Heel contours--structural considerations

I decided that I wanted to individualize my guitars in several ways, and one of the areas I changed was the heel of the neck. I went to a more rounded shape, with less material at the base of the neck shaft, but larger in section [similar to Mike Doolin does on his necks].

The last 3 or 4 guitars that I used this shape on, seem to exibit an excessive action height increase after being strung up and played for several weeks. I have had to re-set the necks on 3 of them, which also required shiming of the fingerboard tongue over the soundboard to prevent a downward pitch at the 14th fret. The action itself has been restored by doing these resets, and they have needed just a tweek of the truss rod to fine tune it.
Could the lack of material under the neck shaft, where the heel block meets the shaft, be the culprit? I know that a new instrument needs time to 'settle' under tension, and the action height might rise a bit, but these occasions have been excessive; triming the saddle height would have resulted in a total loss of the saddle over the bridge!

===========================
I don't know much about Mike Doolin, but a visit to his site shows fairly ordinary heel shapes, at least as far as the pictures there show. They look a lot like old Guild guitars and like my old partner, Michael Millard of Froggy Bottom Guitars, makes them. There are countless thousand of successful guitars made with that wide and narrow kind of heel shape. That would lead me to dismiss the notion that your heel size, mass or contour is the culprit.

Without knowing what the rest of your internal structure looks like, it would be hard to diagnose your problem. But I believe that the precise heel contour is of little importance to the rate at which the guitar distorts under tension. That is because the largest force vector of the neck acting on the body points straight into the body, just under the fingerboard. The remaining, smaller force vector is rotational--the force pulling the tip of the heel out of the guitar. So heels can be surprisingly small, even (as in some low-end Taylors) nonexistent. If my considered perception is true, then the major concentration of forces is dumped onto the soundboard, just under the fingerboard tab, in the region between the front of the box and the soundhole. That area is receiving massive compression from equal forces originating from both ends of the stretched string--from the nut and from the bridge. Those forces meet in that location. if that area yields, the box will bend, causing the string action to rise. So to counter those forces, I would pay attention to whether your structure is adequate under the fingerboard tab (including, of course, the stiffness of the fingerboard tab itself). Over-structuring that area, unfortunately, is a common ham-fisted response of many builders who recognize this problem. So the trick is finding the "minimum adequate structure"--not placing a box girder there, as some seem to want to do.

Keep in mind that I may be completely off base here, because I don't have your guitar in front of me for a valid evaluation. Your guitar design's propensity for neck rotation can be due to any of a number of other shortcomings, such as inadequate soundhole bracing (is your soundhole beginning to look like a potato chip?) or even an insufficiently arched back. Yes, the back's rigidity has an important role in the soundbox's over integrity. The entire soundbox can bend longitudinally under stress, and its rigidity is in part determined by the rigidity of the back. That rigidity is imparted by the size and arch of the back braces.

So any successful diagnosis of the architectural efficacy of your guitar design has to consider the entire guitar "holistically." It is short-sighted to assume that all your problems stem from just your heel profile.