The real problem with divisions of species isn't that they are just guidelines, but that nobody can agree on the rules. Evolutionary biologists have proposed almost 30 different species definitions (called "species concepts") over the years, and none of them apply to all organisms. For example, the "no fertile offspring" definition is a paraphrase of one of the most popular species concepts called the Biological Species Concept (BSC), which actually states that a species is "a population or group of populations reproductively isolated from other such groups". "Reproductively isolated" in this case can mean "no fertile offspring", but it can also mean "they never encounter each other" or "they don't see each other as potential mates". The problems with the BSC are: 1. it doesn't work for organisms that reproduce asexually, 2. lots of organisms we would otherwise call different species have hybrids from time to time, and 3. reproductive isolation is actually really difficult to test (many organisms will hybridize in captivity but almost never in nature). As I said, there's like 30 more definitions of species that have been proposed, and most of them are slight modifications of other species concepts to fix problems with them (and introducing new problems lol). Some researchers have tried to come up with a definition that unifies all these different definitions (e.g. "a species is an independently-evolving metapopulation lineage"), but unified definitions fall apart because they're almost impossible to apply in the real world (how do you determine if something is "independently-evolving" or part of a metapopulation?).