I like to see success from people who grit their teeth and delve into these forms of depiction. Much of digital photography, and the effects and manipulation we do, is rather straightforward. On the other hand, some of it is darned tricky, as if we have to outwit our own 'help.'
This was the case with CombineZP, which I urge everyone to tackle at least once. I can talk you through it if you're interested. But, on topic, I found that CZP, and CZM before it, do much better, and have less of a perimeter of 'waste', if you take pains to orient the lens and detector as close to the same from shutter release-to-shutter release as possible. The more you fudge your aim, and the more grossly 'discrete' your focus shift from image-to-image, the worse the result will be.
What this means is that I got used to shooting strictly in manual so that I could control the depth of focus of each image, but that I also took pains to ensure the camera was restored to a jig of sorts before the timer tripped the shutter (oh yeah, you have to use a timer). So, I would go into macro mode, put the camera in a cradle of some kind, half press the shutter release to get the camera to focus, and then start the timer. I'd take up the camera, switch out of macro as appropriate, and then repeat, always using the timer, always restoring the counting-down camera to its cradle before the shutter released. This way, CZP would have minimal 'crud' to crop as it placed erroneous digital information at the edges of the photo, and you'd have a wider field of view to work with when the final focused rendering is finally released by the software, which in the case of CZP can take up to 2 minutes if you have a lot of images to integrate.
Keep it going, boys...