Spam can be defined in various different ways, but for the purposes of a StackExchange site, I have always been of the opinion that a broad definition is generally desirable:
Irrelevant or unsolicited messages sent over the Internet, typically to large numbers of users, for the purposes of advertising, phishing, spreading malware, etc. (ODO)
This definition fits well with the SE Content Policy note on spam users:
Users that do not publish meaningful content, use deceptive means to generate revenue or traffic, or whose primary purpose is affiliate marketing, will be suspended.
But when you flag something as spam on ELU, the explanatory text (and thus, presumably, the definition of what constitutes spam on ELU) is:
Exists only to promote a product or service, does not disclose the author’s affiliation.
– which is quite a different definition.
We get a certain amount of questions and answers that are almost certainly made by spambots and has only more or less irrelevant, auto-guessed, or downright nonsensical content, even if they don’t actually contain any links or try to promote any product or service as such.
A good example is this (now deleted) answer. Yes, the question is about fountain pens, and yes, the answer is also a picture of a fountain pen … but answer visibly consists only of a picture of a fountain and a (nonsensical, ungrammatical) ‘headline’, and the full text in the answer (which, apart from the header, is hidden in the link text box for the image) is:
My Fountain Pen My Favorite One (And my only one)
This is my exquisite fountain pen I saved up for. it is a mess. Why? Well, because it's a fountain pen of course. Fountain pens are awful, but awesome at the same time. My friend tricked me with some disappearing ink on my B-day.
This, to me at least, reads like a very obviously automated answer made by a spambot. The user who posted it had just signed up when this was posted—in fact, as far as I can tell (and I’m not even sure how this is technically possible), the answer was posted the same second the user account was created, and the user has not been seen since.
Consequently, I flagged the answer as spam.
The flag, however, was declined (“a moderator reviewed your flag, but found no evidence to support it”). This is not, if memory serves, the first time I have flagged a similar post as spam and subsequently had the flag declined, which leads me to believe that the moderators are basing their definition of ‘spam’ on the text in the flagging box: if it doesn’t actively try to promote a product/service, then it’s not spam. (Please correct me if this belief is mistaken.)
Presumably, this type of question/answer should more appropriately be flagged as not an answer instead, though criteria for what that category entails are even more fuzzy; the words of one or two moderators on a different question somewhere would lead me to believe that at least some moderators would decline such a flag for the answer in question here, too.
Is it just me, or is that shooting ourselves in the foot somewhat? Would it not make more sense to reserve ‘not an answer’ for real, actual users who post answers that do not try to answer the question; use the ‘spam’ category for anything that appears to be automated, irrelevant, meaningless, nonsense, deceptive, or otherwise not what you’d expect of a real user; and update our definition of spam to be more in line with the general SE Content Policy definition?