Website Performance & Speed Optimisation Services.
We audit real pages — not a PageSpeed screenshot of the homepage — and ship a fix list you can see move the needle in Search Console and GA4. Core Web Vitals, not vanity scores; mobile first, because that's where most traffic lives. No 40-plugin WordPress Jenga.
What the audit catches
in the first week.
The same six failure modes show up in almost every site we audit. Here's what we usually find on day one, and what we replace them with on day four.
- What usually happens
Hero image is a 2.4MB JPG loaded full-size on mobile, blocking LCP for 3–5 seconds.
What we do insteadResponsive WebP/AVIF, served in the right pixel density, preloaded. LCP under 2.5s on mobile 4G.
- What usually happens
Six blocking scripts in <head> — analytics, chat, A/B tool, pixel, tag manager — all before paint.
What we do insteadOne GTM, deferred. Everything else is async or lazy. First paint is uncluttered.
- What usually happens
Render-blocking CSS from three theme files, four plugin stylesheets, and a builder's runtime.
What we do insteadCritical CSS inlined, the rest deferred. Unused CSS purged. Render starts within the first 500ms.
- What usually happens
INP is 500ms+ because a carousel library is doing layout work on every scroll.
What we do insteadInteraction handlers audited. Long tasks broken up. INP consistently under 200ms on mid-tier Android.
- What usually happens
Cache is 'installed' but the page cache is empty, minify is fighting a CDN, and HTTP/2 is off.
What we do insteadCache layer configured to the host. Minify + combine disabled where it's hurting. HTTP/2 + Brotli on at the edge.
- What usually happens
PageSpeed homepage scores 92; the product pages score 34 and nobody's looked at them.
What we do insteadEvery template audited against real traffic. The pages that earn the revenue get the attention.
Four principles.
No slogans.
The score is not the goal
A 92 on PageSpeed that drops to 34 the moment a user logs in or opens a product page isn't a win. We optimise for the pages that carry real traffic and real revenue, and we measure on real user data.
Mobile first, mid-tier Android
We test on a throttled Moto G, not your M1 MacBook. If the site is fast on a 2021 Android on 4G, it's fast everywhere. The reverse is not true.
Delete before you optimise
A plugin that shaves 100ms off its own load time is still a plugin you didn't need. First pass is always 'what can go', second pass is 'how do we make what's left faster'.
Speed decays without maintenance
Themes update, plugins bloat, someone drops a chat widget in. A site that hits green in month one without ongoing attention is yellow by month six. We plan for that from the start.
How we work, start to finish.
- 01Real-user audit
We pull CrUX field data, crawl the site, and profile the top five templates on throttled mobile. You get a one-page ranked fix list — biggest impact first, effort rated — by end of week one.
- 02Quick wins
Images, fonts, scripts, and cache configuration. These are the fixes that usually account for 60–70% of the improvement and can be done without touching the theme.
- 03Template-level work
CSS audit, JS cleanup, layout shifts, INP fixes, and template-specific changes. This is where we touch the theme, in a child theme, version-controlled.
- 04Measure & monitor
Before/after on real user data, not just lab scores. We set up RUM monitoring so the next regression surfaces in a dashboard, not in a customer email six months from now.
Numbers from the last two years.
Averaged across speed engagements where we had before/after CrUX data. Ranges honest, not headline — the ones where we saved 200ms and the business didn't care are in here too.
Before you
ask us.
Find out what's slowing your site.
Send the URL. You get a ranked fix list for the three pages that carry the most traffic — effort, impact, and which fixes we'd do first. No 70-page PDF, no sales pitch attached.
or email hello@genvoid.com