Speed
AA Sends: Framer Landing Page Speed Optimization
- Client
- AA Sends, an Arabic landing page in Framer. Client project (Upwork).
- Site
- aasends.com/ar ↗
- From → To
- Mobile PageSpeed 35, desktop 45 → Mobile 90, desktop 98
- Stack
- Framer, Framer code components (React, TypeScript), Wistia, Tally, Self-hosted fonts
- Timeline
- 3 days, October 2026
- Key result
- Mobile PageSpeed 35 → 90, desktop 45 → 98

01 / The brief
AA Sends runs a small Arabic landing page in Framer. Visitors arrive from links, not search, so the client wanted one thing: a higher performance score, ideally 90+ on mobile. SEO wasn't part of the brief.
The first test came back at 35 on mobile and 45 on desktop.
02 / What was slowing it down
- Three heavy third-party embeds all loaded at once on page open: a Wistia video in the hero, a Tally form, and four YouTube videos at the bottom.
- Wistia and Tally load from their own servers, so their speed can't be improved directly, only when they load.
- The Desktop, Tablet and Mobile layouts were built as three separate copies instead of one shared component, so phones were downloading the same content twice.
- Several Google Fonts loaded straight from Google, blocking the network at the start.
03 / What I did
- Kept the original page as a backup and worked on a copy.
- Rebuilt the duplicated sections as one component across all three breakpoints. The score improved immediately.
- Added code overrides so the Tally form and YouTube videos load only when the visitor scrolls near them.
- Built a custom Framer code component for the Wistia video. It takes the video link, shows a matching poster right away, and loads the real player only after the visitor's first scroll, touch or mouse move. The player then plays a muted loop, and a click plays it with sound. Visitors see no difference, but Google's test does.
- Downloaded the fonts and self-hosted them as custom fonts, keeping only the weights the page uses.
04 / The result
| Before | After | |
|---|---|---|
| Mobile performance | 35 | 90 |
| Desktop performance | 45 | 98 |
| Accessibility (mobile / desktop) | 89 / 89 | 91 / 93 |
| SEO | 61 | 100 |
| Best Practices | 100 | 100 |
| Agentic Browsing | 1/2 | 2/2 |
Google PageSpeed Insights, same page before and after. SEO wasn't in scope, but it moved too.
05 / Where Framer stops
Framer adds its own JavaScript to the head of every site, and it can't be removed or deferred. That script is what holds the first paint and the largest content back now. Past this point, more speed means moving the page off Framer to a code-first build like Astro, which is what I did for my Everline Advisory rebuild (mobile 78 → 99).
06 / Client feedback
What I'd do for you
Is your Framer page scoring in the red? Send me the link and I'll tell you what's slowing it down.
Send me your site

