Skip to content

Selected work

Six builds around real business problems. Read them as case studies in technology and business thinking: the interesting part of every one of these was the decision made before any code was written.

Case studies

Case studies

6 of 6 shown

  • Fashion retail · D2C

    Powerlook

    Challenge
    A fast-growing fashion brand whose storefront had to keep getting faster while the catalogue and the traffic behind it kept moving.
    Approach
    Headless architecture: the storefront decoupled from the commerce engine and served over a storefront API. Front-end releases stop waiting on back-end ones, and performance becomes a budget the front end can actually hold rather than an outcome it inherits.
    Technology
    HeadlessStorefront APIPerformance
    The Powerlook headless commerce storefront, a mens fashion brand
  • Horticulture · D2C

    Future Roots

    Challenge
    An online plant store selling a product that is fragile, seasonal and hard to photograph, to customers who abandon at any friction in the buy.
    Approach
    Shopify, chosen for speed to market rather than for flexibility, with the customisation spent where it converts: OTP login, a one-page checkout through GoKwik, and product pages built for a catalogue that does not behave like apparel.
    Technology
    ShopifyGoKwik checkoutCustom PDPOTP login
    The Future Roots Shopify storefront, an online plant store
  • Health & fitness · Retail

    S3Buy

    Challenge
    Sports nutrition is a category where authenticity is the purchase decision and delivery speed is the repeat one. The storefront has to carry both.
    Approach
    A commerce build organised around catalogue clarity and a short checkout, on the principle that in a trust-led category the product page is doing the selling and everything after it should get out of the way.
    Technology
    Commerce buildCatalogueCheckout
    The S3Buy storefront, an online sports nutrition retailer
  • B2B procurement · Platform

    Nxtby

    Challenge
    B2B buying is not a cart. It is a request, a quote, an approval chain and a budget holder, and a consumer checkout models none of that.
    Approach
    A custom procurement platform built around the approval structure rather than around the catalogue: purchase requests, quotes and multi-level sign-off with role-based permissions, wired into the systems that already held the order data.
    Technology
    B2B workflowsApprovalsRole-based permissionsIntegrations
    The Nxtby B2B procurement platform
  • Multi-category retail · Marketplace

    ShopUnicore

    Challenge
    A wide multi-category catalogue, which is the point at which catalogue size stops being a number and starts being an architecture problem.
    Approach
    Magento 2, which is the right answer precisely here and the wrong one for a small single-store catalogue: the platform earns its complexity when the category tree, the attribute model and the scale are all genuinely hard.
    Technology
    Magento 2Multi-categoryScale
    The ShopUnicore Magento 2 multi-category marketplace
  • Manufacturing · Engineering

    Synergy Water Slides

    Challenge
    An engineering manufacturer whose buyers research for months and never fill in a form, on a site that search could not read.
    Approach
    A brand platform built for discovery rather than for transaction: technical content structured so a crawler and a buyer can both follow it, and an enquiry path short enough to survive a long consideration cycle.
    Technology
    Web platformTechnical SEOContent
    Outcome
    3× traffic growth and 45% better engagement
    The Synergy Water Slides engineering brand platform

A note on proof

No client logo wall, no invented percentages, and no testimonials. One of the six has a published measured result and it is the only one showing a number. The other five say what was built and stop there, which is less impressive and considerably more useful.

The same rule governs the filters. There is no published AI case study with a measured outcome yet, so selecting that facet returns nothing rather than returning something reframed to fill it.

A replatforming you have been putting off, a checkout that loses people you already paid for, a catalogue that has outgrown its architecture, or an AI idea that needs someone to tell you which half of it is real.

Yuvraj RauljiRead my thinking