Retail software case study

Ecommerce website development that grew online orders 2.4x

A retailer's ageing online store crawled on phones and fell over during its biggest sales. We rebuilt it as a fast, cloud hosted store that stays up when the crowds arrive. Within a year of launch, online orders had grown 2.4 times.

A shopper paying by card for an order on an online store

A faster store that stays up at peak

A slow, ageing store rebuilt in the cloud to be fast and reliable.

2.4xonline orders within a year

What we built

  • Faster pagesPages load fast on phones and desktops
  • Marketing controlPromotions go live without a developer
  • Steady at peakServers added automatically as traffic rises
  • Stock accuracyOrders go straight to the warehouse
Discuss a similar project
  • Fixed estimate before any work begins
  • Reply within one business day

The problem: a store that could not keep up

The store ran on an older platform the business had outgrown. Most shoppers came in on their phones, and on a phone the pages loaded slowly. In the busiest sales, when traffic jumped, checkout slowed to a crawl or stopped working altogether, on the very days the retailer expected to sell the most.

Changing anything was slow too. Every new banner or promotion needed a developer, so the marketing team waited days for simple updates. The stock levels shown on the site also drifted away from what the warehouse actually held, and some orders had to be cancelled.

Six months to launch

How we worked phase by phase

Project phases in months

Study how people shopWeeks 1 to 3: analytics and interviews showed where shoppers dropped out
Design for speedWeeks 3 to 7: mobile first pages, simple filters and a short checkout
Build and move dataMonths 2 to 5: two week build cycles, with products, accounts and orders moved in rehearsed batches
Load test at peakMonth 5: load tests imitating thousands of shoppers, including a major sales day surge
Launch and tuneMonth 6: trading switched to the new store, then pages were tuned using live data
StartMonth 2Month 4Month 6
Research and designBuild in two week cyclesTesting and launch

Daily trading carried on as normal while data moved across in rehearsed batches.

Under the hood

How the new ecommerce website works

How the online store fits together

PeopleWho uses it
ShoppersPhone and desktopMarketing teamPage editor, no coding
Shop windowWhat shoppers see
Next.js storefrontFast, readable pagesContent delivery networkServed from nearby
EngineLinked by APIs
Key partCommerce servicesNode.js basket and checkoutRedis cachePopular data kept readyPayment providerHolds card details
Systems
Stock and fulfilmentExisting systemsAWS autoscalingAdds servers at peaks

We used what's called a headless commerce design. The shop window your customers see is kept separate from the engine behind it, which handles products, prices, stock, orders and payments. The two talk through APIs (application programming interfaces, the standard way one system asks another for data). That means the marketing team can publish new pages without going anywhere near the checkout.

When traffic rises, Amazon Web Services (AWS) adds servers by itself. Orders go straight through to the warehouse, so nobody has to type them in again.

Protecting search rankings

Search engines and shoppers' bookmarks still pointed at the old product and category addresses. We set up permanent redirects so each old address sends visitors and search engines to its new page, and we checked the site in search tools before and after launch.

Security and quality checks

Card details never touch the store: they stay with the payment provider. Staff only get the admin access their role needs. Before any change reaches shoppers, it passes automated tests in a CI/CD pipeline (an automated process that tests and releases each update).

Measured after launch

Results after launch

Online orders (before = 100)

Before100Year before the rebuild: index 100
After240Within a year of launch: index 240, which is 2.4 times as many orders
Old storeNew store

Online orders grew 2.4 times within a year of launch.

What changed for the retailer

  • Faster pages. Product pages now load quickly, whether shoppers are on a phone or a laptop.
  • Steady at peak. Since launch, the store has stayed up through every peak trading period.
  • Marketing in control. A promotion someone suggests on Monday can be live the same week.
Next.js React Node.js Amazon Web Services Redis

What we learned

Speed wasn't the only thing that moved sales. Letting the marketing team change the store themselves did as much for orders as the faster pages did.

If your online store is holding you back, our custom software development team can plan a rebuild, and our cloud migration services cover hosting that scales with demand. See also the software we build for retail and commerce and our guide to what custom software costs.

Has your store outgrown its platform?

Tell us where your store lets you down, and we'll agree a fixed estimate before we start.