Shops, built or rescued

What it is

An ecommerce build here is one person designing, writing and coding a shop end to end, and finishing the parts a theme leaves half done: payments, shipping, tax, tracking, and the policy pages nobody reads until a dispute.

A shop is not a website with a buy button on it. It is a catalogue, a set of tax and shipping rules, a payment path that has to survive a bad connection, and a measurement chain running from an advert to a paid order. Any one of those quietly wrong costs money daily and shows nothing on the shop front, which is why a store gets read before it gets redrawn.

Both build rungs carry the same work rather than selling it back later: analytics and consent configured lawfully, and a handover with the accounts, the theme files and a written account of the build.

Who it is for

Four kinds of buyer. Small brands with real stock and a first shop to build, where the catalogue is tens of products rather than thousands. Owners of a store that went live and then went quiet. Shops that have outgrown a bought template. And sellers crossing borders, where currency, shipping zones and tax settings are the whole job.

The work goes to the Gulf, Europe, the United Kingdom and North America, in English, in writing, across time zones. A written process survives a nine-hour gap and leaves you a record of every decision instead of a memory of a call.

What it costs

These are the shop rungs of the studio ladder, published so you can rank this studio against others before writing to anyone. There is no hourly rate here or anywhere else, because an hourly rate makes working quickly a penalty.

USD is the anchor on every quote and invoice. The figure in parentheses is the same price in dirhams at the pegged rate of 3.6725 to the dollar, rounded to the nearest hundred and given as one local equivalent rather than a currency list, correct as of 5 September 2026. Prices exclude any VAT, sales tax or withholding your own country applies to imported services, and the platform's own costs: a Shopify subscription and a payment provider's percentage are yours, never marked up here.

Shopify or WooCommerce

The choice is made on how the business runs, not on which platform is fashionable this year. Shopify where payments, tax rules and shipping have to keep working without anyone thinking about them, and a monthly fee for uptime somebody else maintains is a fair trade. A custom theme stops the shop looking like the four hundred others on its template.

WooCommerce where the shop belongs inside a content site you already publish on, where the catalogue rules are unusual enough that a platform's assumptions get in the way, or where a monthly subscription is the wrong shape for a seasonal business. It costs more attention afterwards, and that trade is named in the recommendation rather than found later.

The recommendation arrives in writing before the quote and you can overrule it. A studio that only ever recommends one platform is describing its own convenience.

The audit first

Most people writing about an existing store believe it is basically fine and needs a fresh coat. The cheap way to test that is the audit: USD 900, one-off, the same rung sold on the search page pointed at a shop instead of a site.

It reads what a shop front cannot show you. Whether the pixel and the server-side events fire on an add to cart and a purchase or only appear to. Whether the catalogue is the one you think you sell, whether the tax flags are set on every product or on two of them, whether the policy pages still carry placeholder text, and whether the compare-at prices are real. It is written so your own developer can act on it without me, and plenty of audits end there.

Your checkout, and the risk of touching it

Every owner of a working shop asks this, and it deserves a mechanism rather than a reassurance. The checkout is not rewritten. On Shopify it belongs to Shopify and theme work does not reach into it. On WooCommerce the payment gateway keeps its own plugin, settings and updates.

The build happens on a duplicate of the live theme, so the shop your customers are using is untouched while the new one is made. Before publishing there are test transactions through the real payment provider in test mode, on a phone as well as a desktop, and a check that the order confirmation, the stock decrement and the tracking events still happen. Publishing is one action, undone by publishing the old theme again.

Speed, and what actually causes it

Store speed is rarely the theme and almost always the weight the shop agreed to carry: apps injecting a script into every page for a feature used on one, product images at the size the camera made them, widgets loading before the price does. A speed pass is subtraction, and every removal is written down with what it was doing, so you can put it back.

The measuring bench is this studio's own, and the scripts that read frame cost and paint behaviour run on the work before it ships. There is no before-and-after store speed figure from a client here, because no such job has been sold and measured yet.

How it runs

1. You write the brief in a sentence. The card below takes it: what you sell, roughly how many products, and the address of the store if one exists.
2. I reply in writing, usually inside a day, saying whether this is a build or an audit, and which platform it belongs on.
3. The audit first, where a store exists. Inside two weeks, findings ordered by worth against cost, then a call to walk them. A build starting without it is a guess.
4. A fixed written quote, scope, exclusions and payment stages named. Catalogue size, currencies and languages are priced on its face.
5. The build on a duplicate theme, with pages and drawings arriving as they are made rather than at a reveal. Products, tax rules and shipping zones loaded and tested.
6. Launch and handover. Accounts, theme files, analytics, Search Console, training, and the written account of how the shop works. The first live orders are watched.

What it has proved

One shop built, one audited, each named with what it does and does not prove. Two projects are a start, not a track record.

• La Soie, built. The client Search Console export, handed over 28 August 2026 and covering 25 May to 23 August 2026, shows an average position of 6.75 across the run with impressions up elevenfold. That proves what the shop did in search after launch, not the build.
• Pink Salt, audited. Eight months live at zero orders, roughly half the catalogue off its own niche, the tracking pixel never firing, placeholder policy text in place, phantom compare-at prices on nearly every variant, tax flags off on all but two products. That is what an audit finds. It is not a result, because nothing was rebuilt.

The build side of La Soie is my own description rather than a measured claim: six bespoke theme sections, a Meta pixel and server-side events verified as firing with zero errors, GA4 and Merchant Center connected. The dormancy line on Pink Salt is that owner's own account, recorded in research note 07 and repeated with that attribution rather than as a measurement of mine. There are no testimonials on this site and no star rating in its markup, because it has neither.

Who it is not for

Questions people ask before they write

Sources

1. Client Google Search Console export for La Soie, handed over 28 August 2026 (Chart.csv, Queries.csv and Pages.csv), covering 25 May to 23 August 2026.
2. Pink Salt store audit, this studio's own reading of that shop. The eight months at zero orders is the owner's account, recorded in ShepherdUX research note 07, section 2.4.
3. Pricing rungs and buyer behaviour, ShepherdUX research note 18, September 2026, drawing on published agency and freelance rate surveys.
4. United Arab Emirates dirham pegged at 3.6725 to the United States dollar, Central Bank of the UAE, rate as of 5 September 2026.