3 posts
Lighthouse 98, INP under 150ms, LCP under two seconds — on a real Plus store with twelve apps installed and a merch team that won't stop adding Hero sections. The wins, the losses, and the audit method that finds the next bottleneck before the merchant complains.
Six months. Eight Plus stores. Fourteen A/B tests. The four interventions that won every time, the three that lost, and the four that depend — with sample sizes, lift, and what you should actually run on your store.
Forty-plus Liquid themes, twelve Hydrogen builds. The seven questions we ask in kickoffs to decide which platform layer fits — and the cost reality most agencies don't put in the proposal.