I think the question of the very low conversion rate of editathon editors into regular editors is a well-studied problem.
I'm fairly relaxed about it.
Anecdotally I know some people I've trained
have returned to editing after a long period (a year or more), or have
continued editing occasionally, as IP editors. Others may not have
edited again, but act as advocates, telling other people that they can
edit (and this includes librarians and archivists I've trained, who speak to potential new editors every working day). Even if all we achieve is to get someone to stop decrying Wikipedia as unreliable, or to donate in our yearly
drives, that's something.
One of my trainees, for example, wrote
<https://en.wikipedia.org/wiki/Lisel_Haas> and never edited again (at least, not from that account).
Wikipedia is still better for having that article, and I'm glad that
the effort I invested resulted in it.