Why don't web pages use UDP for their "tracking pixels"?

A significant number of network calls from web pages are merely for statistics.

The most widely used practice is a GET/POST request with header or URL parameters about the state being tracked.

Can someone explain why this is done with HTTP and not UDP?

• There is nothing actionable if the tracking GET/POST fails.
• Using HTTP means waiting for a useless response. Granted modern browsers reduce the pain somewhat by having a small pool of outstanding requests, but ... why?