target to fill. Before allocating the work we'll confirm the current product list, which originals to use as references, and how crops and variants count.

Phase 2 also covers nine responsive page templates, three annotated customiser screens, six key-page copy treatments, up to eight category introductions and 20 priority product descriptions. No Logo's development scope is separate. These figures are from the revised proposal dated 22 October.

The design direction presentation offers three visual routes, with product and cart or quote page examples. This page explains the buying principles each route has to support. Simon's feedback picks the direction.

Top priority · M-18

Before we can say the redesign worked, we need to agree what counts as a sale.

These are the figures Simon gave us: 91 days either side of the redesign that went live on 31 July 2025. They're useful context. They are not proof.

Purchases per 100 visits

91 days before91 days after

Why we can't read this as lost sales

×It doesn't separate the design from the time of year.
×It doesn't separate the design from who was visiting.
×Changes to minimums and range happened around then too.
×Orders placed with help from Clifton may not show in these counts.
×Tracking itself may differ between the two periods.

So we won't turn the gap into a rand or pound figure. We'll set a clean baseline instead.

Six things we need to agree on counting

"A sale" means different things on different routes, especially for orders Clifton helps with. Each of these needs one written definition.

Thirty days after launch we compare like for like, with the numbers behind every percentage. If orders are too few to say anything, we'll say the result is inconclusive and keep watching rather than promise an uplift.

What happens next

Choose a direction. Then make every page work together.

Read this alongside the design direction presentation and the product and cart or quote examples. Together they show the opportunities, three possible looks and how buying can work.

Review the three routes together

Choose the direction that best expresses Clifton's quality, then judge it against the product and cart or quote examples.

Agree one clear set of feedback.
Agree the audience split, and cut the range to match.
Check a real anonymised quote, design proof and order pack.
Settle style, fabric, fit, minimum, price and market rules.
Confirm the payment, approval and delivery sequence for online and assisted orders.
Confirm current product priorities, imagery and use rights.
Map which garment, gender, fabric and design combinations exist.

Tandem presents. Clifton selects. No Logo and Hassan verify facts.

Refine the chosen direction

Use the agreed feedback and the verified garment and order facts to refine the examples, then carry the system across the wider site.

A useful gallery and a style and fabric explanation.
Selected price, minimum, delivery and design-help route.
The customer's design, sizes, delivered total and next commitment.
Clear notes showing what search information stays.

Tandem designs and writes. Clifton reviews the facts.

Start the independent technical work

These can move forward while the final look is developed. The full list is in No Logo's part.

Accessible size-guide buttons and form labels.
Loading delays, page movement and script costs.
Selection handoff, edit states and size defaults.
Showing a chosen design on the model straight away.
Delivery rounding, the account-history link and sitemap checks.

No Logo leads. Hassan checks anything search-related.

Test the whole journey

Safe orders and real devices. A good-looking screen still has to carry the right order through.

Mixed sizes, back navigation and design changes.
A second decision-maker reading the quote.
Guest continuation and payment recovery.
Search checks, accurate tracking and a 30-day review.

No Logo builds. Tandem reviews the experience. Hassan checks nothing was lost.

What we'll look for after launch

Compare like with like, by device, market and sales route. Track product-to-designer progression, completed orders and qualified enquiries. For assisted sales, follow quote versions through to acceptance and cleared payment, and allow enough time for groups to decide.

At 30 days we review order quality, recurring questions, technical measures and search outcomes alongside sales. More clicks alone isn't success, and small datasets may need longer.

What we need from Clifton

The facts only Clifton can supply.

Until these arrive we'd be guessing, and we'd rather not. They're also what unlocks the lead-time line, the price labels and the quote layout.

Ref	What we need	From	Unlocks	Status

The detail

Our jobs, in plain English.

Every job where Tandem leads. Priorities are proposals. Nobody has signed them off yet. Tap a square to jump to a job.

Part 2 · For Hassan

SEO

What search needs us to protect, and what we need from Hassan.

Hassan's seven rules are the checklist. Most of what they ask for already works on the live site. The job is to keep it working through the redesign, settle the few places where his rules and the site disagree, and get his five exports in before anything is built.

