Project Ivy: Much Ado About Nothing

The Russian invasion of Ukraine has occupied much of the discussion over the past week. It is easy to forget some developments that were considered newsworthy just a short while ago — especially when they are stories that pertain to domestic politics exclusively.

From the last few election cycles, one thing has become clear. The ability of President Obama to turn out his base during both of his election victories is by far the most important factor in his two wins — especially his second one. Many think that the main reason he had been so effective in this regard was his campaigns’ use of data mining and microtargeting.

Obama’s success in the GOTV (Get Out the Vote) effort was quite extraordinary in his last campaign. So much so that in the weeks leading up to the 2012 election, Republicans were dismissing polls that showed him winning as “skewed,” because they could not accept the fact that Democrats would turn out in stronger numbers than Republicans. And turn out they did, handing the president his reelection victory.

As if it were not enough that the Obama team had invested a lot in this area, the Republican disadvantage was compounded by the fact that the Romney campaign’s effort to compete on this front was an epic failure. Their attempt to outdo the Obama campaign in the tech arena ended up actually costing them votes as an unreliable and often wrong “Project Orca” led them to the misallocation of resources on Election Day.

So the RNC decided to address this disadvantage by creating an in-house tech incubator called Para Bellum Labs, where they could make use of the data they had already spent years compiling. When consultants hired by the presidential campaigns (as was the case in 2012) used that data in an inefficient way, it neutralized whatever advantage the information would have been able to give them. It also meant that campaigns, which had to spend incredible amounts of money building the platform through which they would use the data, stood a chance of wasting all those financial resources as well.

This new effort to bring the entire operation in house pays off in two distinct ways for the GOP as a whole. Having a full-time staff working on this project helps refine the tech they will be working with and makes the probability of a glitch-filled launch on Election Day (as was the case with Orca) less likely. The fact that this is an ongoing project also means that every single campaign that the party gets involved in will have access to this resource — something that was usually cost-prohibitive for all but a national campaign.

In many ways this groundbreaking step for the RNC is their way of advancing after having been “out-teched” in the last two presidential elections. During the Bush years, the Republicans had a great advantage in both voter roll data (the expertise of Karl Rove) and microtargeting (the field of Ken Mehlman). As is often the case, winning leads to complacency and stifles innovation — which was responsible for the improvements in this field by the Obama campaign being able to leapfrog the Republican candidates, who had grown accustomed to relying on what they thought was a microtargeting machine that was the most efficient. The hope was that the same sense of contentment with where they stand in relation to each other would have led the Democrats to suffer the same fate.

But last week the Democrats unveiled “Project Ivy” — their answer to Para Bellum Labs. The two projects are very similar in scope and nature. The difference between the two, and why it would seem to be a matter of concern for Republicans, is that it is apparently being built on top of the Obama technology that was so effective in the past.

But in all likelihood, there is nothing for the Republicans to be worried about.

Time’s Zeke Miller, in a post about the launch of “Ivy,” says that “the DNC spent much of 2013 testing and adapting the Obama developments on Terry McAuliffe’s Virginia gubernatorial campaign. The DNC officials said the tools were not available to New Jersey gubernatorial candidate Barbara Buono.” This statement is curious for several reasons.

While the intention is clearly meant to convey the effectiveness of Ivy (saying “we used it in the race we won, but not the race we lost”), in reality it does nothing of the sort. The Virginia governor’s race featured a well-funded Democratic candidate in McAuliffe against a Republican candidate, Cuccinelli, who was easy to demonize on social issues and who was abandoned by the donor class. It was a race where Cuccinelli was outraised $34 million to $20 million, and was outspent 10–1 in media buys in the final weeks of the campaign. Despite all this, McAuliffe, who was leading by an average of almost seven points in the polls closest to the election, could only pull out a two-and-a half-point victory. Under-performing polling by more than four points is not the performance you want to point to in order to gain confidence in your GOTV efforts.

Moreover, the fact that this “turnout machine” wasn’t made available to Buono raises even more questions. As has been pointed out before in this space, the main difference between Christie’s 2009 close win and his landslide reelection was Buono’s inability to turn out the Democratic vote. If it is true that the DNC chose not to share these tools with her campaign, she can lay a lot of the blame for her defeat at the feet of party leaders.

There’s another curious thing about this. In an Ars Technica story from February 7 about Para Bellum, just over three weeks before the DNC revealed that they had spent “much of 2013 testing and adapting the Obama developments,” Carol Davidsen, who was the Director of Integration and Media Targeting for Obama for America (OFA), said that she “was unaware of any effort by the DNC to do anything with… [any of the] tech developed by OFA’s technology team.” Similarly, after the 2012 campaign was finished, OFA campaign manager Jim Messina said, “I want to be firm about this — you can’t just hand this to the next candidate for president.”

Messina also said that “this organization [OFA] was built for people who support this president.”

That is the most overlooked flaw of the thinking that they can recycle what was successful in turning out voters for Obama for any other race. The other races aren’t running Barack Obama, and Barack Obama is the only candidate this technology was ever remotely successful for. Messina, at the POLITICO Playbook Breakfast where he made those same remarks, said that the tech from the Obama campaign was also utilized in the Walker recall election in Wisconsin. The results in that race and the Virginia gubernatorial race in 2013 make it quite clear that in any race where the president’s name isn’t on top of the ballot, the Democrat gets beaten in the GOTV effort —despite the vaunted Obama tech.

And finally, the financial situation of the two organizations needs to be considered as well. In the DNC fact sheet about Ivy, they say that “the DNC will invest millions of dollars in Project Ivy in 2014.” That is understandable; an operation of that magnitude can certainly cost millions of dollars and for many years would have been cost- prohibitive for both the RNC and the DNC. But at a time when the RNC (which has been consistently out-raising the Democrats) has $9.8 million cash on hand and no debt, and the DNC has only $5.7 million on hand and $15.9 million in outstanding debt, which organization is better suited to make actual gains on this front? The DNC may have no choice but to be content with the Obama model — no matter how ineffective it is for other candidates.

They might just not be able to afford to do any differently.