SEIU's Data Footprint In 2008

Part two in a three-part series of posts: The 2008 Data WarsRe-examined. Yesterday: Data suggests that Democratic microtargeting efforts were successful. Late yesterday: Catalist's after action report, posted in full. Today: what the SEIU learned from its data-crunching.

For several cycles now, the Service Employees International Union (SEIU) has been among the most active progressive groups in the liberal firmament, spending in excess of $80 million to influence the 2008 election alone. According to an analysis of their efforts, the SEIU was in regular contact with more than 4.5 million voters in ten battleground states, including more than 1.2 million in Virginia alone.

For the first time, thanks to the data crunched by Catalist and the Analyst Institute, we can now figure out what, besides money, distinguishes SEIU's efforts from others.

Primarily, the data shows that 64% of all SEIU voter contact work was focused on live telephone calls to prospective voters and 24% on neighborhood canvasses.  Both percentages are much higher than the average for all progressive groups, which tended to use less personal forms of contact, like automatic "robocalls" and direct mail pieces. But SEIU, able to pay for hundreds of its members to work full-time, largely eschewed mail and robocalls.

"Text messaging and face to face contact are really the keys to success," said Jon Youngdahl, the SEIU's national political director, in an interview. "Phones - [robocalls] --  dropped off in terms of effectiveness. They don't work for get-out-the-vote. They might be better as persuasion tools, rather than as get-out-the-vote tools," he said.
In several states, the number of unique contacts - that is, the voters contacted only by SEIU and not by the Obama campaign or any other group - exceeded Obama's margin of victory, including Indiana and Virginia. In Indiana alone, SEIU was responsible for contacting more than 40% of all independent voters who received attention.

Youngdahl said that the after-action report provided valuable insights for the SEIU's ongoing campaign to influence the health care debate. Based on the Catalist data, it has identified several million voters who identified health care as the issue that most persuaded them to vote. After the first week of the August congressional recess, when organizations opposed to health care reform pushed their members to attend town hall meetings, the SEIU tapped its health care voter database, and, according to its own calculations, managed to mobilize tens of thousands of them by the end of the month.  These include SEIU members, to be sure, but a large number of non-members identified through the union's independent expenditure program in 2008. (Catalist was able to assign each voter a "health care propensity score" -- the higher the score, the more likely the person would turn out.")

In 2010, aside from its efforts to turn out SEIU members in critical congressional districts, SEIU plans to focus on newly registered voters and those who were registered, voted in 2008 but did not vote in 2006. These "drop off" voters are generally the toughest to reach - they tend to be less politically motivated than the average member of the Republican or Democratic base, and they have weaker partisan attachments.  This is where the SEIU's contact data comes in. It now has several million pieces of information about what motivated these first-time or presidential-race-only voters to show up, and so, Youngdahl says, the literature and language used to persuade these voters will be better targeted to their worldviews.

Youngdahl, who has studied the Catalist data and SEIU's own internal reports, offers up several other lessons from 2008. Some are surprising. Those ubiquitous door-hangers that many groups left when the voter they attempted to contact wasn't home - turns out that the voters propensity to turn out increased by as much as 10% in some neighborhoods. And the act of asking voters when they planned to vote - either early or on election day - how they would get to the polls - and where they'd be coming from - their home or their worksite - "more than doubled the impact of a standard GOTV call," Youngdahl said.

Gentle psychological manipulation also seemed to bear fruit.  In general, once someone makes a promise, they are more likely to follow through if that pledge was made in the presence of someone else - even a stranger. Which is why, Youngdahl said, the SEIU made sure to recontact people who had pledged to their call-takers to show up and vote. Reminding people of their pledge increased the likelihood that they would go ahead and vote.

The Obama campaign's data now resides at the Democratic National Committee's Organizing for America arm, which, like SEIU, is trying to use the information to build support for the president's legislative agenda. Progress has been slow. As powerful as the models are, they are better at identifying people than figuring out what will motivate them to participate in a movement. Catalist is now working on this -- what it calls "response modeling," which Catalist CEO Laura Quinn calls "the next level" of data manipulation.

Republicans emain adept at using organizing and narrow-casting to mobilize activists -- niche targeting -- but their tactics tend to be useful for the blocking of a political objective, rather than moving people to embrace a big concept or policy change. This may, however, be a function of being the party that's not in power. The GOP's "Voter Vault" is state of the art, but Republicans have yet to set up a data consortium like Catalist that can store and model the data -- while maintaining legal and practical firewalls. Several senior party strategists said that had urged the Republican National Committee to study the problem, and outside the party, several targeting specialists are trying to find a business model to create a Catalist-Right, as one strategist called it. In theory, conservative groups cooperate through list exchanges involving affinity organizations like the Chamber of Commerce and the National Rifle Association. But in practice, legal and technological  hurdles have hampered the synchronicity.

A senior Republican official, who agreed to speak about internal party operations on the condition of anonymity, conceded that the Catalist model -- a for-profit, outside-the-party consortium -- provides the Democrats with an advantage, although the official said it was mostly financial. "They're essentially using money outside the party structure, and we're fighting with federal, hard campaign dollars.

The official pointed to several GOP innovations in 2008, including a new phone banking system that instantaneously updated data on the voter vault. 28 million voters were contacted using the system, the official said.