Lessons from My Mistakes: Start with MySQL

+1 to premise of the title, but these specific reasons don't hold much water.

* "no tooling": searching Google for "mongodb ui" turns up - even before the organic results - three PAID results for different tools. I'd suspect you'd see similar results for most any reasonably popular NoSQL option.

* table scans: table scans can be expensive in both approaches. You'll get exactly the same performance problems doing the queries described against a SQL database; the number of records required to be considered "a lot" may be different. The fix is similar in both: indexing.