9 guides
Almost every performance problem gets diagnosed before it gets measured, which is why so many of them get fixed in the wrong place. The guides in this cluster are ordered around measuring first.
The distinction that resolves the most confusion is lab versus field. A Lighthouse run is a single test on simulated hardware, and it is repeatable and useful for comparing changes. Field data comes from real visitors on real devices and networks, and it is what Core Web Vitals reports. The two disagree routinely, and neither is wrong when that happens; they are measuring different things. Reading a score without knowing which one you are looking at is how teams end up optimising something their visitors never experience.
From there the guides work through the layers where time actually goes: images and media, fonts, third-party scripts, render-blocking work, and the data fetching behind the page. Most sites are slow in one or two of these, not all of them, which is why the diagnosis is worth the hour it takes.
A note on scores. Our own commercial terms use a 90+ Lighthouse acceptance target on named pages, verified across recorded runs before handover. That is a lab acceptance criterion, not a promise about rankings, traffic, or field metrics after later changes.
To see where a specific site is losing time rather than reading about it generally, start with a free performance audit of your own site.
PandaCodeGen builds from $1,500 at a fixed price agreed before work starts, has no minimum project size, and the source code, design files and production accounts are yours at the end.
Topic introduction written by Hassan Jamal, Lead Engineer, from the scoping and handover work behind these guides. How claims here are sourced, dated and corrected is set out in our editorial policy.
Published May 13Updated 2026-08-1914 min read
Find the slow layer before paying for the wrong fix. This measurement-first guide covers field data, lab traces, templates, media, scripts, hosting, and platform constraints.
Published Apr 30Updated 2026-08-1914 min read
Twelve route-level performance diagnostics, how to measure field and lab evidence, and when optimization or migration is the more defensible investment.
Published Feb 17Updated 2026-08-1911 min read
An eight-step Lighthouse and Core Web Vitals process for improving representative pages without false 100-score promises.
Published Apr 16Updated 2026-08-199 min read
What LCP, INP, and CLS measure, how field and lab data differ, and how to investigate business impact without inventing a ranking or revenue number.
Published Mar 9Updated 2026-08-1910 min read
How Core Web Vitals fit within page experience, how to separate field and lab evidence, and why performance does not map to a fixed ranking or revenue outcome.
Published Aug 13Updated 2026-08-1911 min read
The web got measurably less accessible in 2026, ending six years of improvement. What the measured data says about failure types, platforms and legal exposure, and what it cannot tell you.
Published Jul 16Updated 2026-08-1324 min read
A source-linked map of 23 AI-search, rendering, structured-data, and web-performance terms, including what each term does and does not prove.
Published Aug 13Updated 2026-08-1323 min read
Plain-English definitions of the accessibility terms that appear in audits, quotes and procurement forms, with what each one does not prove and what to ask the person selling it to you.
Published Mar 13Updated 2026-08-0226 min read
Vercel Hobby is free for personal, non-commercial projects only. Compare its published usage guidelines with Vercel Pro pricing for commercial Next.js hosting in 2026.
These guides describe the decision. If you have made it and want the work scoped against your own site, the commercial terms, acceptance criteria, and exclusions are on one page.
See the free site audit