I think they interrupted her at home. They were reportedly armed, so that could explain how they got her into the car. I presume they were disguised in some way in case they were seen at the house. As for the phone, Sherri could have been listening to music as she was wrapping gifts or whatever; I prefer to use earbuds when listening on my phone, as th sound quality is better. The abductors muscle her out the door and at some point pull her phone away and toss it.
And I also think she could have identified Latinas without getting a good look at them. I seriously doubt they ever revealed their full faces but even a partial face and/or speaking in Spanish could be clues. Yes, people can fake accents but that strikes me as far-fetched. Why bother if your victim can't see enough of you to make an ID?
I would also imagine they probably blindfolded her early on.