Chelsea Supplies

Chelsea Supplies is a UK online retailer of designer radiators, composite fencing and composite decking, and I was brought in to review the whole WooCommerce stack, then stayed on to make the database fast, build the two product configurators that launched their new product lines, and keep the advertising measurement chain accurate across every platform they spend on.

Key Takeaways

  • Took database queries from 3 to 4 seconds to under one second on a store carrying over 100,000 orders, through an HPOS migration, index keys across ten tables and a disciplined plugin cleanup.
  • Built the slatted fencing configurator that launched a new revenue line: style, colour, height and post options priced from the client's own specification sheets, feeding clean structured orders to the sales team and CRM.
  • Built the composite decking configurator on the same architecture, giving the business a repeatable pattern for configurable products.
  • Delivered four colour-specific ranking URLs off a single decking product, each carrying its own SEO title, meta description, share image and breadcrumb, without splitting the catalogue.
  • Rebuilt conversion measurement across Google Ads and TikTok with advanced matching, deduplication and verified purchase payloads, so paid media optimized on accurate data.
  • Supported the payment stack through a processor change, keeping checkout options and the sales workflow intact.

Organization

Chelsea Supplies Ltd trades from Bolton in the UK. The catalogue runs column, towel, electric and designer radiators, valves and accessories, plus composite fencing and composite decking as two newer revenue lines. The store carries a seven-figure order history with live card and instalment payments, a sales team working from a shared inbox, and a CRM that mirrors the product catalogue.

Challenge

The store was doing real transaction volume on a platform nobody could see inside of: no performance baseline, no visibility into what the database was doing, and an advertising stack spread across multiple vendors with no single owner of measurement. At the same time the business was launching two configurable product lines that off-the-shelf product pages could not represent.

Solution

  • A full stack review first, then a database pass: HPOS migration verified through a full day of live transactions, index keys added across ten tables, and background processes scoped down to what the store actually needs.
  • Product configurators specified directly from the managing director's pricing spreadsheets, with conditional options, per-option pricing and structured order data.
  • Rewrite-rule routing for colour landing pages, with the gallery, metadata and breadcrumbs all resolving per colour.
  • TikTok measurement built to the platform's events specification, with advanced matching, logged-in customers excluded, and duplicate-fire protection.
  • Google Ads conversion tracking verified in the browser, and the consent and server-side tracking layers kept consistent through vendor changes.
  • A written infrastructure plan giving the group a clear path to dedicated hosting and, longer term, a purpose-built commerce application for configurator-led products.

Results

  • Database queries under one second, from 3 to 4 seconds.
  • Conversion rate up an estimated 5 to 15%.
  • Paid media efficiency up an estimated 15 to 25%.
  • 10 to 20 hours a month of sales and admin time returned through structured configurator orders.
  • Organic visibility on the decking line up an estimated 10 to 20%.