2024 February 2
How fast can a modern website get?
It turns out, pretty fast. I recently ran across this great series of five posts by Taylor Hunt from 2022 about a skunkworks project to explore the limits of the possible in the context of the grocery chain Kroger’s website:
- “Making the world’s fastest website, and other mistakes”
- “The weirdly obscure art of Streamed HTML”
- “Speed Needs Design, or: You can’t delight users you’ve annoyed”
- “Routing: I’m not smart enough for a SPA”
- “So what?”
This is the kind of writing that I love, and maybe you will too!
In a few very specific ways, Hunt’s story makes me jealous of the corporate world: in a company like Kroger, you can sometimes prove that certain technical projects — like making the landing page faster — will directly lead to making more money, and sometimes someone actually follows through on the obvious cost-benefit implications. In academia, the “benefits” associated with technical projects are often less tangible, and even when that’s not an issue, the resources often simply aren’t there to follow through on obviously desirable projects. Of course, decisions in the corporate world are often far from rational even at the purely microeconomic level, but it seems like it would be nice to at least have the opportunity to tell someone “this project will make you X dollars, so please give me up to X - 1 dollars to pursue it”.
When that happens at the scale of a large business, you can get stuff like Hunt’s project. Buying specialized hardware (old, slow phones) to test the actual user experience? Check. Isolating crucial system architectural choices needed to achieve demanding UX goals? Check. Design clarity driven by the need to focus on those goals? Check. Big-picture insights springing from a relentless focus on the details? Check. Although I don’t expect to ever need to pursue web performance quite as fanatically as Hunt did, there’s a ton of food for thought in the writeup.
It’s also a little reassuring, in a certain sense, that even in the allegedly hyper-rational world of business Hunt’s successful but unorthodox project didn’t actually get adopted. It can be frustrating in science to see how often people just stick with what they know, rather than what actually works the best, but at least we’re not the only ones with that problem.