An Interview with Obama for America Targeting Guru Dan Wagner

_­Happy Holidays from the CRI team! Since a recurring theme in our research is that more data and precise targeting would improve the distribution of independent films, we interviewed former Obama analytics guru and entrepreneur Dan Wagner. Dan revolutionized targeting for political campaigns by working as the Chief Analytics Officer on the 2012 Obama campaign.  Previously in 2008, Dan had worked as the National Get Out the Vote (GOTV) Targeting Director for Obama and as the National Targeting Director at the Democratic National Committee (DNC) for the 2010 election cycle. In 2012, Dan and his 54 person team of analysts worked to create the analytics models that targeted undecided voters and supporters through media outlets and offline canvassing and phone calls.  You can read more about Dan's work for the Obama campaign and how it relates to targeting for film, in our blog post here.

After the campaign, Dan received funding from Google's Executive Chairman, Eric Schmidt, who was a technology advisor for the campaign, to start a new firm called Civis Analytics. Headquartered in Chicago, Civis uses the technology and analytics strategies from the Obama campaign to help companies, non-profits, and campaigns utilize their own data to target their audience/consumers and build stronger data driven organizations.

We interviewed Dan to get his thoughts about how an analytics model could be applied to help distribute film, and the conclusions were fascinating. Here, we’ll share some of what we learned, with accompanying quotations from Dan.

1) Start with who likes you, not who doesn’t like you.

There’s an automatic assumption when we think about data that the starting point should be to gather as much as possible about the entire pool of potential consumers/audience members/voters. But Dan told us that during the whole endeavor of gathering data on the Obama campaigns began in Iowa in 2007, focusing on turnout – in other words, getting as many likely supporters to turn out for the caucus as possible. Dan and other team members used statistical models to identify people likely to support then candidate Obama, then integrated these models to the voter file and what the other operations of the campaign were doing (ie constituency outreach, volunteer recruitment). In other words, you can do more by perfecting the profile of someone who does like what you’re offering than you can by trying to deal with the whole sea of data out there.

In the film world, Dan yielded the notion that probably, through things like AMC rewards cards, etc., some large consumer information company is indeed analyzing that large sea of data about filmmakers in general, for the sake of the exhibitors and the distributors. But in the independent film world, why worry about coming up with a system to try to do the same? Which leads to the second conclusion:

2) The most valuable data is the data under your feet.

Dan encouraged independent filmmakers to look wherever they could and do whatever they could to gather data sets about the audiences going to their films. That could be as simple as a Facebook page; that could mean trying new things like a sign-up sheet after screenings. By pooling and cross-referencing these data sets, you can come up with a more and more finely attuned profile of the kind of audience member that likes films like yours. To further this process, Dan says the question becomes “When we create a product, how can we display it to a range of people to see how we can maximize the potential of a more targeted approach?” In other words, “test drive something specific, and see who likes it.” Michael suggested limiting the test art house theaters in New York, like BAM and Nighthawk, but Dan warned against a regional approach.

You might be able to do a limited release of an online panel of your movie, and then have a Google consumer survey where people can report back on their movie, and then you could say these are the people the movie appeals to across the country. So you're looking at it more as broad customer feedback, and looking at that feedback in terms of who likes this and how you could go wider. You could go in New York and look at who liked it but that's going to be a limited subset of people who already go to see independent films in New York City.

3) Surprise surprise: a consortium of data-gathering films would greatly enable the ability to use that data wisely.

A recurring theme (or pipe dream) in our research is the possibility of a grassroots collective of filmmakers that pool their resources – in other words the equivalent of an Obama for America infrastructure, but for film. Dan hit on this when he talked about how each state used to be doing its own data management, until the DNC stepped in:

Many years ago every state was doing its own thing… which is trying to define people that like Democrats. And what we did in the DNC in 2010 was we said: …[W]e're going to combine all this data and define for you who we think like Democrats and you'll have more precision because we'll have more time to spend on architecture and it'll get faster and cheaper. And before… the type of research they could do was bounded by their own capacity and resources.

And that's the same position someone who's a filmmaker and a small team is in. They can only collect the kind of research that is within their resource set which is probably small… But working together as a lot of people who are thinking about a similar audience you could probably do that.

The fundamental genius of the DNC was they went to these people and they didn't say give us your data. What they did is they said, give us your data so we can poll it. And you're going to get the technology infrastructure and the historical list of ID's that we will hold in that store for you. That's a good deal. So when you think about these people who are participating in these data agreements, they need to have an incentive to do it.

Dan connected the dots already, but a consolidated grassroots consortium could give filmmakers a similarly solid incentive to join: data management handled in house, and a more targeted idea of what an audience member who could like their next film would look like, for the sake of craftier, more efficient marketing efforts next go-round. As Dan puts it: “A bunch of people working together, like a mini-studio, could get consumer info – ‘we looked at 20 films, these are the people who like them… These are the people who like independent movis’ and then over time validate that.” In other words, this mini-studio or collective, could create pretty accurate profile of someone who likes independent movies in general, “and then generalize it for lots of promotion afterwards.”

4) This sharing of data amongst films necessitates a broader, more shared, more abstract notion that an audience could opt in to.

Dan spoke to the power of creating a broader banner under which a collection of films could solicit this data. Giving information on behalf of a film is a strange ask of an audience member; giving information on behalf of a bigger idea about what film could be is not so strange and potentially more appealing. Dan cited the example of Beasts and the limited grassroots work we did with that film.

You could imagine a case where you have a second or third movie that in principle is as good as Beasts. And you formed a way to share with a second audience and you say that this is not just about Beasts this is about bringing you into a community of independent art and you're going to be a part of that. And people say, 'Oh, I'm in.' And you're kind of building a community of people not just for that one movie but about all these other movies that are going to come out. You're building this community of people and also a set of data in terms of a list that grows over time of people who care about this. To use an organizing model that suggests sharing data between films over time. Because once a film is done, it's done, but you can capture that behavior and pool that into new films that are coming out overtime. You could offer them previews, fundraise for Kickstarter, all sorts of things.

In conclusion, Dan asserted that there is a different feeling one has when watching a film like Beasts than there is when watching a big blockbuster like Superman. Perhaps the starting place for the creation of a grassroots collective would be: what is that feeling? How can we articulate it in a mission statement for a large organization that could appeal to audience and induce them to offer information about themselves, in the name of that mission statement/feeling?