Yesterday I admitted distribution is my weak spot. So before asking anyone to visit my landing page, I checked what they'd actually get when they arrive.
The honest answer: a page that was heavier than it needed to be, especially on phones.
So last night turned into a speed session. Here's what changed.
1. Right-sized images
The three illustrations were 280 KB on phones. Properly sized WebP versions brought that down to about 50 KB. If you serve one big image to every screen, phones pay for pixels they never show.
2. No render-blocking stylesheet
The stylesheet was the only render-blocking request on the page, so I inlined it. The browser can paint without waiting on a second round trip.
3. Self-hosted fonts and long caching
Fonts now come from our own domain, and the CSS, JS and fonts are cached for a year. Fewer third-party connections, and repeat visits barely touch the network.
4. Lazy-loaded diagrams
Four diagrams (45 KB of SVG) now load when you scroll to them instead of in the first response. The first screen only carries what the first screen needs.
5. Contrast fixes
I fixed some low-contrast text I'd been ignoring for a while. Not a speed fix, but it's the same idea: don't make people work to read the page.
Why bother
None of this is exciting. Nobody is going to clap for a font table. But if I'm going to work hard to get someone to click, the least I can do is not make them wait.
Earlier the same day I shipped v0.1.1 of Virling, the open source revenue analytics tool I'm building, with a one-command installer and a docs site. I also started tracking where visitors leave a page, so soon I'll be able to see whether this speed work actually changes anything.
My takeaway: distribution doesn't start when you hit post. It starts with what people find when they land.
What's one boring fix that made a real difference on something you shipped?
Top comments (0)