Many online news papers are using disjunct categories for their news channels: an article belongs to either «Politics» or «Local News» or «Business» or «Sports» and so forth. Problems arises when you've got an article dealing with the financial situation of the local football team and the political implications of it. Choose your category ;-)
There are many other examples where disjunct categories are introduced to collect digital things. I do tend to think that this is quite a popular mistake or at least a missed opportunity.
For each and every artificial, disjunct set of categories, anybody is able to derive examples which can not be associated to one single category only. This is a very easy task to do.
tl;dr: The real world does not fit into disjunct categories.
The human desire to come up with a strict hierarchy order results from the interaction with the physical world, with physical objects. Classical example for this are the order of books in a library.
As a result, we do suffer tremendously with our strict hierarchical computer file systems because of bad decisions from the early stages of IT: http://karl-voit.at/tagstore/ or read my PhD thesis on that.
The topic is well investigated and the common conclusion is that strict hierarchies don't work very well for storing arbitrary concepts.
My personal guess is that companies introducing categories for their information portfolio do so because it reflects their internal department organization.
Disjunct categories work for physical items and their digital representation which has to stick to this physical limitation without having all the advantages of a digital alternative.
Whenever you do have a very well-defined and limited area of things, disjunct categories can be used as a common (and forced) agreement.
To be clear: disjunct categories will »work« also for digital things. For example, news web pages will not be totally unusable per se whenever each article is associated to one category only.
However, in many cases, those categories do not provide a benefit for navigation (in what category does the article I am looking for fit?) or search.
Let me give you an example: a store selling electronic hardware does have an area in his physical shop where digital cameras can be found. Every customer knows this and generally does not have an issue with it.
Now consider an online version of an electronic hardware shop. Usually, customers do recognize this: there is the possibility to stumble on digital cameras when they are looking for office stuff, children toys (cameras for kids), outdoor equipment (water resistant ones), tools (ruggedized ones), and so on. Of course, digital cameras do have a category on their own. In case a camera overlaps with, e.g., camping stuff, the very same water-proof camera can be found within the digital camera section and the camping equipment as well.
Nowadays, every customer expects this. Every decent online-shop tries to implement it.
It's subtle and it's the right thing to do if you want to optimize your online web shop. Online shops without these links will work as well. However, you can imagine the difference clearly.
If someone would like to improve any kind of digital representation of things, among other methods I'd suggest to introduce multi-classification, usually named «tagging». And I'd pledge to use a so-called controlled vocabulary. This is a small, well-defined set of tags where authors and readers have a common understanding to avoid singular/plural issues or the vocabulatory problem in general.
Any chunk of information is associated with one or more tags. I can imagine the introduction of one RSS/Atom feed per tag which I also want to implement for my blog system: http://karl-voit.at/tags/
Using tags, information providers could also provide useful search/filter functionality where a reader is able to look for all articles of August 2016 that had the tags «security» and «Android» associated.
So many possibilities with tags.