Are there advantages to having multiple databases rather than keeping everything in a single one, such as the Global Inbox? Yes, indeed!
All together, I manage more than 250,000 documents among a number of DEVONthink Pro Office databases. Some of those databases are archives of old information that I hardly ever need to review. Even on my current MacBook Pro Retina with 16 GB RAM, the performance of a single database holding everything would be terrible, and I wouldn’t be satisfied.
I create topical databases, each holding information that meets a particular interest or need. I treat them like information Lego blocks that can be opened or closed as needed.
Normally, in addition to the Global Inbox, I have a set of 5 open databases, that have a total word count of about 40 million words (comparable to the word count of Encyclopedia Britannica). That’s a LOT of information, and as there’s plenty of free RAM left on my MacBook Pro (currently about 6 GB free RAM, after meeting the memory needs of my databases and several other open applications), my computer always runs at full speed, with no pageouts and no use of Virtual Memory swap files.
The full Search window (which I use for all my searches in DEVONthink) lets me search across all open databases, or only within a specific database, or even with a specific group in a database. And as the option to provide indexes of my databases to Spotlight is checked for all databases, I can do Spotlight searches across all my databases, whether open or closed. (Tip: After a Spotlight query, choose the option to view all results. They will be displayed in a Finder window, and all results in DEVONthink databases will display the blue ammonite shell icon. Click on such a result and press the Space bar to view the content in Quick Look, or double-click on the result to open it in its DEVONthink database.)
So the first reason to use multiple databases is to tailor the size of open databases to the computer’s resources, keeping performance at the maximum speed of which the computer is capable. That becomes very important as one’s collection of documents grows.
But there are other reasons to use multiple databases. One that is especially important to me is that topical database design can make searches and the artificial intelligence assistants such as See Also much more efficient and useful.
For example, I have two databases that deal with my professional interests in environmental sciences, policy and management. My main database contains about 25,000 references and about 5,000 of my own notes, and includes scientific papers, case histories of environmental problems, discussions of environmental policy issues and environmental laws and regulations (mostly U.S. and EU).
I have a companion database, also a large one, that holds information dealing with methodological matters, such as environmental sampling procedures, chemical analytical procedures, quality assurance protocols, procedures for evaluating environmental data, and risk assessment and cost/benefit methods.
Suppose I’m researching the health issues posed by mercury pollution in edible fish. I search in my main database and will pull information about mercury toxicology, case histories of health problems, regulatory standards for mercury intake, etc. But I wouldn’t want those searches to be diluted with large numbers of items dealing with how to sample for mercury, how to analyze for mercury in fish, how to develop regulatory standards, and so on. Likewise, I use the AI assistants such as Classify and See Also. Splitting those environmental databases makes those assistants much more efficient and useful to me.
I rarely send new content to the Global Inbox. Most of the time, my setting in Preferences > Import—Destination is Select group. That allows me to choose the appropriate open database and even a specific group within that database, in which to file a new item.
I don’t spend a lot of time in filing new content. My hierarchical structures of groups in my databases usually don’t go very deep, but emphasize main topical groupings. I’ll confess that I’m cynical about the return on investment of my time and effort in applying tags to documents. I almost never tag anything as it is added to a database. I’ve been working with information management since the 1960s, back when the only way to find anything added to a computer collection of items was by searching for keywords/tags, before full text indexing became available. I became all too familiar back then with two fundamental problems of applying keywords/tags: comprehensiveness (were all the important elements of information identified?) and consistency (different people often used different terms, and even the same individual was likely to do so at different times).
I fell in love with DEVONthink precisely because it provides tools to help me find the information I need, without forcing me to spend a lot of time and effort in organization or tagging of content. 
Do I use tags? Yes, usually at the stage in which I’m working on a project and want to identify resources for that purpose. I’ll apply tags to remind me of those resources and help classify them. Often, when the project is finished, I’ll delete those tags. Why? Because the next project may use some of the same documents, but in a different context, so that older tags may be counterproductive.
) so you already do data segregation in your life. You don’t want the dishes in the bathtub and you don’t want to bathe in your sink. I would think long and hard before I would just accept the single room approach.