Virtually anything created by a human being is fallible. It can be a good source, but not always. That depends on what information you are trying to glean from it:
It can (not always) be a good source for:
- The name of the individual
- Their dates of birth and death
- Their relationships (for instance, if buried with a wife or parents)
- Their location of burial
- An image of the grave marker
Other information is typically supplied through the research of the profile manager (much like WikiTree) and as such, it tends to have greater room for the introduction of errors. Still, just because some profiles contain erroneous information does not mean that they should all be tossed aside. It can also be important because those profile managers may have done research from which you have benefited - as such you are ethically obliged to cite that. That's also part of the WikiTree ethic, embedded in our Honor Code:
VII. We give credit. Although most genealogy isn't copyrighted, researchers deserve credit for the work they've done.
Genealogy is also about corroborating sources. It is never ideal to have just one single source to go off of, however good it may seem. It's better to have multiple sources that are mutually affirming. Of course, sometimes not all of the sources agree... but trying to access the truth of what actually happened despite divergent accounts is part of the telling of history and genealogy. (And sometimes all we can do is tell each of the divergent accounts and permit the reader to decide.) So even if you have other, better sources, it is always good to have one more one to corroborate.
Additionally, it is helpful to others to have a cross-reference between one profile and another. That helps to ensure that other profiles do not get misidentified with those representing yours: Think of it as aiding others in disambiguation of similar names. There can be thousands of people named John Murphy. By linking together those profiles which represent the same John Murphy, we help every other John Murphy that is being identified since it reduces the number of unidentified matches. A very similar problem is faced by scientists trying to figure out who wrote a given scientific paper and we are now starting to use a numbering system for scientist-authors to help address the problem:
In scientific publishing, the situation is particularly bewildering in Asia where so many names are similar. According to Nature in 2011, there were 3,926 publications by various authors named Y. Wang, more than ten per day. Papers by Chens, Lees, Zhangs, and Lis were almost as common. So how can we be sure of who is who?
https://www.enago.com/academy/experts-take-on-orcid/
WikiTree's database-lovers (Aleš, I think) have been doing automatic data validation checks, comparing WikiTree info to FindAGrave info. It helps maximize the accuracy and can function as a double-check on an errand keystroke that might introduce a name or date error.
Finally, WikiTree provides a template to help with the cross-referencing of FindAGrave profiles. This reminds us that such a citation can either go under the Sources section or just past the Sources in a section titled, See Also, which exists for other stuff that may relate to the individual profiled.
To summarize:
- It can be a good source for certain information.
- It gives credit to the person whose effort in creating the FindAGrave profile has advanced your genealogy research.
- It can function to corroborate other sources.
- Linking provides a cross-reference between WikiTree and FindAGrave which helps with disambiguation.
- WikiTree's database team can use it to perform a helpful double-check for accuracy, because "We care about accuracy".
I hope that answers your question.