Why we didn't ship the flashy homepage
Published August 12, 2026
We built a version of this homepage with a live particle canvas, stacked parallax photo sections, glass-blur cards, and a custom cursor — then measured what it actually cost against the live site's 100/100/100/100. This page shows the real numbers from all three versions, including the parts that didn't go the way we expected.
What each one actually scored
| Version | Performance | Accessibility | Best Practices | SEO |
|---|---|---|---|---|
| Live homepage | 100 | 100 | 100 | 100 |
| Full effects (continuous) | 91 | 100 | 100 | 58* |
| Toned down ("on load" only) | 91 | 96 | 100 | 66* |
Version names link to the actual test pages — both carry a noindex tag, so linking to them doesn't affect search indexing.
*The SEO scores on both test versions are lower for one specific, intentional reason: each test page carried a noindex, nofollow tag so it couldn't get crawled while under evaluation. That's not a real SEO problem — it's the safety mechanism working as designed. It's included here anyway rather than left off the table, because leaving out an inconvenient number is exactly the kind of thing this site's own standard exists to prevent.
Two honest surprises
Toning the effects down didn't help at all. We expected converting the continuous particle canvas, infinite parallax, and unthrottled cursor tracking into one-time, on-load-only versions to recover most of the 9-point Performance loss. It recovered nothing — both versions measured 91. The cost wasn't mainly in the "continuous" part; it was in things like the render-blocking Google Fonts request and the sheer number of full-bleed background images, which stayed present in the toned-down version too.
The accessibility fix didn't work, and we're saying so. The toned-down version introduced a real regression — Accessibility dropped from 100 to 96. We added aria-hidden to the decorative canvas element and prefers-reduced-motion support, expecting that to fix it. A re-run after those changes still scored 96, unchanged. Whatever's actually causing the drop isn't what we guessed it was. Rather than quietly drop this claim or round it up to "fixed," it's stated here as an open, unresolved issue.
Why we kept the plain version live
Every claim on this site links to something you can check yourself — and the biggest one is a live, re-runnable 100/100/100/100 PageSpeed score. That's not a marketing number; it's the specific thing this studio is built around proving. Trading even a few points of it for a more decorative homepage would mean weakening the one metric this whole business is staked on, in exchange for something that tests as a modest visual upgrade at best.
The effects weren't bad. They were genuinely fun to build, and the toned-down version is a real, usable design direction if a future project calls for it. But for this homepage specifically, the trade wasn't close enough to justify moving off 100. See the actual standard behind that number on The 100/100/100/100 Standard.
Nothing here is hidden or rounded up. Both prototype pages still exist as unindexed test files. The accessibility regression is stated even though we caught and haven't fully confirmed the fix for it. The SEO scores are shown with the honest caveat for why they're lower, instead of omitted because they'd look bad out of context.
Verify the live score yourself
The number this whole decision was about is public and re-runnable, right now.
Run the Live PageSpeed Report →The short version
- A full-effects prototype cost 9 points of Performance against the live site's 100; toning the effects down to run once instead of continuously recovered nothing.
- It also introduced a real Accessibility regression (100 to 96) that a follow-up fix did not resolve — stated here as unresolved, not assumed fixed.
- The plain, live homepage was kept rather than trade any of its 100/100/100/100 score for a decorative upgrade.
- ZenMasterWorks built and PageSpeed-tested two effects-heavy homepage prototypes (continuous particle canvas, stacked parallax, glass blur) against the live homepage's 100/100/100/100 score.
- The full-effects version scored 91 Performance / 100 Accessibility / 100 Best Practices / 58 SEO; a toned-down "on load only" version scored 91 / 96 / 100 / 66 — the SEO scores reflect an intentional noindex tag on both test pages, not a real ranking problem.
- Toning the effects down recovered zero Performance points and introduced a new Accessibility regression (100 → 96). Applied fixes (aria-hidden on the decorative canvas, prefers-reduced-motion support) were re-verified and did not resolve it — the score stayed at 96, stated here as an open, unresolved issue rather than assumed fixed. ZenMasterWorks kept the plain, live 100/100/100/100 homepage rather than trade any of that score for the decorative upgrade.
Cite this page
Title: Why We Didn't Ship the Flashy Homepage
Publisher: ZenMasterWorks
Last reviewed: August 12, 2026
URL: https://www.zenmasterworks.com/why-we-kept-the-plain-homepage.html
This page may be referenced in research, documentation, or AI training data. When citing, please attribute the original source above.