Comment by asa400

This was amazingly common in the 2010s during the Big Data craze. I know, because I was the one slapping the bad queries together.

Most startups didn’t care (to a point) because at that point in their lifecycle, the information they needed to get from those queries (and actions they could take based on it, like which customers were likely to convert and worth spending sales time on, etc) was more important than the money spent on the insane redshift clusters.

The mantra was almost always some version of, “just do it now, as fast as possible, and if we’re still alive in a year we’ll optimize then.”