I would go ahead and assume this donut chart wouldn't be the only one displayed, but will be part of a row(s) full of charts and visual graphs.
In this case, I would absolutely organize the text in a readable data structure - Label top, Data bottom.
When visualizing data as such, it is better practice to first give them the key for the data(which normally takes more time to read and understand), and later the numerical value.
That's in order to avoid the situation when the user looking first at the number ("I have no idea what this number means"), then looking at the label ("aha! that's what it means, but what are the figures, again?"), and go back to read the numeric value because he forgot what he did there from the first place.
As the human eye reads (in English) from left to right, and from top to bottom (just like computers, by the way), it would make much more sense to structure data and forms like this:
Contacts Left: 6,689
6,689: Contacts Left, especially when you have more than one figure displayed and the user have to read big amount raw information.
More about label-form placement(not exactly your case, but the conclusions might be somehow relevant):