There are a lot of different ways to approach this, but for me, I find the ideal number is six to eight. Unlike the other posters, I prefer simultaneous reads. (Note: unlike E.Murray, I do include some family members and friends in this group: the ones who are excellent writers themselves, who read in my genre, and who I know will be brutally honest with me.)
Why?
Well, first off, you'll get a lot of different opinions thrown at you. Some of them will be easy to make a decision on: you'll see someone's comment and think, "OH! Why didn't I see that?!" and immediately scuttle off to fix it. But for other things, you may think, "Hmm, Billy Joe Bob says my pacing on this chapter is slow. But I don't think it is. Is that my author bias? Or BJB's 3.5-second attention span?" Or maybe Delilah tells you she got totally confused about what was happening in the car chase scene. Is it unclear? Or does D always struggle with visualizing city street layouts?
If I have six readers and one of them gets confused on a point, but it seems very clear to me and to all my other readers, I might ignore that note. If three or four of them are confused, it doesn't matter how clear I thought I was being; that section is getting rewritten.
I also find that multiple readers seeing the same draft helps me to pinpoint the most glaring problems in my manuscript. Which is useful, because if I'm going to make major changes, I want to make sure to hit the most important issues first, before I tidy up the minor ones.
Also, as Neil Gaiman said, "Remember: when people tell you something’s wrong or doesn’t work for them, they are almost always right. When they tell you exactly what they think is wrong and how to fix it, they are almost always wrong." For me, seeing feedback from different people on the same draft is often a good reminder of this. Three people may tell me a chapter's not working for three different reasons, and two of them may propose very different solutions. Seeing a variety of takes on a single problem can be helpful to me; I'm pretty sure there's been at least one case where the real problem/solution wasn't something anyone articulated--but because I had a variety of comments around one section, I was able to take a closer look and ask myself, "Okay, so why is this causing different problems for different people?" And work out the solution from there.
If I've made a lot of changes to a manuscript after beta feedback, I may run it by a second round of readers--actually, I'm in the process of doing that now--just to make sure I didn't create new problems when solving old ones. But that's a much less intense process, and the revisions from that one (at least in this case) are relatively minor.