Gallery One · Old vs New Site · April 2026
An independent comparison of both Publifai audit reports
The new Gallery One site is the stronger site on every measurable axis we audited — desktop speed jumped from 45 to 99/100, mobile from 49 to 90/100, layout shift dropped from 0.845 to 0.002, and SEO went from 69 to a clean 100.
The honest caveat: the new site has 7 pages, the old site has 218. The visible 114-artist roster and the 48 individual artwork pages from the old site have not yet been ported over — they are the next round of work, not part of this comparison.


| Performance (desktop) | 45 | → | 99 | major win |
| Performance (mobile) | 49 | → | 90 | major win |
| SEO score | 69 | → | 100 | major win |
| Accessibility score | 80 | → | 93 | improved |
| Largest Contentful Paint (mobile) | 4.15s | → | 2.89s | improved |
| Largest Contentful Paint (desktop) | 3.97s | → | 0.76s | good |
| Cumulative Layout Shift (mobile) | 0.845 | → | 0.002 | major win |
| Cumulative Layout Shift (desktop) | 0.899 | → | 0.002 | major win |
| Homepage <title> tag | ✓ | → | ✓ | unchanged |
| Homepage meta description | ✗ | → | ✓ | fixed |
| Open Graph share image | ✗ | → | ✓ | fixed |
| Homepage H1 heading | ✗ | → | ✓ | fixed |
| Structured data (JSON-LD) | — | → | LocalBusiness, ItemList, BreadcrumbList, Person | fixed |
| Image alt-text coverage | 14/17 | → | 7/7 | n/a |
| llms.txt present | ✗ | → | ✓ | fixed |
| AI crawlers allowed | ✓ | → | ✓ | unchanged |
| Google Business Profile linked | ✗ | → | ✗ | still a gap |
On speed, desktop performance climbed from 45/100 to 99/100 and mobile from 49/100 to 90/100 — a 54-point and 41-point jump respectively. The biggest single change is layout stability: cumulative layout shift on the homepage went from 0.845 to 0.002 on mobile and from 0.899 to 0.002 on desktop, which means the new homepage no longer jumps around as it loads. Desktop largest-contentful-paint dropped from 4.0 seconds to 0.8 seconds. SEO score went from 69/100 to 100/100, accessibility from 80/100 to 93/100. The homepage now carries a real title, a 144-character description, an Open Graph image (so WhatsApp previews finally work), an H1, four types of structured data (LocalBusiness, ItemList, BreadcrumbList, Person), and an llms.txt file for AI crawlers — none of which existed on the source site. Image alt coverage on the homepage went from 14/17 to 7/7.
Mobile largest-contentful-paint improved from 4.1 seconds to 2.9 seconds, but that is still above Google's 'good' threshold of 2.5 seconds — the homepage hero image needs to be smaller and preloaded before mobile LCP is fully solved. The Google Business Profile is referenced via sameAs in the structured data but is not yet a visible 'View on Google' link in the footer, which is the same gap the old site had. And the catalogue depth — 114 artist pages and 48 artwork pages — has not yet been rebuilt. The new site has 7 pages to the old site's 218.
The old site's strongest asset was always its catalogue: 114 artists and 48 artworks all live, browsable, and individually addressable. It also had a healthy sitemap, a working robots.txt, canonical URLs on every artwork page, and an accessibility score of 80/100 — a solid baseline that the rebuild has now built on rather than replaced. The catalogue itself is the next round of work; the rebuild has put a stronger foundation underneath it, not erased what was there.
The rebuild has replaced a 218-page WordPress catalogue with a 7-page editorial foundation that is faster, cleaner, and properly structured for both Google and AI search. The real next question is not whether to keep the new site — every measurable axis says yes — but whether the next round of work should be content depth (rebuilding the artist and artwork pages) or visual polish (mobile LCP, GBP footer link, social handles).