Publication Revisions

I just got back reviewer feedback on a conference proceedings article on assessing quality of critical thinking in online discussions (an initial application of an assessment framework which I think will go far so the application isn't extensive yet but the idea is pretty good even if I am saying so myself). It was a double blind process. And the good news is that it is "accepted with revisions." The bad news is that I am confused about making the revisions.

One liked the paper a lot (and gave some great recommendations for clarifying a few points), and the other really hated it (and said that it should only be published if we completely changed the analysis from a general statistical picture of the process, structure, and quality of the overall discussion to descriptions of specific discussion posts).

I think the real problem for the second reader is that I failed at making the analysis accessible to him/her as he/she basically panicked every time there was a statistic. I had those spread through the paper rather than isolated off all in one place that could be skipped by number-phobes. But, I really think that the integration of the results into the overall piece flows nicely; so, I don't really want to separate it out and lose that logical flow. And, my opinion there was validated by the first reader who really thought that part of the paper was presented well.

So, now I feel somewhat stuck as I am supposed to make revisions and am not really sure which revisions to make. Do I need to respond to absolutely everything both said to fulfill the revision requirements? Because I don't think that is possible. If I just hit some of the minor points from the second reader but ignore the huge change, is that acceptable?

Anyway, I just thought I would share my conundrum with others since we are all headed down this publication path. Any suggestions are welcome as is commiseration.