The incoming CEO who wanted to make his mark (let's call him Derek) had an impressive background in marketing decided that the right thing to do was to rationalise everything into one huge database.
They employed a small army of consultants who covered the entire office (it was a big office) in huge database schema diagrams. This took several months. Eventually they "discovered" that the only data that the schemas had in common was a reference, a description, and a created date field.
So they immediately cancelled the project and we all breathed a sigh of relief. Haha, only kidding. Of course they didn't. The right solution was to go schemaless. Enter mongodb. They would achieve full flexibility going forward, and wouldn't have to subside a small diagramatic wallpapering industry as a bonus.
At this point I was no longer involved and just watched the slow train wreck from afar. Years later all the applications had finally been migrated to use the new OneDatabaseToRuleThemAll™ system. I don't think any new functionality was delivered over that period. There were persistent performance issues for the larger data sets. Changing anything required the code to support all previous possible schemas because they never migrated any data (it's schemaless, it's all super flexible!).
I think Derek left to go ruin somewhere else, with a huge success story on his CV.