This is my attempt to explain how the Ancestral Regions work (and obviously is my interpretation of Ancestry's explanation).
The first thing to understand is that there is no such thing as "German" DNA, any more than there is English, Scottish or Irish DNA. All there is, is DNA which is more or less prevalent today in people with Scottish, Irish, German or English descent.
The second thing is to consider what Ancestry actually does - which may not be quite what the marketing blurb sometimes suggests.
What Ancestry does, is form reference panels of, in their words "people with deep family roots in a specific geographic area or cultural group". What that means is they use people whose families can be shown to have lived in a particular region for several generations. So their German reference panel (which at 2000 members is the same size as the Scottish and Irish reference panels) consists of people who have established roots living in Germany for several generations.
They then analyse the DNA of the people in the reference panel, measure the percentages of people who have each of 1001 separate segments, and use these percentages to decide on the "origin" of any particular segment.
Let us call one of these segments the "Abba" segment. It turns out the Abba segment is quite common in the Scottish reference group - about 40% say. It is also quite common in Germany (let's say 38%) and Norway (62%), but as I'm sure you've guessed already, I'm going to make this particular segment most common in Sweden (75%, say). (Just in case it's isn't obvious, I'm making these numbers up, I have no idea what typical ranges they get.) The point is, now that they have have measured the percentage abundance of the said segment, the "Abba segment is now categorised as "Swedish".
Now they analyse my DNA. It turns out I have the said Abba segment, so Ancestry now allocates 1 of my 1001 segments as Swedish. This does not mean I must have gained that segment from a long lost Swedish ancestor - it just means I have a particular segment of DNA which is more common NOW in people whose families lived in Sweden for several generations in the past than in any of their other reference groups.
Should Ancestry find 10 such "Swedish" segments, that is 10/1001, ie roughly 1% so it now classes me as 1% Swedish. BUT, because each of those MAY in fact be from Scottish ancestors given that, as with my earlier example of my entirely fictitious Abba segment, those segments may also be relatively common here in Scotland, they put an error range of between 0 and 1% on my Swedish ancestry.
Given that the UK has during it's history been invaded by Romans, Angles, Saxons, Jutes, Vikings from various parts of Scandinavia as well as Norman French, has provided refuge to many exiles from religious or other forms of prosecution, and has in turn fought over or invaded large parts of the globe, it is perhaps not surprising that we share a lot of those segments with many other groups.
What all that means is that what used to be called ethnicity estimates are just that - glorified guesses based on statistics and probability - which only indicate that you share DNA with people whose ancestors are from a particular region in the past. If you want the fuller picture look at the ranges on the given identified regions.
Most importantly, always remember that DNA is only one more piece of evidence that might link you to a given ancestor.
