The Wrong Tool Is Expensive Even When It's Free

The deployment for a new product was built on Cloudflare Workers.

Workers are designed for dynamic edge logic. They require explicit routing code to serve static assets. The correct tool for a static HTML site is Cloudflare Pages, which handles routing natively with no configuration.

The site was built. The pages broke. The debugging cycle ran for hours across multiple sessions. The root problem was not in the code. It was in the choice of infrastructure made before the code was written.

Cloudflare Workers and Cloudflare Pages are different products with different mental models. Both are free at low volume. Choosing between them costs nothing. Choosing wrong, then discovering it at deployment, costs time. And time in an early-stage company is the one resource that doesn't replenish.

The lesson is not about which tool to use. The lesson is that free options are not equivalent options, and the cost of the wrong choice is almost never discovered at the moment of choosing.

Architecture decisions made under time pressure tend to get made by default: use what you know, use what's familiar, use what's easiest to set up. The cost appears later, when something breaks in a way that takes hours to trace back to a decision made in thirty seconds.

The fix was deleting the project and starting a Pages deployment. Twenty minutes. The discovery cost was the expensive part.

Make the infrastructure decision deliberately, before the build starts. The tool selection conversation is not overhead. It is the build.