A WooCommerce boutique was choking on JS sliders and a custom select. Swapping them for native CSS selects, scroll-marker carousels and scroll-state queries cut a second off Time to Interactive.
An agency gallery held 14 years of high-resolution assets and took 12 seconds to load. Caching the resolution URLs in a transient made it worse; generating the links from a CDN naming convention fixed it.
A team page that broke every time someone joined pushed me to rebuild it with sibling-index(), a double rotation transform and trig inside calc(). The ring now spaces itself, cut-out mask included.
A bespoke furniture shop was A/B testing gray buttons against other gray buttons. The fix was per-page color palettes injected from post meta into CSS variables, so the products set the tone instead of the theme.
A client spent six months AI-powering their onboarding flow and ended up with hallucinated shipping dates. The prompt was never the problem, the context was. Here is what UX has to own before the model ships.
An 85% churn score with no reason attached is a support ticket waiting to happen. What worked instead: a plain-language because statement, the charts behind a link, and code that maps feature names to words.
A client once brought me Dribbble-ready mockups and no answer for the 45% drop-off at his shipping step. He had a goal, not a strategy. Here are the six UX strategy components I look for before anyone writes code.
Our WooCommerce configurator passed the automated audits and still stopped a screen reader user at step one. Accessible UX research, and Michele A. Williams’ new book on it, is what got us out of the patch-and-move-on habit.