Velocity Commerce
Web Development202412 weeks

Velocity Commerce

Frontend Engineering · Performance

Overview

Velocity's existing storefront was losing 34% of mobile visitors before they reached the product page. The culprit was a 6.2-second LCP on mobile — a legacy of accumulated technical debt and an image pipeline that hadn't been touched in three years. We rebuilt the frontend from scratch with performance as the primary design constraint.

ClientVelocity Commerce
ServiceWeb Development
Year2024
Duration12 weeks
RoleFrontend Engineering · Performance

Project Gallery

Storefront experience
Storefront experience
The Challenge

Speed as a conversion problem

Every 100ms of load time was costing Velocity an estimated $40,000 in monthly revenue. Their Lighthouse score was 31. Their largest contentful paint was 6.2 seconds on a mid-range Android device. The engineering team knew it was bad — they just didn't have the bandwidth to fix it.

The rebuild had to happen without disrupting the existing SEO rankings that were driving 60% of their traffic. Every URL, every redirect, every structured data schema had to be preserved or improved.

Our Approach

Performance budget from day one

We set a hard performance budget before writing a single component: LCP under 1.2s, TBT under 200ms, CLS under 0.1. Every architectural decision — server vs client components, image formats, font loading strategy — was evaluated against these constraints.

We rebuilt the image pipeline with next/image, AVIF format, and a custom CDN configuration that serves appropriately sized images based on device and connection. The product gallery went from 2.4MB to 180KB per page load.

The Solution

A storefront that earns its traffic

The new frontend is built on Next.js App Router with aggressive use of Server Components — only the cart, wishlist, and personalisation layers run on the client. The result is a storefront that renders the critical path in under 800ms on a 4G connection.

We rebuilt the checkout flow with a single-page architecture that eliminates full page reloads between steps. Combined with optimistic UI updates and skeleton loading states, the perceived performance is near-instant.

SEO was treated as an engineering discipline: semantic HTML, JSON-LD structured data for every product, a dynamic sitemap, and canonical URL management that preserved every existing ranking.

Outcomes

The numbers
that matter.

6.2s → 0.8s

LCP improvement

↑ 34%

Conversion rate

100

Lighthouse score

↑ 28%

Organic traffic in 60 days

Our conversion rate went up 34% in the first month after launch. The combination of performance engineering and CRO thinking is genuinely rare.

Priya Nair

CEO, Velocity Commerce

Helix AI

Next Project

Helix AI

→
SaaS Development2026

Start a project

Want results
like these?

Book a 30-minute call. We'll listen, ask the right questions, and tell you exactly how we can help.

20+ products shipped · 2+ years · No fluff