After my star alignment and polar alignment, I park the scope, power off to clear the computer memory, power it on again and select option 2 (don't resume from park) and go through the star alignment and polar alignment routines again. I've had mixed success with it, but sometimes it only takes a couple of iterations to get under 1 arc minute error in each and that's about where I stop. Some nights it takes many more iterations to get there. I'm only using a short focal length (340mm) refractor but that has allowed me to get 5 minute subs. If I stick my C8 on board, I'm lucky to get 1 minute subs. I use an EQ6 but the software is the same.
Like glen says, I believe it is the replacing of the alignment stars - or clearing the memory in my case - that helps hone in to the right place. Also, a few times it has sent me all over the shop at which point I've powered it all off and sat with a pair of binoculars
One other thing...make sure you are using the same stars for your star alignment and polar alignment correction with every iteration. I thought that , in theory, if I was close to good polar alignment it shouldn't matter, but since I've been enforcing that I seem to be finding it faster.
I'm also keen to learn how to get better results without resorting to autoguiding. Personally I find the altitude adjustment on the mount very difficult to operate.