One facet of web sites that often confuses clients who have not been involved with the web before is that they have to pay separately for their domain and their hosting. While the different bills can be annoying and seem like a scam, there is very good reason to keep them distinct and one should actually be wary of any "package deals" that sell both at once.
The difference is often explained as if the domain is a phone number and the hosting server an actual phone, though I think this analogy is a bit limited as we will see in a moment. As long as these can be purchased separately (and the number simply directed to the phone by your telecom company), it's possible to get a new phone every couple of years without having to alert all of your contacts that your number has changed. You contacts don't need to know or care that the hardware you carry in your pocket has changed.
The same is true of hosting: as a company or organization grows, the needs for it's web hosting will change. Perhaps it simply gets more traffic and needs a dedicated server, or the backend tech running it's website will change. Or maybe the hosting company just sucks (let's not go there). In all of these cases, it's possible to keep the same domain and just direct web traffic to a different physical computer.
One would NOT want to change their domain just to switch servers. Unlike a phone number, a good domain is a plain-language pneumatic for an organization, as readable by humans as it is by computers. You don't need to look up anything to contact Facebook. A good domain that has built traffic, public knowledge, and search-engine credence over time has equity that can be measured in real monetary value.
A modified version of this exists for scientific journals, who - in a time where anyone can setup a website that they call a journal and start "publishing" papers - must protect their reputation. John Bohannon, who previously exposed a new underclass of predatory publishers, reports that in several cases, the domain of a journal was "snatched" by predatory entities who then use this domain to fool researchers into publishing in their fake, spoof journal.
I am utterly astonished by this. The only way to snatch a domain is if the original owner allows it expire. This is one thing for a small business like a hair salon or car mechanic, another one entirely for an academic publisher who's sole "product" is information and therefore whose business in 2015 is web-first, print-second, even if they don't acknowledge it.
While Bohannon implies that these journals have not been doing due diligence in web administration and security, this doesn't begin to describe it. The domain now IS the equity and the reputation of the organization - the first identifier a human or a computer associate to an actual entity. to ignore it is not a lack of diligence, it's total ignorance of how people - even academics - interact with journals in the 21st century.
Perhaps this shouldn't be a surprise for an industry that still seems stuck in the print world, where PDFs are still the primary data format for both manuscripts and publications, that charge by the page and impose strict length limits even for methods and references, that distribute "raw data" in MS Word tables, that demand figures at 600 dpi (as if dpi mattered for 90% of their readers), and that demand archane reference formats. The Science news piece reporting this story inexplicably has two sets of comments: one for mobile and one for desktop.
Perhaps the obsession with print, which has meant death in other industries, has actually helped journals retain their exceptional profit margins when they aren't adding any conceivable value to the scientific process. As long as space remains limited in a particular journal, space in that journal will be perceived to have merit. And nothing matters more to academics than merit. God forbid they have the technology to publish an effectively limitless amount of content every month: written, edited and peer-reviewed by people who don't even work for them.
Contrary to Bohannon's claims, buying a domain when the owner let it expire is not "hijacking." While Bohannon says "Anyone can buy a Web domain from private registration companies who neither vet nor care whether the purchaser has a 'right' to it." This is, of course, by design; no one has a "right" to any domain a priori. Just like no one has a right be a gatekeeper of scientific truth, a priori.