The Sabermetrics of Campaigning
Capitol Offense by Dave Mann - The Texas Observer
Kuff tips me off to this article, and not surprisingly, I'll second the "beer AND tacos" solution that he offers.
The article in question is a nice overview of one of the Democratic Party's rising stars among the consultantocracy: Leland Beatty. I've worked with Beatty's product before and deal with him on some current endeavors. All that means is that I'll echo the positives of targetting that he offers in this article.
Traditionally, targetting has had the effect of narrowing the pool of voters in play. But that was back when the Democrats doing the targetting were the majority. Targetting from a minority party standpoint, by definition, has to be about how you grow the base.
Now, full disclosure here ... I've also worked on campaigns where Russ Tidwell was involved. Russ is a great guy, but I'm not as sold on his knowledge of the mechanics of campaign work. Tidwell comes across as a sceptic of Beatty's brand of targetting in the article, though I'm not sure how fair that is to his actual views on the practice. Simply stated, there's a huge difference in knowing the macro and micro aspects of campaign work. Russ gets the macro better than he does the micro. Beatty doesn't confuse what he does as anything more than a micro approach that fits within the broader aspects of campaigning: having a good candidate, having a winnable district, etc ....
Anyone can drop mail off at the post office, or blockwalk from a targetted list. But that doesn't close the sale. From my experience with the Miles campaign, blockwalking with the candidate was much more fun than blockwalking without the candidate ... and not just because Borris is a great guy. Voters will judge a campaign based on all the criteria they're presented with. If they get an overworked campaign flack at their doorstep, that's something, but it's not exactly the be-all of campaigning. When they see the actual candidate on their doorstep, different story. If that candidate doesn't measure up, however ... all bets are off. No amount of targetting can make up for that.
Beatty's work is solid, though. It can put good candidates over the top (as it did with Donna Howard and others). Knowing which doors to knock on and which addresses to send mail to is an important part of the process. I'm willing to be that even Big Russ knows that.
Categories
Texas1 Comments
I read this article and liked it a lot. It is understandable why there is resistance to using statistical methods to focus on narrow segments of the electorate. This is understandable, because what these data mining methods tell campaign coordinators is that you can get more votes by focusing your efforts on fewer voters. The same is true in direct marketing, where at least there is a long established track record of success. It just doesn't pay for Lands End to send their catalog out to everybody. But using rich datasets and statistical algorithms like CART, they can find the consumer segments that will provide them the best return on the investment of a catalog mailing. The same idea applies in politics, Karl Rove famously used it in 2004, and Democrats need to get on board--especially as there are people who reflexively call themselves Republican but can be convinced to cross-party lines--if they can be identified and targeted.