Back to projects

E-commerce Platform

Velvet Vogue

The storefront was only half the challenge. The real work was keeping every product, variant, image, order and stock update trustworthy behind it.

Velvet Vogue project showcase
RoleFull-stack developer and designer
PeriodFebruary 2026 - March 2026
StatusCompleted and deployed

Project overview.

Velvet Vogue is a PHP and MySQL fashion store connecting product discovery, variant-aware shopping, orders, inventory and administration in one system.

12relational database tables
100validated size-colour variants per product
20validated product images per submission

One retail workflow, split across too many moving parts.

Velvet Vogue was built for a fashion retailer needing both a customer storefront and an independent management system. Customers needed clear discovery and ordering, while administrators needed reliable control over products, variants, stock and order progress.

The difficult part was keeping related data trustworthy. One product could contain several sizes, colours, stock values and images, so a failed upload or invalid variant needed to leave no incomplete records behind.

Start with what shoppers and administrators actually need.

QuestionnaireRequirements analysisWireframing

Questionnaire feedback made discovery speed a priority: shoppers wanted to combine size, colour and price criteria without repeatedly opening separate pages.

Design the experience, then protect the data behind it.

The strongest decisions improved catalogue discovery, treated variants as real inventory and made product creation recover safely from failure.

Combined filters instead of isolated searches

ChallengeA user might know the size, colour and price range they want, but separate controls force repeated searches.

DecisionI implemented multi-criteria filtering so several attributes can work together in one catalogue view.

Why it mattersIt shortens discovery and mirrors how people actually describe clothing: not one attribute at a time, but a combination.

Variants as first-class inventory

ChallengeA product-level stock number cannot explain that a black medium is available while a blue medium is sold out.

DecisionEach size and colour combination receives its own SKU, stock count and optional price adjustment.

Why it mattersVariant-level records make availability and ordering precise instead of treating every product as one undifferentiated item.

All-or-nothing product creation

ChallengeThe product, variants and images are written in separate steps. Failure in the middle could leave orphaned data or files.

DecisionI used database transactions, server-side validation, rollback and uploaded-file cleanup as one controlled workflow.

Why it mattersThe admin either gets a complete product or a clean failure. Partial success is not acceptable for inventory data.

A complete retail loop, not a collection of disconnected screens.

Customer and administrator actions now share the same product, inventory and order data from listing through delivery.

Fast product discovery

Search, categories and combined filters help customers narrow products by the attributes that matter.

Variant-aware shopping

Product pages expose size, colour, stock and pricing choices before an item reaches the cart.

Customer order journey

Accounts, cart, checkout, invoices and tracking provide a clear path after product discovery.

Operational admin tools

Products, variants, inventory, customers, orders, reviews and enquiries are managed from one dashboard.

  1. 01

    Admin creates product

  2. 02

    Variants receive stock

  3. 03

    Customer select product

  4. 04

    Add to cart or checkout

  5. 05

    Customer tracks progress

Keyboard controls: Space or K plays and pauses. Left and right arrow keys move five seconds. Home and End move to the beginning and end. F toggles fullscreen.

Video paused

A deployed retail platform connecting customer and administrator workflows.

The application was completed, functionally tested and deployed online. Its strongest result is the end-to-end retail workflow and the reliability built around product creation.

Live deployed customer-facing website

Twelve-table relational data model

Validated submissions supporting up to 100 variants and 20 images

Customer and administrator flows tested across the main tasks

Reflection

The biggest lesson: backend reliability is part of the user experience.

The strongest part of the project is the connection between a custom visual experience and a structured operational backend. Product discovery, inventory and administration operate as one connected workflow.

  • Model product variants before designing their interface
  • Treat uploads and database writes as one recovery problem
  • Use feedback to simplify discovery rather than add decoration
Next project

ETCP

A 50-screen React frontend where traveller and provider journeys evolved through research, sketches and usability feedback.