Why Headless Shopify Finally Makes Sense for Brands Under $5M

Headless Shopify storefronts used to be enterprise-only projects. Large teams, long engagements, massive overhead. Only brands doing $10M+/yr could justify the investment.

That's changed. We migrated a growing Shopify store from Liquid to Hydrogen — one developer, AI-assisted throughout. Mobile Lighthouse score went from 29 to 91. LCP dropped from 23.1 seconds to 2.9 seconds. Every Core Web Vital now passes.

Here's what the build looked like, what made it possible, and what it means for brands stuck on slow storefronts.

The old model: why headless was inaccessible

Until recently, a custom headless storefront meant hiring an agency or building an in-house team. The typical engagement:

• Custom React app from scratch: Storefront API integration, component library, cart logic, checkout flow, search, filters. 600-800 hours of development work.
• Large teams: Minimum 3-5 people — a tech lead, 2 frontend engineers, a PM to coordinate, a QA engineer to test. More likely 6-10 for a production-grade build.
• Heavy overhead: PM coordination, account management, agency margin, bench staff between phases. A significant fraction of every engagement never touched the actual store.

The result was a technically superior store — but only if you could afford the team and timeline. For most brands under $5M/yr, it wasn't an option.

What we actually did

Crate Clothing is a New Zealand streetwear brand doing over $500k NZD in annual revenue across nearly 4,000 orders — 86% revenue growth year-on-year, driven by original designs and standout artist collaborations. 79% of their traffic is mobile.

The problem was structural. Their Liquid storefront had accumulated years of app script debt: 20 third-party scripts loading from 8 external domains on every page. The homepage had a mobile LCP of 23.1 seconds. Total Blocking Time was 7,300ms. Lighthouse score: 29 out of 100.

With 79% mobile traffic and a 23-second load time, the store was functionally broken for most of its visitors. Homepage average engagement time was 19 seconds — meaning a significant share of users bounced before the main content even rendered. Mobile conversion rate: 0.90%.

One lead developer, AI-assisted throughout. 595 commits. 120 pull requests. 199 source files. Approximately 22,500 lines of code.

How one developer replaced an agency team

Four things changed simultaneously. Any one of them alone would have improved efficiency. All four together is what made a one-person build viable.

1. Hydrogen maturity

When headless Shopify first emerged, you were building custom infrastructure: your own React app, your own API client, your own cart architecture, your own hosting setup. Hydrogen didn't exist yet. You were inventing conventions, not following them.

Today, Hydrogen ships with cart logic, Storefront API hooks, server-side rendering, streaming, and edge deployment via Oxygen — all built in. The framework does what used to take months of custom engineering. A developer starts with working infrastructure and builds from there.

2. AI-native development

Roughly 40-50% of the work on this project was tasks AI handles better and faster than a human: test coverage, CSS migration from Liquid to Tailwind, boilerplate component generation, API schema documentation.

The lead developer focused on architecture decisions, UX judgment calls, and client-facing work — the parts that require a human. AI handled the implementation volume. The result was one-person output at what used to require 3-4 people.

3. Zero overhead

Agency projects carry significant non-engineering cost: PM hours, account management, team coordination, bench staff between phases, and margin on top. Rough industry estimate: 30-40% of an agency engagement is overhead that never touches your store.

One developer. No overhead. Every hour goes into the build.

4. Strict scope

This was a migration, not a reinvention. Same brand. Same products. Same checkout. The goal was a structurally better storefront — faster, more maintainable, purpose-built for mobile. Every feature on the scope had a clear ROI justification.

Out of scope by design: custom design language, new features that didn't exist on Liquid, integrations the store wasn't already using. Discipline on scope is underrated as an efficiency lever.

What the performance gains are worth

The performance improvements aren't just vanity metrics. They translate directly to revenue.

Google and Deloitte research associates reducing LCP from 3-5s to under 2.5s with a 5% or greater conversion rate lift. At $506k NZD annual revenue, a 5% lift translates to roughly $25,300 NZD in additional revenue per year.

That's before accounting for improved paid social ROAS (better landing page = lower effective CPA), SEO ranking improvements from Core Web Vitals now passing, or the eliminated app subscription costs from functionality now built natively.

We don't have post-launch conversion data yet — the store went live in March 2026, and a meaningful before/after comparison requires 30 days of data. But the performance lever is well-documented. The math is there.

What this means for brands stuck on slow storefronts

The brands most damaged by slow performance are mid-market: doing $1-5M/yr, spending serious money on Meta and Google ads, converting below their category benchmark. Most have Lighthouse scores in the 20-50 range. Most have mobile LCP above 4 seconds.

Until recently, a headless migration required the kind of team and budget only enterprise brands could justify. That barrier has collapsed. Hydrogen maturity, AI-assisted development, and lean execution models mean a single experienced developer can deliver what used to require a full agency team.

The question isn't whether headless is worth it — the performance data answers that. The question is whether your store's performance is actually the bottleneck. Run the audit first.

The ceiling on this model

This approach has real constraints worth naming.

It works when performance is the problem. If your store is slow and performance is killing conversions, a migration recovers real revenue. If your store is fast and conversions are low for other reasons (pricing, product-market fit, creative), a Hydrogen rebuild won't help. Diagnose before you build.

Complexity scales effort. Crate Clothing has a clean product catalogue, standard checkout, and well-understood integration requirements. A store with complex B2B pricing, subscriptions, custom checkout modifications, and 40 third-party integrations will take longer. Simple stores migrate fastest.

Post-launch matters. A faster store gives you a better baseline. But the real return comes from what you do with it — data-driven optimisation, conversion testing, feature iteration. The build is the beginning, not the end.

If you're a founder forwarding this to someone

The question is simple: Is your Shopify store fast enough to convert at the rate your ad spend deserves?

Run your homepage through PageSpeed Insights. If mobile Lighthouse is below 50, there's a structural problem. If LCP is above 4 seconds, you're losing conversions before visitors see your product. If TBT is above 500ms, the page is functionally blocked on mobile.

We audit stores and tell you exactly what's costing you conversions — and whether a Hydrogen migration is the right fix or a cheaper path gets there. If there's nothing worth fixing, the audit is free.