Framer's median site scores 62/100. Here's the gap, and what closes it.
An independent benchmark of 30 real Framer marketing sites found a median mobile Lighthouse Performance score of 62 out of 100 — compared to 94 out of 100 for 30 comparable hand-coded sites, tested under identical hosting conditions. This page runs on the exact standard that gap requires closing.
Sourced & linked below · not a client buildThe standard every build is held to — including this page. Verify this page's own score.
Framer: what actually changes
The default Framer way
- An independent 30-site benchmark found a median mobile Lighthouse Performance score of just 62/100
- Framer ships its entire JS runtime to every page regardless of what that page actually uses — typically 350–500KB of JavaScript on initial load
- Animation and interaction layers load wholesale, not scoped to what's actually on screen
- Framer's own help docs call PageSpeed Insights "not worth dedicating significant effort to" — a defensible stance for a platform whose median site scores 62/100
The ZenMasterWorks way
- System font stack, zero webfont requests, zero render-blocking <link> tags
- No shipped runtime beyond what a given page actually renders — nothing loaded "just in case"
- Every below-the-fold image ships with explicit width/height and native lazy-loading — zero layout shift
- Analytics deferred until 1.5s after the window load event, never blocks first paint
- Native contact form, no third-party bundle loaded speculatively, every field has a real, associated <label>
Where the 62/100 figure comes from
This isn't an estimate from this page — it's a third-party benchmark of 30 real Framer marketing/business sites (10–40 pages, blog, forms, analytics) against 30 comparable Next.js sites, tested under identical hosting conditions. Read the methodology yourself before taking our word for it.
Why the numbers matter
Google and Deloitte's joint "Milliseconds Make Millions" study found that a 0.1-second improvement in load speed lifted retail conversions by roughly 8% and travel conversions by roughly 10%, measured across 37 real brands. Separately, Walmart's engineering team found every 100ms improvement in page speed produced a 2% lift in conversions. These aren't estimates from this page — they're published, third-party findings, linked so you can read the methodology yourself.
Questions about Framer performance
What's the average Lighthouse score for a Framer site?
An independent benchmark of 30 real Framer marketing/business sites found a median mobile Lighthouse Performance score of 62 out of 100, compared to 94 out of 100 for 30 comparable hand-coded Next.js sites tested under identical hosting conditions.
Why do Framer sites score lower than hand-coded sites?
The primary cause cited is JavaScript bundle size. Framer ships its entire runtime to the browser on every page load regardless of what a given page actually uses, typically 350–500KB of JavaScript on initial load — weight a hand-coded, static page simply doesn't carry.
Does Framer say PageSpeed Insights scores matter?
Framer's own help documentation states that PageSpeed Insights is "an imprecise tool that doesn't impact SEO" and that it's "not worth dedicating significant effort to improving its score," pointing instead to Core Web Vitals. That's a defensible technical position — but it's also the position of a platform whose median real-world site scores 62/100.
Can a Framer site be built to score 100/100/100/100?
The platform itself doesn't prevent it, but the runtime it ships by default makes it substantially harder than building static HTML directly. A Framer-equivalent build held to this standard would need the animation and interaction layer stripped down to only what's actually used on a given page, not loaded wholesale on every visit.
Want a Framer-equivalent build held to this standard?
This is the same standard behind every ZenMasterWorks build. Get in touch and we'll show you what a build like this looks like for your business — built first, paid only if satisfied.
Get in touch