Beta-reading has become a stable and staple part of the indie publishing process, and a good thing, too. Whether you pay for it or not (a few of my peers offer it as a paid service) it is a reasonable and efficient way to get a view on whether a book is "working".
There is a fundamental difference between reading a book that is published, and that you have chosen to pay for and read, and reading a book that someone has asked you to read, or is paying you to read.
Who (of the countless myriad readers of my modest blog) has never picked up a book and begun reading, out of curiosity, boredom, on a recommendation or a whim, only to abandon it after a few pages or even after a few chapters?
When that happens, at least some of the time it can be blamed on the book, and if I want to be really generous, I will say: "that book didn't work for you."
Before you engage the services of a content editor (which ought to be costly - another post is in preparation about that), its good to get some idea of whether the book is working, and the beta reader can do that for you.
So why "unpaid" ?
A volunteer beta-reader can do something that a paid editor cannot do. He can turn round to the author and say "I just couldn't finish it."
To me, the value of this is beyond measure. So often I find that after getting off to a good start, a book starts getting bogged down in details, or the plot loses it's way, or what looked like a promising action thriller becomes a repetitive brassfest. A paid beta reader or a content editor is going to keep going to the end in spite of this. Actually a book doesn't even need to go downhill for a reader to give up. Some books are so patchy in the quality of the narration that eventually the good parts - where the narrative flows well or the author really settles into a groove - no longer outweigh the weak, clunky or "experimental".
An unpaid beta-read should be a "test read". The beta reader should read "as if" reading for pleasure. And you should tell him that if at any time he wants to give up, then he should do so, say so, and try to explain why.