Owner: HassanNo Logo implementsTandem annotates and writes

Before design is locked

What Hassan needs to settle or supply.

His document is live and may still change. Some of these exist because the rules, read literally, collide with something the site does today. None of it is a criticism of the rules.

Promised, not received

Five exports he's promised

The approved title, description and headline for every page.
Category copy for every live category.
The current redirect map and sitemap spec.
Business and price data templates.
A crawler test script the team can run.

Eleven questions, not yet sent

Rulings where the rules and the site disagree

The home page headline is a slogan. Rule 3 asks for a subject.
"Clifton" and "CLIFTON" appear in home copy. Rule 7 says "Clifton Clothing".
Slash and capital-letter versions of an address load as duplicates. Rule 1 says no new slashes.
Business details for search sit on the home page only.
If five rugby products fold into one page, which address and headline it keeps.
Whether new audience pages can sit over the existing garment addresses.
Two blog posts showed no headline or preferred address in one crawl.

Needs booking

Reviews to put in the diary

One-pass mockup review of the home page, a category, a product and the 404 page, with a 24-hour turnaround.
Staging access, so he can run the same crawler tests on the test site.
A freeze list that every page loads its text from.
A launch checklist and same-day recrawl requests.

Everything we're waiting on from Hassan

Ref	What we need	From	Unlocks	Status

What we keep

What's already working, and worth protecting.

The live check on 1 November found that the site does most of what Hassan's rules ask. The risk is losing these in the redesign, so they're on the protect list.

✓Search-ready snapshots on all 8 page types tested
✓Honest "page not found" answers for removed pages
✓One real-text headline on every page checked
✓Category description in the page code (golf sample)
✓Business details for search on the home page
✓Price details for search on all 79 live product pages crawled
✓Fabric, fit and measurement detail
✓Guest checkout and assisted help

What this means in practice

We can bring products higher up a page without deleting the descriptive text. We can rename a menu label without changing the address behind it. We can shorten text on a phone, as long as it is still in the page code when the page loads.

What this audit is not saying

Neither audit proves that the July 2025 redesign caused lost sales. Nothing here shows Clifton has lost rankings. We tested how the site responds to search crawlers. We have not measured how Google actually ranks it.

Rule 4 in pictures

Your customer and Google are shown two different pages.

The Clifton site is a JavaScript app. That means a browser gets a nearly empty shell and builds the page on screen. Search crawlers can't always do that, so the site hands them a ready-made printed copy instead. Try it.

The shop window

What customers see in their browser. Built live, looks great, changes all the time.

The printed menu

A ready-made snapshot of each page, handed to Google and ChatGPT at the door. It carries the headline, the description and every word.

The risk

A new design can look perfect in the window and leave the printed menu blank. Customers wouldn't notice. Google would.

Pretend the new build forgot the printed copy

cliftonclothing.com/all/custom-golf-clothing

The eight pages we tested on 1 November

Page type	Address	Words handed over	Headline (H1)

Each one returned a successful response with the "prepared" stamp (x-prerendered: true), one text headline, a title, a description and a preferred address. We asked as a Googlebot and as a GPTBot. An ordinary browser request to the same pages gets a 21-word shell. These tests copy how a crawler introduces itself. They don't prove how the real Googlebot is treated.

Hassan's seven non-negotiables

Seven rules that keep Clifton findable.

Hassan's document says: break these and rankings reset. Tap a rule to see what it means, why it matters, what the live site does today and what it changes for the redesign.

Hassan's four design objectives

One real Clifton page, labelled.

Hassan's point is that a page that is clear to people is usually clear to machines too, as long as the important facts are real text, not baked into pictures. Tap a number to see what each part of a product page needs to be.

Mobile screenshot of a Clifton Clothing product page for Custom Cotton Rugby Shirts, showing photos, a description, a from-price and a Design and Order button.

Screenshot of the live mobile site, captured 21 October. Prices and minimums may have changed since. The audit's 28 October samples show a minimum of 10, so treat this page as an example of structure, not a source of facts.

M-16 · Downgraded to next-in-line

The sitemap still advertises 145 pages that are gone.

