There's no shortage of campaign strategy advice in this year's Kentucky Senate race. In that spirit, I'll add my own two cents.
If I were advising the McConnell campaign, I would say...
Don't screw up.
You're the incumbent and incumbents already enjoy somewhere between a 5%-10% advantage right off the bat, although this does tend to fade over time so it's not going to be worth as much as it was in the past when you first ran for reelection. Also, the economic and political "fundamentals" are on your side, which is why all the numbers geeks are giving you anywhere between a 78% and 99%+ chance of winning (see here, here, and here). So basically, just make sure to keep up with the fundraising and campaigning, give your conservative Republican base a reason to turn out to vote for you by railing on Obama and by talking up the strong possibility of a GOP Senate takeover.
And don't screw up.
If I were advising the Grimes campaign, I would say...
You and I both know that you have an uphill battle to fight. You're a Democratic challenger in a red state where the sitting Democratic president is very unpopular. But then, your incumbent opponent is also very unpopular in your state, but that tends to matter less than the economic and political fundamentals which are currently giving you a 1-in-5 chance, at best. You'll need a strong campaign combined with some luck to come out on top this year.
Right now it seems that one of your key strategies is trying to appeal to women, presumably in an attempt to entice Republican women over to your team (see here, here, and here, e.g.). While it makes for a great media narrative and may possibly work, there are strong reasons to think that this may not be the most effective strategy. To put it bluntly, women simply don't tend to be swing voters. Oodles of political science research has shown that, after controlling for partisanship, there's not much of a difference between men and women in their voting patterns. In other words, women are just as reliably partisan as men. The fault lines of American politics do not tend to fall around gender, but rather partisanship and ideology. Thus, there are likely not very many Republican women who are going to "defect" in this high-profile partisan election.
So who are more likely targets where you could concentrate your efforts? I took the liberty of doing some number crunching on an exit poll of Kentucky voters from the 2008 Kentucky Senate election where McConnell narrowly beat Bruce Lunsford 53%-47%. In that election, only about 14% of Republicans voted for Lunsford, and they made up only 5% of all voters total. Further analysis shows that these Republican defectors tended to be a little younger than their loyal partisan counterparts (about 22% of Republican defectors were under age 30 compared to 15% of Republicans who stayed in the fold). They also tended to be poorer (46% of Republican defectors made less than $50K/year compared to 33% of loyal Republicans) and more ideologically moderate (56% of those Republican defectors identified as moderate and 34% as conservative, while those who stuck with McConnell were 37% moderate and 70% conservative).
Perhaps most importantly, there was ZERO difference when it came to gender. 50.7% of Republicans who voted for Lunsford were women compared to 50.4% who voted for McConnell - a statistically indistinguishable amount. This suggests that women are very likely not the persuadable demographic among Republican partisans. Instead, it seems to be younger, poorer, more moderate Republicans.
On the other hand, nearly a quarter of self-identified Democrats switched sides and voted for Mitch McConnell in 2008. They made up a full 11% of all voters in that election. What did these Democrats look like? They were more ideologically conservative (34% of Democratic McConnell voters said they were conservative compared to only 15% of Democratic Lunsford voters), more likely to be white (95% of Democratic defectors were white compared to 72% of loyal Democrats), and more likely to approve of George Bush (34% compared to 10%). They were also slightly more likely to be men, making up 48% of Democrats who voted for McConnell compared to 41% of Democrats who voted for Lunsford. There were also no differences when it came to age, education levels, income, or religiosity. This suggests that in 2008, Lunsford lost Democratic partisans who looked a lot like Republicans - conservative white men who were more approving of President Bush. This suggests that you might have success keeping your Democratic partisans "in the fold" by veering toward the middle and appealing to cultural conservatives in Kentucky as much as possible.
That presents a tough choice: appeal to younger, more moderate Republicans who might be persuaded to defect or appeal to conservative white Democrats who may be likely to switch sides. Given that there were more than twice as many voters in the latter category (11% of all voters) than the former (5% of all voters) in 2008, it stands to reason that veering toward the middle and trying to retain moderate Democratic partisans may be the option with the higher pay-off. That being said, you don't want to veer too far toward the middle or you might risk alienating your loyal liberal base so much that they don't care enough to turn out to vote on Election Day. Trying to balance that tightrope walk will be a delicate endeavor indeed.
One thing is for certain, at least: there is little evidence from the 2008 Kentucky Senate election that Republican women were a persuadable demographic in that campaign. It's possible that the 2014 Senate campaign will be different, but given how consistent and predictable American voting patterns are, I wouldn't bet on it. Perhaps consider altering the approach slightly. Forget about "peeling off" Republican women and instead focus on loyal Democratic women (to make sure they show up to vote on Election Day) and moderate or conservative-leaning Democratic women (to encourage them to stay in the fold).