If it has been spoken, it is a word.
Like, a couple seconds ago I vocalized the sounds “xzoubadoo”, but I didn’t attach any particular meaning to it, it was just an arbitrary collection of syllables, “spoken” to demonstrate a point, but without the intent that the sounds stand for anything.
I don’t think “xzoubadoo” is really a word. At least, not yet. I suppose one might (likely temporarily/in an ad-hoc way) assign it a meaning as a response to this comment, but I have not done so.