Guest blog: From a data standpoint, the movie industry is pretty much blind — here's how Hollywood can fix that
There has been an avalanche of articles about the lessons marketers can learn from Barack Obama’s 2012 presidential campaign.
There has never been a political digital campaign of such scale and scope, and the technology and data mining techniques applied by Obama’s team should be of special interest to Hollywood’s marketers. There are striking similarities between campaigning for the next box office hit and for the election of the next president of the United States.
Let’s begin by observing some similarities:
First Weekend vs. Election Day
There is no other industry where the rules of market economy condense with such brutality and drama into one single moment of truth as film. Years of financing and development rounds, months of production, weeks in the cutting room, and then one opening box-office weekend that determines much of the success or failure for a several hundred million dollar investment.
The whole film industry is watching the first numbers coming in on Friday evening from the east coast, first exit polls analysis are evaluated and predictions for the weekend’s box-office results and overall success of a movie are made on Saturday.
The election-day experience is of noteworthy similarity. Years of campaign fundraising, months of campaigning, weeks of unremitting voters mobilization by campaign staff and then one moment of truth — election day. First, numbers from the east coast come in, exit polls, first predictions, calculations based on electoral college and popular vote numbers, all to measure the success of a campaign investment. The exact same experience.
Swing Audiences vs. Swing States
Over half of movie tickets are bought annually by only 10 percent of the population. Many of these frequent moviegoers decide on their movie choice very early, and some of them decide what they want the moment they see their options at the ticket booth (“Swing Audiences”).
Additionally, there are the people who normally do not go to the movie theater at all but when they do, make a significant impact on results (“Sluggish Moviegoers”).
All of these audience segments are important for the studios. And every movie marketer must achieve two goals to see success with these groups: to get them off the couch and into a theater seat for opening weekend, and to get that prospective audience to make the right choice at the ticket booth.
The same applies to presidential campaigns: The nominees and their staff must get people to make the correct, favorable choice and then mobilize them to actually leave their homes and fill out the ballot on election day.
Needless to say, there are major differences. While political campaigning is based on programs and hard facts that will directly affect the future life of voters, the only selling proposition in movie marketing is creativity and storytelling: to build a strong emotional connection with potential audiences based only on a narrative, casting, artwork and a trailer.
It is by far the most advanced industry in creating this magic connection with its audience. Data and technology will never change that. But it might help to deliver the magic in a more impactful and efficient way. In this sense there might be a few interesting lessons to draw from this year’s Obama campaign:
Use Data to Develop a Holistic View
The biggest change in Obama’s '12 campaign was that his CTO, Harper Reed, and his staff integrated the existing databases from prior elections, polling, fundraising activities, field work and much more into one massive database. Only by doing this, was the data analytics team able to develop a holistic view on voting and non-voting America.
By cross-validating the patterns, they were able to draw conclusions from the insights the organization was getting across the United States. The profiles of people signing in to www.barackobama.com with Facebook Connect could be cross-referenced with the profiles of followers on Twitter and pre-registered voters, then further with the fund-raising supporter lists, and door-to-door surveys conducted across the US.
With this holistic view, it was possible to segment American citizens not only by location, ethnicity, household income and political interests but also by two scores that were calculated in the system: a score between 0-100 on how likely these segments were to vote for Obama, and a second identical scale gauging how likely these segments were to actually show up at the polls come election day.
The movie industry does not have this holistic data view yet; there is precious little data on moviegoer's profiles per title. The awareness and intent of audiences to watch a specific movie are measured in small samples by age and gender, and even then only weeks before opening day.
Further, they do not cross-reference with the other angles and data insights the marketers are able to access, so they are difficult to validate in terms of ethniticity, location, household income, interest, etc. The studio’s tracking data do not cross-reference with their Facebook fan relationships, Twitter data, advertising click through rates, TV consumption patterns, home-entertainment purchase data, media affinities, and so on.
From a data standpoint, the movie industry is pretty much blind.
Test, Measure and Iterate
From the beginning, Obama’s campaign manager Jim Messina had promised a totally different, metric-driven kind of campaign in which politics were the goal, but political instincts might not be the means. With this approach, the Obama team transformed marketing into engineering. Political instincts were used as hypotheses that could be tested with real time data.
The process of campaigning became an iterative process with detailed feedback metrics, making investment and resource allocation decisions much more accountable.
While the movie industry has its tracking techniques that respond to big TV campaigns, this tracking data is rather coarse and un-detailed. And since it is mainly conducted via telephone landline calls, it is not monitoring some of the most important age groups and ethnicities that do make a difference at the box office.
If you only measure and monitor a small fraction of the market with a very rough methodology, it is very hard to allocate your marketing budgets effectively and efficiently. You neither know who you activated through your advertising nor what was the approximate cost per activation.
While instincts and experience always will be important in the movie industry, a more data and structured signal-echo process would reduce a lot of wastage and lead to a higher marketing efficiency.
Concentrate on Scope, Not Only Scale
Once Obama’s team had modeled reliable probability scores within their big sample set, they could extrapolate the findings to a broader population and prioritize their target segment. Cross-referenced with latest polling numbers, Obama's team could estimate which target segments of non-voters or undecided voters might actually make a difference in terms of winning the election in particular states, regions or even neighborhoods.
With this differentiated probability map the team could much more efficiently allocate their media budgets and campaign activities.
They could hone in in targeting those segments and driving up their probability scores with the most effect per spent campaign dollar or hour of volunteer work. To be fair, this is not at all a new approach, and on a macro level, Gov. Romney’s team did a similar job. They did not spend their media budgets on unwinnable states like California, where media budgets would have had a very low return on investment.
The key is, Obama’s team optimized this targeting technique on the street level, rather than simply the state level.
The movie industry does not have enough data to model its audiences by probability scores in a differentiated way. It concentrates parts of its campaigns to reach particular audience segments like “moms” or “gamers” where the strongest affinities are supposed. But this is done in a very undifferentiated way — and more importantly, it does not have any feedback process to ensure the effectiveness of these more targeted parts of the overall campaigns.
With a thorough, data-centred approach that monitors the market in an elaborate way, marketing becomes a metric-driven feedback loop that can be used to hone in on the right positioning, and to deliver the most effective message to different market segments.
The key in this process is to begin early. Collect data as soon as possible, derive hypotheses from early population samples, test those hypotheses against broader populations, optimize, and verify the hypotheses. The value of early iterations of campaigning is not to win early votes but to gain data insights that can be used for the following campaign iterations.
Although the numbers were not accessible at this time, one would safely assume that Obama spent bigger portions of his budgets earlier than Romney, moving away from a “Total Awareness Buyout” Strategy one or two weeks before election towards a more engineering-typical approach of Agile Advertising.
The vast majority of the marketing budgets in the movie industry are spent within two weeks before the theatrical release. While that is understandable — and probably similar to Obama’s timing of spends, it means that by the time the big chunks are invested into TV advertising, the marketers do not yet have a clear picture of their most important audience segments and on the best technique to win them on opening weekend.
By spending a bit earlier movie studios would enter the metric-driven feedback loop earlier and could improve their market understanding. Every penny spent months before release would — in a data-driven approach — make every dollar spent closer to release more efficient and effective a thousand times over.
(Note: Some information was gleaned from this article.)