Karl Rove may suffer from a couple of common problems. First, he seems focused on demographic data that describes WHO voted for whom but not WHY they voted. The latter type of data is more difficult to obtain but it is much more valuable for creating strategy. You can get absorbed in data even though that data is not especially useful.
The second problem is failing to notice that the opposition party runs candidates that are absolutely nuts and it doesn’t seem to slow those candidates one bit. Rove’s concern that some candidates need to be eliminated before the general election overlooks the possibility that the criteria he is applying to determine if a candidate is electable may not be important factors in their electability.
A third problem Rove may have is confusing the power of donations with election success. Rove’s party had lots of money in the 2012 campaign and did not spend it as effectively as they could. In the end, you need votes. A few big donors may not translate into sufficient votes. It would be nice to have millions of little donors—more voters. Getting out the vote is important. Getting donations has to be linked to getting votes.
The result is that Rove is attacking conservatives and losing elections. I suggest that running a candidate that doesn’t match your demographic trend predictions may result in continuing losses.