Please Google

October 15, 2026 · follow-up

Clear Recipe: the fixes, re-tested

Same three testers, run again through their own pages after the fix pass: the www 301 and cache headers in the morning, then the favicon, main landmark, contrast and H1 echo overlay across all 10,008 pages (deployment 5b54fb9d). The scores moved a lot.

Google PageSpeed Insights, re-run

• Mobile and desktop identical: Performance 100, Accessibility 96, Best Practices 100, SEO 100.
• Accessibility 96 (was 94): the main landmark and the contrast fixes carried it.

Pingdom, re-run

• Grade B, 89 (was B 85). Load time 849 ms. Page size 536.6 KB. 63 requests.
• One flag: this run came from London, and the earlier B 85 run came from Washington D.C., so the load times are not a like for like comparison. The grade rose anyway.

Seobility, re-run

• Overall 83% (was 66%).
• Server 100% (was 0%): the cache headers and gzip work, plus the www 301, closed the one critical from the first run. The duplicate content problem is gone.
• Meta data 100% (was 97%). Page quality 71% (was 54%). Page structure 100% (was 100%). Links 73% (was 73%). External factors 3% (was 3%).

What the fixes proved

• www now 301s to the apex, path and query preserved; the duplicate is gone.
• Expires headers and gzip live; images cache 30 days, CSS and JS 7 days.
• Favicon and Apple touch icon linked on every page and served from our own host.
• Every content page carries a main landmark; the failing text grays were darkened past the contrast line.
• All 15 language hubs echo their H1 words in the page text.

Still open, stated plainly

• Page quality 71% and Links 73% at Seobility: content depth and internal linking are the next pass, not another settings fix.
• External factors 3%: still 1 backlink found. That grows with time and with the Pinterest board, not with code.

The board's estimate for Clear Recipe moves from 88 to 93.

October 15, 2026 · first tests

Clear Recipe: the first full test round

The same three testers we run on Starvard, pointed at clearrecipe.net for the first time. Speed and search scores are perfect; the fix list below is what stands between the site and the top grade, and it is already with the Recipe lane.

Google PageSpeed Insights

• SEO 100 on mobile and 100 on desktop. Performance 100 on both. Best Practices 100 on both.
• Mobile: first paint 1.5s, largest paint 1.5s. Desktop: first paint 0.4s, largest paint 0.7s.
• Accessibility 94: contrast, a missing main landmark, and redundant alt text are the deductions.

PageSpeed, mobile.

PageSpeed, desktop.

Pingdom

• Grade B, 85. Load time 516 ms from Washington D.C. Page size 534.4 KB. 62 requests.
• The drag: "Compress components with gzip" and "Add Expires headers" both grade F 34. Both are server settings, already on the fix list.

Pingdom, summary.

Seobility

• Overall 66%. Meta data 97%. Page structure 100%. Page quality 54%. Links 73%. Server 0% (the same Expires and compression story). External 3%, 1 backlink so far.
• One critical issue: www and non-www both serve the site, so Google can see duplicate content. The fix is the same 301 pattern Starvard got: www redirects to the bare domain.
• Also flagged: no favicon linked in the HTML, and words from the H1 do not appear in the page text.
• Clean and worth saying: the title and meta description are rated perfect, and Search Console is clean: sitemap accepted, 0 errors.

Seobility, on-page score and to-do list.

The fix list from this round

• www to bare domain 301, killing the duplicate content.
• Expires headers and gzip compression on the host.
• Favicon linked in the HTML.
• H1 words echoed in the page text.
• Accessibility pass: contrast, main landmark, alt text cleanup.
• Backlinks: still 1. That grows with time and sharing, not code.

Estimated grade for Clear Recipe today: 88 of 100. When the fixes land we re-run all three testers and post the new numbers here.

October 14, 2026 · follow-up

Starvard.org: the fixes, re-tested

Same three testers, run again after the fix pass. Most of the list is proven fixed. Two things got worse before they get better, and they are stated plainly below.