A sitemap is the list of pages you hand Google so it knows what to look at. Ours was last stamped in May and June 2024. It still lists nearly all the removed pages and leaves out 14 live ones.

What the sitemap lists, by type

Not found dead in our crawlReturns "page not found"

The "146 dead pages" correction

This number sounded alarming in the first audit. Here it is in proportion: a 404 is the right answer for a page that's gone. Nothing is broken for crawlers. The real issue is tidiness. The sitemap still tells Google to visit 145 of those 146 dead addresses.

What we're doing

With Hassan's sitemap spec, we regenerate the sitemap from the new build so it lists only live pages, including the 14 missing ones. We also confirm each alias points to its approved preferred address, and agree one single "page not found" template. Today the two versions carry different titles and addresses.

Hassan's four gates

Four checkpoints he wants in the plan.

They're reviews, not hurdles. Booking them early stops his 24-hour turnaround from stalling the design. Whether the home, category and 404 mockups sit inside Tandem's agreed scope is still to be settled.

Before buildMockup review

Hassan reviews the home page, one category, one product and the 404 page. One pass, 24-hour turnaround.

Before handoverStaging access

He runs the same crawler tests on the test site that he runs on the live one.

Start of buildFreeze list

One definitive export of addresses, titles, descriptions, headlines and category copy. Everyone loads from it.

Before go-liveLaunch checklist

Snapshots pass on every page type, redirects resolve, the sitemap regenerates, robots.txt is unchanged, recrawls requested the same day.

For Hassan

Eleven questions we're taking to him.

None has been sent yet. They exist because Hassan's rules, read literally, collide with a few things the site does today. Tap one for the plain-English version and why we're asking.

The detail

The search jobs, in plain English.

Every job where search leads. The seven rules above are the acceptance criteria, and these four are the work that makes them stick. Priorities are proposals, not sign-offs.

Part 3 · For No Logo, the developers

No Logo

What's broken or unproven underneath, and what we'd like fixed or checked.

These are the engineering findings from both audits: speed, access for keyboard and screen-reader users, order logic, tracking, and what the new build has to keep doing for search. Each one is either reproduced or marked as not yet reproduced. None is a verdict on how the site was built.

Owner: No LogoTandem specifies the lookHassan checks search

What we picked up

Five things underneath that need engineering.

None of this starts from a new platform. It needs careful fixes and some checks, and most of it can start now while the final look is still being developed. The designer changes are the biggest unknown, which is why we're asking before we promise anything. Tap one to jump to it.

These can start now

Accessible size-guide buttons and form labels.
Loading delays, page movement and script costs.
Selection handoff, edit states and size defaults.
Showing a chosen design on the model straight away.
Delivery rounding, the account-history link and sitemap checks.

Top priority · M-15

The site is slower than Google's own targets.

This is real-visitor data from Google, covering the latest 28 days, as recorded on 28 October. Each bar shows how long visitors waited, with Google's pass mark as the black line. Three of every four visits were this fast or faster.

Failed Google's assessment

Why it matters

Slow pages lose impatient buyers, and Google counts speed. The cause isn't known yet. It could be scripts, third-party tools, images or layout. No Logo will trace it before anyone assigns blame.

What we're doing

No Logo traces what slows the page. Tandem supplies image sizes and stable layouts. We then repeat identical tests after each real change. One catch: crawlers get the snapshot and people get the app, so every speed fix is tested the way people see it, not only the way Google does.

The stress test, kept separate on purpose

A mid-range phone (Moto G Power) on a slow 4G connection, two runs per page, until the main content appeared. It finds problems. It isn't what real visitors feel, and the two datasets shouldn't be averaged.

0 s5 s10 s15 s20 s25 s

What the three measures mean

LCP, loading. How long until the biggest thing on screen, usually the main photo or headline, appears. Google's pass mark is 2.5 seconds.

INP, responsiveness. How quickly the page reacts after you tap or click. Pass mark: 200 milliseconds.

CLS, jumpiness. How much the layout shifts while loading, like a button moving just as you tap it. Pass mark: 0.1.

These pass marks are the project's working targets, carried forward from the audits.

Top priority · B-04

The designer shows nothing until you pick a colour.

