All of mine were subject to my whims (including my doctoral one) so whilst I had something of a plan and a flow in mind from the off (necessary, I think, for so thing like philosophy), I didn't write them all, in their totality, in order. I'd switch focus depending on mood, or depending on an idea I had. Then later, I'd find a way to weave things together.
At one point, I stopped writing my PhD thesis for about 2 months to fanny about with translations instead, and at the time I didn't even know if I could get these translations into the thing.
If you're in a discipline which needs data analysis and discussion of results then in my mind it's easier to do the sections in order. But I never have been so it wasn't so thing I cared about.
Conclusion and then intro last, though - that's good advice from Mike.