I am actually on the direct consumer side, but I feel for you. Occasionally we'll take intermediate data though from third parties, and good god the quality is atrocious.

This is especially true for datasets that don't have a single authoritative source, like corporate calendar actions, where you have to consume data from 3-4 vendors and reconcile for a 'shared consensus' on simple shit like..."When is a company going to release quarterly earnings?"

I'm ever grateful that I do not work in asset classes that are not centrally traded and centrally cleared. I recognize there's more money to be made in the uncertainty, but holy crap it drives me insane.