I say yes, keep Jacobsdochter, and merge her duplicates as soon as possible, because she is the key profile that seems to be holding up the vast majority of Van Couwenhoven descendants from being cleanly merged away.
Our only other choices are Peters, which is clearly wrong, because that is simply applying the surname of her father.
Or Jansen, which was in a hand-entered profile, but with no parents attached, and no source about that origin. The origin of that Jansen variant may have been the LDS Ancestral file from way back in 1998, which listed her father as Neeltjen Janse. And so that is wrong on two counts:
First, it simply applies the surname of her father as her surname, as in the wrong Peters already mentioned.
And second, the identity of the father as Neeltjen Janse simply seems to have been Ancestral File confusion with this woman, the daughter of Metgen. The Ancestral File named her Neeltje Jansen.
So I would discard the Jansen as unlikely, and as simply an AFN error, although it does leave open the question of where the Janse / Jansen name had come from to begin with.
The father as Jacob Petersz seems to be much more recently widely accepted. So that makes this Neeltje properly the daughter or Jacob.
And since you had opened this way back on March 5, I have no problem with closing it now a day or two early. It has already been nine days since the first query.