In screenshots from Tandem's own walkthrough of the live site, choosing a design leaves the model plain grey until a colour changes. It happens on every product, so the designer can look broken. We haven't reproduced it yet, so treat it as a strong lead, not a verdict.

Picking a design doesn't draw it on the model until a colour is changed, on every product.
The women's shirt is shown on a male model, and offers two style options.
Fabric is a small dropdown inside the Colours tab, where its effect can't be seen.
The design is picked before the fabric, so the designer can't filter designs by fabric.

What to check

Can the renderer draw a design's base colours as soon as it is selected? Can the model image change per range? Is there a record of which styles, fabrics and designs work together? The answers decide whether this is a small fix or a bigger change.

Next in line · M-03, M-07, M-17, M-19

Four small things that could cost trust at the worst moment.

None of these is a confirmed fault in a finished order. Each one was seen once, or in an initial state, so the first job is to reproduce it.

Size defaults · M-07

What we saw

In an earlier test all ten units started as Mens / Classic Fit / MENS-2XS, and a blank design was added. That's an initial state. It doesn't prove a wrong final order, because later safeguards weren't tested.

What to check

Mark sizes unconfirmed until reviewed, show assigned against total, and test mixed sizes and edits. Find out whether plain, undecorated garments are valid before warning about a blank design.

Delivery charges · M-17

What we saw

One cart showed a one-cent difference in the express price. One South African cart charged standard shipping even though a threshold was on display. Only one audit saw this.

What to check

Reproduce one identical basket and market. Confirm the threshold rules and the rounding, use a single price source, and show the destination and currency near the offer.

Clifton sets the delivery policy. No Logo fixes the calculation. Tandem labels it.

Account-history link · M-19

What we saw

The Shipping and Returns page links to account history with literal parentheses in the address, which probably breaks it. What happens on click wasn't established.

What to check

Confirm the supported address, fix the link, test the return after sign-in and keep guest order tracking working. Check the corrected link is in the redirect map.

Forwarded baskets · M-03

What we saw

One test basket opened in another browser behaved differently through a plain link and a share link, and sizing and checkout controls were visible. That's one observation, not a security breach.

What to check

Write down who may view, edit and pay for plain baskets, shared baskets and issued quotes. Improve what the recipient sees: who sent it, what it is, and what they can do.

Hassan's rules, from the engineer's side

What the new build has to keep doing for search.

Hassan owns the rules and the checks. No Logo owns the build. These are the parts that land on the engineering side. The plain-English version, with examples, is in Part 2.

Snapshots on every page type (rule 4)

Every new template is tested against the snapshot service from the first day of the build, not patched in later. Test: request the page as Googlebot and expect the full page and the header x-prerendered: true.

Today this passes on all eight page types tested, for a Googlebot and a GPTBot request. Snapshots can be up to a day old. These are spoofed user-agent tests, so they don't prove how the real Googlebot is treated.

Real 404s (rule 5)

A removed page must return a genuine 404 status, with a helpful page for people. This passes today on a removed product, a removed blog post and a made-up category.

Two gaps: the two removed-page versions have different titles and preferred addresses, and the /404 page itself sat behind a Cloudflare check that we didn't try to get around, so it wasn't tested.

Category text in the first HTML (rule 6)

The category paragraph has to be in the page code on load, even when it's restyled, shortened on a phone or collapsed. Hiding it behind a "read more" that fetches it later is the one thing that isn't allowed.

Not checked: whether the text stays in the code when the page's expandable sections are collapsed.

Addresses, metadata and one text H1 (rules 1 to 3)

Keep every live address exactly, including spelling, capitals and slash conventions. Load titles, descriptions and headlines from Hassan's approved list through the CMS. Each page has one real-text H1 that names the subject, and hero lines go in a separate element.

Schema and footer identity (rule 7)

Keep the Organization and Product schema and the footer line "Clifton Clothing is operated by Clifton Concepts LLC". Today the business details are on the home page only, and the saved product schema sample is the twill rugby shirt, not the cotton rugby pricing example.

Sitemap, aliases and offers (M-16)

Regenerate the sitemap from the new build so it lists only live pages and includes the 14 live pages it misses. 145 of the 146 dead addresses are still listed, and 35 of 79 live product pages name a different preferred address.

