I ran Lighthouse on my app and got a pile of warnings and scores. I don't know which ones matter or what to do first. Triage my Lighthouse (or PageSpeed) results: 1. Ask me to paste the report or the key numbers and opportunity list, then organize findings into: - Real user-facing problems (slow LCP, layout shift, blocked interactivity) with the specific element or resource responsible. - Easy mechanical wins (image formats, text compression, cache headers, unused CSS/JS) with expected payoff. - Noise to ignore: diagnostics that don't apply to my stack, opportunities with tiny estimated savings, lab-only artifacts that won't affect users. 2. For each real finding, give me the concrete fix, not a restatement of the warning. 'Properly size images' becomes which images, what sizes, and the code or settings change. 3. Prioritize: what improves the experience of an actual user on a mid-range phone over a mediocre connection, ranked by effort versus impact. 4. Note which lab findings I should verify with real-user data before spending time on them. Don't chase a 100 score. Don't recommend rearchitecting the app over a metric that doesn't affect users. The deliverable is a ranked fix list with the specific change for each item, plus the list of warnings you've told me to ignore and why.