All4feet

The customers are pedicurists and salons who order the same things again and again, often on their phone between two clients. The shop made that hard: tier discounts only on the invoice, every colour and size a separate product, a core gel polish range in no category at all, and no quantity field on the phone.

The shop isn't a checkout. It's an order request.

The name is used with permission. This is a concept and a proposal; the shop was not rebuilt.

I pitched version 2 to the owner on 18 October. He liked the direction, and corrected almost every assumption behind it:

• Product sales are the core, not the repair service I had put first
• There is no stock system and no online payment: an order is a request on account, and staff call when something is short
• Prices only for logged-in customers; some must never be seen by consumers
• The customer file and SnelStart (bookkeeping) matter most
• Two category levels at most, no expensive plug-ins
• Calmer and more elegant: pedicure and high-end spa, not deals and prices

So I rewrote the brief and started version 3 from those answers.

One organising principle: type of product. Brands get their own page and a filter, offers become a label on the product, the divider lines disappear and the catch-alls get a real place during migration.

The owner's answers turned the cart into an order list. The customer sends it without paying; the team checks what is in stock, calls if something is short, and invoices from the bookkeeping system. The site says so at every step.

Every version was saved with screenshots, so each change can be compared with the one before. The biggest turn came from the pitch, not from a test.

The owner found version 2 too busy. Three style directions on one board led to style A, Salon Calm. On screen it read as muted, so a header, logo and colour board led to a fresher version: a near-white page, white cards and a clearer green. Six new logos didn't beat the simple wordmark, so it stayed.

Instead of a few sample products I rebuilt two complete brands from the live shop. Gehwol: 53 listings, one per size, became 34 products with the size as a choice and its own price and photo. Gelamour: 107 listings became 11, with 91 gel polish colours as one product, a swatch for each colour and a search by number.

36 of the 51 original findings were solved. But the new design had problems of its own, and I fixed the quick ones before calling it done.

A near-white shop with one green for everything you can do. Logged-in customers start with their regular products; every price is their own; every deal is counted for them.

Most orders happen on a phone between two clients, so I compared eleven tasks at the same screen size, the old shop next to the concept.

The board uses the prototype's own stylesheet, so it can't drift from the screens: colours with measured contrast, type, buttons, fields, labels and the product card.

• An audit of the live shop and a platform advice (WooCommerce, lean)
• A clickable prototype with prices per group, deals, an order list without payment and a staff page
• Two complete brands as data, ready to show the migration
• An audit of my own design, with the quick fixes done

Not done yet: no tests with users (a card sort and a remote task test with pedicurists are planned), the SnelStart link is unconfirmed, product photos and texts still come from the old shop.

AI crawled the shop, measured contrast and targets, wrote the prototype code and turned 160 product pages into data. I set the brief, chose every direction and rule, and checked the claims. One check caught a wrong number: the menu has 39 lines, not the 35 the first audit said.