Not directly PGM, but hopefully in support of it:
I adopted a lot of old profiles willy-nilly from 1600s New England before figuring out how to tell who was really an ancestor. I am now working through my watchlist, looking for profiles without dates.
I add the dates where I can, and estimate where I can't find them. After as much cleanup as possible, I look to see if there are any potential matches, and propose merges. Often, then I can go back and add dates that have become clear.
If the profile is for someone from New England in the 1600s, I usually add as much in the way of history as possible, and try to make a list of the person's children, to either link now or facilitate links later, if the children don't have profiles yet. Then I do as much connecting as I can to spouses, parents, children.
If the profile meets criteria, I'll add the PGM template if its not already there. After cleanup, if the profile isn't for a likely direct ancestor or their sibling/spouse, I remove myself from the trusted list, adding the unsourced template if needed.
Of course, in the process, I'll find people who aren't connected where they should be, merges that should happen, etc. The latest one I've been working on is Jonathan Boardman, who was probably born in 1660/61. He wasn't linked to his PGM father because of a difference in spelling of their surnames, and his father wasn't listed as PGM. Jonathan's not on my trusted list; I found him while trying to estimate a date for someone else.