"Don’t worry, your DNA doesn’t change. What changes is what we know about DNA, the amount of data we have, and the ways we can analyze it. When that leads to new discoveries, we update your results."
In June my initial DNA findings were 70% England, Wales and Northwestern Europe. 25% Germanic Europe, 3% Norway, 2% Sweden.
In August, they revised it to: 59% England, Wales and Northwestern Europe. 35% Germanic Europe, 2% Sweden, 2% Ireland and Scotland, and 2% Norway.
I'm not concerned about the update in the DNA breakdown since I always only thought I was Dutch (my father was born in Holland), and English because my mother was born in Canada, and early research I did showed part of her family came from there. The other parts, Germany, Norway, and Sweden were a surprise. I am wondering though about the Italian ancestors I have found in my family tree. My 9th great-grandfather Peter Caesar Alberti was from Italy. I've rechecked the earlier connections, and the records all match. He was born in Venice in 1608, and is hailed as the first Italian immigrant to New York. His mother was supposed to be Veronica de Medici. His father Andrea Piero Alberti. Peter and his wife Judeth Jan Menjie were killed by Indians on his plantation in Brooklyn near New Amsterdam on November 9, 1655. As far as I know there is no proof that his father was Andrea Piero Alberti, or that his mother was a de Medici. Those are the "potential" parents cited in hints from Ancestry.com and other people's family trees. The problem is that there hasn't been any DNA connected with Italy. Most of my DNA connected lines peter out in the mid-to-late 1700's, and since he was born in 1608, the DNA connection on that family line ends at my 5th great-grandmother who was supposed to be his direct descendant. I know that Native American blood lines, even if they are supposed to be there, according to Ancestry.com may not show up because either it was so little that it didn't register, or that my ancestor never inherited the Native-American DNA from their predecessors.
Once the DNA connections end in the family tree, anything beyond that is speculation at most. So like you, I take all these connections further up the tree with a grain of salt. They can't be proved.
On FamilyTreeDNA I found a couple of third cousins (same set of great-great-grandparents). One of them was one of my closest matches; the other was way down the list sharing a much smaller number of centiMorgans with me than the other man does.
I did the Ancestry test in 2016 and have had two revisions since then--in October 2018 and a few weeks ago. The most recent one seems the most accurate (which is not to say I am 100% convinced by it).
Sometimes the matches are useful--I have had some interesting exchanges with distant cousins I would not have known about otherwise. I was able to help a fourth cousin in Australia, adopted at birth, to find her biological mother and they had a happy reunion. Of course many of the matches are too distant to be able to find a paper trail documenting the exact relationship, and sometimes I wonder how I could possibly have a relative in that country.