The price a crawler reads has to match the visible price for the same product, configuration and market. These are dated sitemap findings, not proof of lost rankings.

Merged product pages (B-07)

If the five rugby product pages fold into one, each retired address needs a permanent redirect to the merged page with its fabric pre-selected. Hassan decides which page keeps the address and headline, and the redirect map is his. Carry each product's price into the merged page's price data, and keep the category text in the page code.

Staging access and launch checks (the gates)

Hassan runs the same crawler tests on the test site before handover and again before launch. Give him access, regenerate the sitemap, keep robots.txt as agreed and request fresh visits from search engines on launch day.

For No Logo

Seven questions we'd like answered.

None has been sent yet. Each decides whether a change is small or big, so we'd rather ask before we design around it. Tap one for the detail.

The detail

The engineering jobs, in plain English.

Every job where No Logo leads, including the one on hold. Priorities are proposals, not sign-offs. Tap a square to jump to a job.

What we need to unblock this work

Ref	What we need	From	Unlocks	Status

Everyone

What we're waiting on, and from whom.

The full list across all three parts. Filter it to see only what you owe.

Ref	What we need	From	Unlocks	Status

Straight talk

What we're not doing, and what we haven't checked.

Not doing

×A rebrand
×Replacing the platform
×Replacing the designer. How far its behaviour can change is still being scoped
×A photo shoot
×Changing anything on the live site. This was read-only.

Not proven

×That the redesign caused lost sales
×That rankings have dropped
×That the audience front door, the merged rugby page or the home page changes will lift sales. None has been tested with buyers.
×How the real Googlebot treats the site. We copied how it introduces itself.
×How often AI tools mention Clifton
×That Hassan's example "$41.99 each (50+ pcs)" is an offer. It's his illustration.

Not tested yet

×The /404 page, behind a Cloudflare security check we didn't try to get around
×Redirects, basket and quote pages, any test site
×Real phones and full screen-reader journeys
×The designer's blank preview and the male model on the women's shirt. We saw them in one walkthrough's screenshots and have not reproduced them.
×More than one or two samples per page type
×Hassan's approved lists, which haven't arrived. His document is live and may change.

Jargon decoder

The words, in ordinary language.

Search the jargon

Two problems, one plan.

Three parts. Each reader only needs their own.

Tandem

Show the garment. Explain the choice. Make approval easy.

Make the quality visible

Help people choose

Make the order complete

Make it work on a phone

Seven moments where a little more clarity helps.

Five decisions, and what we recommend.

Who the front door is for

Which products to cut

What happens to the five rugby pages

Where gender is chosen

Where the logo row goes

Nobody can say what happens after you order, because the pages disagree.

Make the quality easier to see.

One family, not identical pictures

Explain the material where people choose

Make help part of the offer

Use proof people can trust

Before we can say the redesign worked, we need to agree what counts as a sale.

Choose a direction. Then make every page work together.

Review the three routes together

Refine the chosen direction

Start the independent technical work

Test the whole journey

The facts only Clifton can supply.

Our jobs, in plain English.

SEO

What Hassan needs to settle or supply.

Five exports he's promised

Rulings where the rules and the site disagree

Reviews to put in the diary

What's already working, and worth protecting.

Your customer and Google are shown two different pages.

Seven rules that keep Clifton findable.

One real Clifton page, labelled.

The sitemap still advertises 145 pages that are gone.

Four checkpoints he wants in the plan.

Eleven questions we're taking to him.

The search jobs, in plain English.

No Logo

Five things underneath that need engineering.

Speed

Keyboard and screen readers

Designer preview

Order logic

Tracking

The site is slower than Google's own targets.

The designer shows nothing until you pick a colour.

Some customers can't get to the size guide or know what a box is for.

Four small things that could cost trust at the worst moment.

What we saw

What to check

What we saw

What to check

What we saw

What to check

What we saw

What to check

What the new build has to keep doing for search.

Seven questions we'd like answered.

The engineering jobs, in plain English.

What we're waiting on, and from whom.

What we're not doing, and what we haven't checked.

The words, in ordinary language.