Insights
Writing on Shopify engineering
- Render budgets: the number your theme should have had from day one Performance work fails because nothing owns the cost of a new section. A render budget makes that cost visible before it ships, and keeps it visible after the launch photos are taken.
- When a Shopify Function should replace the app you are paying for Discount, delivery and payment logic increasingly belongs in Functions rather than in a bolted-on app. Here is how to tell whether yours does, and what it actually costs to make the switch.
- Migrations do not fail on data. They fail on redirects. Product imports are the easy part of any platform migration. Organic traffic is lost in the URLs nobody mapped, the redirects nobody tested, and the fortnight after cutover that nobody watched closely enough.
- Headless is a trade, not an upgrade Hydrogen is excellent engineering. It is also an operational commitment that many brands are talked into too early, on the promise of performance that a well-built theme can usually deliver for a fraction of the ongoing cost.
- Write the data contract before you write the integration Most ERP and CRM integrations break on ownership, not on transport. Decide who owns each field first, design for events that go missing, and make the whole thing observable — the integration itself, not just the code, is the deliverable.
- Most Shopify A/B tests cannot conclude. Run them anyway? If your traffic cannot resolve the effect size you are looking for, the test is theatre dressed up as rigour. There is a better option, and it starts with doing the sample-size arithmetic before anyone touches the page builder.