agency-wordpress-performance-engineer

Start by profiling with Query Monitor and Lighthouse to find where time actually goes, then fix the biggest costs first: bound slow WP_Query calls, trim autoloaded options, add a Redis or Memcached object cache, configure page caching for anonymous traffic, and defer render-blocking assets. Verify every change against mobile Core Web Vitals.

Identify the failing metric first: LCP usually traces to slow server response or an unoptimized hero image, INP to heavy JavaScript, and CLS to dimensionless images. Preload the LCP image, add explicit width and height to all images, defer non-critical JavaScript, and measure on throttled mobile.

Both work as persistent object cache backends via the object-cache.php drop-in, and the Skill supports either. Redis is more common on managed hosts and offers richer eviction policies like allkeys-lru. Verify the cache is actually hitting with a target above 90% on warm cache.

Yes, caching dynamic pages is a privacy risk, not just a bug. Cart, checkout, account, and logged-in pages must be explicitly excluded from page cache and CDN edge HTML caching, otherwise one user could see another user's data. Always verify the bypass at the edge.

A cache plugin pointed at the wrong layer hides symptoms without fixing costs. Common remaining causes include bloated autoloaded options loaded on every request, unbounded meta_query calls, heavy plugins dominating per-request PHP time, and render-blocking page-builder assets. Profile with Query Monitor to find the real cost.

Speed plugins cannot fix architectural problems like unbounded queries, missing indexes, or a heavy page builder, and misconfigured ones backfire through over-minification, broken jQuery deferral, or cached dynamic pages. Measurement-driven manual optimization with correct caching layers produces durable results.