Lesson 3.9 · 55 minSpeed and search basics

Purpose: record a baseline of the live site's speed before changing anything.

Context: the live site is https://[your-domain].

Done when:
- docs/lighthouse/ contains before.report.html and before.report.json for the
  home page, produced by:
  npx lighthouse https://[your-domain] --output html --output json
  --output-path docs/lighthouse/before --chrome-flags="--headless=new"
- you ran it three times and kept the run with the middle performance score;
- you give me a table from the JSON: the performance, accessibility, best
  practices and SEO scores; LCP, CLS and Total Blocking Time; which element
  is the LCP element; and the top five opportunities, each explained in one
  plain sentence.

Constraints: measure only; change no code. Default (mobile) settings.

Purpose: fix the biggest speed problems in the baseline without changing how
the site looks or behaves.

Context: baseline in docs/lighthouse/before.report.json. Read the image and
font guides in node_modules/next/dist/docs first.

Done when the top opportunities are fixed, and you have checked and reported
on each of these:
1. the hero image is requested at a size suited to the viewport and is the
   only eagerly loaded image; other images lazy-load; source files are no
   larger than needed;
2. only the font families and weights the design uses are loaded;
3. no unused libraries or imports remain, icons are imported individually,
   and only components that need interactivity run in the browser.

Constraints: no visual changes, no new libraries, no removed features. Do not
chase 100: stop when LCP, CLS and Total Blocking Time are in Lighthouse's
good range, and tell me what you left and why. Before reporting: build and
lint; compare screenshots at 375px, 768px and 1280px with the live site; list
each change and the baseline item it addresses. Then commit and push.

Purpose: measure the result of the speed fixes.

Context: the fixes are deployed. Baseline: docs/lighthouse/before.report.*

Done when: docs/lighthouse/after.report.html and after.report.json exist,
produced the same way (three runs, middle one kept), and
docs/lighthouse/summary.md has a before/after table of the four category
scores, LCP, CLS and Total Blocking Time, with the date, the URL, and one
line on what remains unfixed and why.

Constraints: same command and settings. First confirm that the latest commit
is the one deployed.

Purpose: make the site easy for search engines to read.

Context: siteUrl is in content/site.ts. Index: /, /privacy, /terms. Do not
index: /thank-you, /styleguide. Later sections add /admin and /api, which
crawlers should not crawl. Business facts are in content/site.ts and
docs/brief.md. Read the sitemap, robots and JSON-LD pages in
node_modules/next/dist/docs first.

Done when:
- app/sitemap.ts generates /sitemap.xml with only the three public pages, as
  absolute addresses built from siteUrl;
- app/robots.ts generates /robots.txt that allows crawling in general,
  disallows /admin and /api, and points to the sitemap;
- the home page includes JSON-LD of the fitting schema.org local-business
  type with only: name, description, url, telephone, email, address or area
  served, and opening hours if docs/brief.md states them;
- every indexable page has a canonical address.

Constraints: no ratings, reviews, prices or awards unless they are in
docs/content.md and visible on the page. Escape the JSON as the Next.js guide
describes. Before reporting: build; start the production server; show me the
full contents of /sitemap.xml and /robots.txt and the JSON-LD as rendered in
the home page's HTML; re-run Lighthouse and report the SEO score.