When we started building BrightSite, we faced a fork in the road. Every other website builder on the market is built on the same playbook: PHP, plugins, and 500ms page loads. We took a different path.
The WordPress trap
WordPress powers 40% of the web. It also causes 90% of the maintenance headaches. The plugin architecture that gave WordPress its flexibility also gave it its fragility — every plugin is a potential security hole, every update is a coin flip, and every page load drags through dozens of database queries.
We saw this firsthand running websites for clients at K Squared Consulting. Every WordPress site needed three things to be production-grade: a maintenance retainer, a security plan, and a backup of the backup.
Why Phoenix LiveView
Phoenix LiveView is an Elixir framework that renders pages on the server and updates them over a persistent WebSocket. The result: page loads under 100ms, no JavaScript build step, and concurrent requests handled by lightweight processes that scale to millions.
For a marketing website — which is what BrightSite is built for — this is the right architecture. Server-side rendering means your pages are fast on day one without you tuning anything. LiveView means the click-to-edit experience works without a heavyweight React app.
The trade-off
The cost: fewer developers know Elixir than know PHP. If you want to fork BrightSite and self-host, the contributor pool is smaller. We accept that trade-off because the platform is managed by us — you don't need to hire an Elixir engineer to use it.
The benefit: BrightSite sites load in 30-80ms. Not 30-80ms in a lab, on a CDN edge in Iowa. 30-80ms from your visitor's browser to a rendered page. That's faster than most CDN edge cache hits on competing platforms.
What this means for your site
Speed is a ranking factor in Google. It's a ranking factor in user attention. It's a ranking factor in conversion. A 1-second delay reduces conversion by 7%. Most website builders deliver pages in 1-3 seconds. BrightSite delivers them in milliseconds.
That's the engineering bet. Three years from now, we'll know if it was the right one.