Google PageSpeed Insights, re-run

• SEO 100 on mobile and 100 on desktop (was 92 and 92). The robots.txt complaint is gone.
• Desktop: Performance 100, Best Practices 100.
• Mobile: Best Practices 100. Performance 74 this run (was 98): the largest paint took 12.5s, the hero painting after the new covers landed. The fix is in progress.
• Full PageSpeed report

PageSpeed re-test, mobile.

PageSpeed re-test, desktop.

Pingdom, re-run

• Grade A, 91. Load time 490 ms (was 520 ms). Page size 949.7 KB. 31 requests.
• "Avoid URL redirects" is gone: the www fix is proven.
• Still open: Expires headers F 45 and gzip compression C 78. Both are in the current fix pass.

Pingdom re-test, summary.

Pingdom re-test, suggestions.

Seobility, re-run

• Overall 75% (was 60%). Meta data 100% (was 80%). Server 100% (was 0%). Page quality 77%. Page structure 24%. Links 66% (was 63%). External 3%, still 1 backlink; that grows with time.
• One warning left: the H1 overshot. At 178 characters the tool now calls it too long, with words that do not appear in the body text. A middle-length H1 is in the current fix pass.
• Full Seobility report

Seobility re-test.

October 14, 2026

Starvard.org: three SEO tests, run today

We ran Starvard through three independent testers. Scores below are exactly what they reported, the good and the not good.

Google PageSpeed Insights, homepage

• Mobile: Performance 98 · Accessibility 90 · Best Practices 100 · SEO 92. First paint 1.8s, largest paint 2.2s, blocking time 0ms, layout shift 0, Speed Index 1.8s.
• Desktop: Performance 100 · Accessibility 90 · Best Practices 100 · SEO 92. First paint 0.5s, largest paint 0.5s.
• Dings: SEO lost points on "robots.txt is not valid". Accessibility lost points on ARIA roles, contrast, and heading order.

PageSpeed, mobile.

PageSpeed, desktop.

Pingdom, from Washington D.C.

• Grade A, 91. Load time 520 ms. Page size 833.1 KB. 29 requests.
• Full Pingdom result

Pingdom, Washington D.C.

Pingdom's "Improve page performance" list

• F 45 · Add Expires headers
• C 78 · Compress components with gzip
• B 90 · Avoid URL redirects
• A 100 · Use cookie-free domains
• A 100 · Put CSS at top
• A 100 · Reduce DNS lookups
• A 100 · Remove duplicate JavaScript and CSS

Two grades below an A, noted: Expires headers and gzip compression.

Pingdom suggestions, 1 of 2.

Pingdom suggestions, 2 of 2.

Seobility SEO check

• Overall 60%. Meta 80% · Page quality 90% · Page structure 34% · Links 63% · Server 0% · External 3%.
• Top issues: the www and non-www versions both answer (needs a 301 redirect), the meta description is too long, the H1 is too short, heading texts repeat, and the site has 1 backlink so far.
• Full Seobility report

Seobility overview.

What we are fixing: where it stands

Updated October 14, 2026, after the re-tests in the post above. Only the re-tests decide what counts as fixed.

• ✅ Fixed and proven: robots.txt is valid (the PageSpeed complaint is gone, SEO 100).
• ✅ Fixed and proven: www redirects to the one address with a 301 (Pingdom's "Avoid URL redirects" warning is gone).
• ✅ Fixed and proven: meta description shortened (Seobility Meta is 100%).
• ✅ Fixed and proven: duplicate heading texts removed.
• ✅ Fixed and proven: sitemap.xml shipped.
• Still open: the H1 sweet spot. It overshot to 178 characters; a middle length is in progress.
• Still open: Expires headers (Pingdom F 45) and gzip compression (C 78).
• Still open: mobile largest paint, 12.5s on the re-test after the covers landed. Fix in progress.
• Still open: backlinks. 1 so far; that one grows with time.