AYVA.

App-stacked Shopify stores lose money to latency.

Industry pattern: the average Shopify store runs 12–18 third-party apps. Each adds CSS, JS, and DOM weight. Result: mobile Lighthouse scores in the 30–50 range, bounce rates above 60%, and conversion drop-off of ~7% per additional second of load time (Google data). The brand's ask: a Scandinavian-minimalist store where every byte and every millisecond is owned. That eliminates 80% of the shortcuts most Shopify developers take.

Fork Dawn. Rewrite every section. Carry nothing forward.

What ships: hand-written Liquid for homepage, collection, and about-page sections; custom section schema exposing merchant-friendly settings (non-technical staff update copy and layout from Shopify admin); hand-tuned CSS under a 50 kb budget per section; responsive image srcset pipelines; deferred fonts; zero third-party JavaScript. 50 products migrated with metafields intact; pre- and post-migration SKU diff verified none slipped.

What it replaces: a typical $2,000–5,000/year third-party app stack (reviews, cart, upsell, wishlist, loyalty) — every feature that would normally come from an app was built directly into the theme, converting recurring SaaS subscriptions into a one-time theme investment.

Non-obvious decisions.

• Zero-app discipline — every feature the brand asked for (cart upsell, dynamic badges, filter logic) was hand-coded in Liquid + vanilla JS rather than bolted on via app.
• Performance budget — CSS budget was set at 50 kb and enforced manually per section. No section ships that breaks it.
• Migration safety — catalogue was diffed pre- and post-migration with a Python script; not one SKU slipped.

Why this one matters.

A Shopify build that proves the counter-intuitive thesis: if you want a store that loads fast, converts well, and stays maintainable — you need a developer who can actually write Liquid, not a theme-buyer. AYVA is the clearest example of that on the portfolio.