What really happened is people wanted something new but did not want to change the ways they DESIGN their application and processes. There is place for NoSQL but if you are going to use it as if it was SQL then you would be better served by an SQL database.
Also, I think MongoDB tried to be everything and failed to be good at anything. It offers neither stellar performance nor scalability and I guess for most projects there is not much advantage over regular SQL database. Certainly nothing to fight over when there is much more technology choices to make.
Just a few thoughts. A lot of little companies like picking hyped tools because they think it will differentiate them and server them well later, but most companies never get to that stage and don't really need what is being offered. Seems much safer to take the tried and true standard solution that has worked for decades and just make your UX outstanding rather than try to put together a "dream team" of new tech...
Another thought. Most companies have absolutely no idea how to select the technology. A lot of people that are in position to decide don't have all that much hands on experience with the new technology and so they will decide based on other factors like whether his manager would like or would like not see new technology, what other companies are doing, etc.
Another example is "Agile". Everybody is doing it yet I still wait to see a single company that understands what the term means. My current boss is big promoter of "Agile" which in his language is synonym to "Scrum". Yet when asked he has never heard of Pheonix Project, The Goal, theory of constrains or basically any theory at all. So what the people are doing is fighting fires almost 100% of the time with not much project work done for the effort and absolutely no improvements. Yet, because everybody complies to do daily standups and Jira updates we are 100% agile.
Also, I think MongoDB tried to be everything and failed to be good at anything. It offers neither stellar performance nor scalability and I guess for most projects there is not much advantage over regular SQL database. Certainly nothing to fight over when there is much more technology choices to make.