Legendary Applied Analytics: Making Movies the ‘Moneyball’ Way

How big data and analytics can change the traditional movie industry, and how far it can go?

Hollywood believes more and more in big budget ‘tentpole’ movies, and Legendary Pictures is one of the group, making many blockbusters such as Jurassic World ($1.7 billion box office revenue) and two The Dark Knight movies ($2.1 billion revenue in total). Legendary may spend up to $100 million promoting a new movie, a process it goes through four to eight times a year[i]. Big data and analytics became a secret weapon for Legendary to improve its economic performance in this risk-undiversified sector.

In 2012, Matt Marolda sold his sports player analytics business Stratbridge, which established six months before the publication of book Moneyball and focused applying the same analytics to predicting future performance of athletes to XOS Digital, joined Legendary building and heading the firm’s Applied Analytics team.

 

Value Creation and Capture

Marolda’s team uses deep analytics against multiple data sources to inform decisions from choice of actors, content of trailers to media buying.

First of all, the team assembled a database containing information including name, email address, demographics, interest, social media activity and movie viewing and game playing histories from different sources, such as Twitter, Wikipedia, blogs and vertical websites/platforms. The group surveyed samples from this database and scored them for their interest in different genres of entertainment. Then they analyzed what predictors in the database could best explain these scores and applied the result to score the whole database.

This proprietary database helped the team to micro-segment potential customers to thousands of groups, accordingly deploy different combinations of promotion elements on different media (majorly digital), and continuously monitor, test and optimize results on a daily basis during a movie’s campaign window.

By doing so, the Applied Analytics team made As Above So Below, a low budget thriller film, Legendary’s most profitable movie in 2013. The big data and analytic strategy also helped mitigating losses, for example, the team successfully predicted the disappointing box office revenue of Blackhat, and thus Legendary timely cut media spending before movie release and saved cumulatively over $20 million.

 

Competitive Advantages

The conventional movie production is done by alliances of relatively small independent producers who brought together stuff on a project-by-project basis, though ordinary audiences can only recognize names of the six dominant studios (Warner Brothers, Walt Disney, NBC Universal, Sony, Fox and Paramount), whose role is financer, marketer and distributor. And the studios’ standard marketing approach is to concentrate spending most of their budget 6 weeks before the opening weekend, of which the box office revenue is proved to be as an indicator for ultimate revenue performance.

Two major problems exist in this conventional practice. Primarily, customers usually don’t buy tickets till they go into cinemas for movie watching, this is especially true in the U.S market. By then, most of the promotion budget was spent by the studios, thus too little could they do to change ultimate performances.

Furthermore, movie marketing budgets in the US tended to concentrated in non-addressable traditional media such as TV and radio, which makes it difficult to measure marketing effectiveness and make timely adjustment.

US Marketing Budget Allocation

Trailers

5%
Television 60%
Internet 10%
Other Media 10% (includes radio, magazine, billboard)
Others 15% (market research, publicity, creative services)

Data source: HBS case: Legendary Entertainment: Moneyball for Motion Pictures, John Deighton, May 2016

 

Compared with the conventional approach, Legendary applied analytics strategy and tactics is a source of competitive advantage for the company. In January 2016, the Chinese investment group Dalian Wanda acquired Legendary, presented greater opportunities for the team to play in more markets. The industry believes that the analytics capability is an important strategic rationale behind Wanda’s acquisition. Moreover, thanks to the alliance between Wanda and two internet giants in China, Tencent and Baidu, Legendary is able to access huge amount of high quality data of Chinese customers. The application of big data and analytics contributed a lot to Warcraft’s success in the China market, according to Marolda (when visiting HBS).

 

Challenges

Even though Legendary is ahead of competition, it is still facing challenges. One major challenge is that in the U.S. market, due to above noticed ticket purchasing behavior, there is limited transaction data that can be collected timely. The Applied Analytics team can only optimize on something less than purchase, aka referral data rather than descriptive data. And this constraint poses a discount on the team’s work.

What’s more important, Legendary still needs to deal with disappointing box office revenue for movies like The Great Wall. This reveals the fact that movie production and marketing is essentially a matter of art and science, in which data and analytics can only go that far. In this sense, big data and analytics is not a magic mirror that can answer every question.

Related reading:

https://www.bostonglobe.com/business/technology/2016/03/31/making-movies-moneyball-way/Uzgwh2cdGthA1N3nZHqz0N/story.html

[i] http://data-informed.com/big-data-takes-a-star-turn-at-legendary-entertainment/

Previous:

Netflix: Reinventing Hollywood with Data

Next:

FiveThirtyEight and the Big Data Fail: Election 2016

8 thoughts on “Legendary Applied Analytics: Making Movies the ‘Moneyball’ Way

  1. Hi Yao. Thanks for the post. I had a few questions regarding Legendary’s strategy:
    1. Movies are making a lot of revenue from international markets. For e.g. Legendary’s movies like Dark Knight, Jurassic Park were big hits outside US. Do you think the localized and granular data approach that Legendary takes right now would remain relevant given the very diverse and wide international audiences that movies have to cater to?
    2. This approach appears to be similar to Netflix’s approach of creating relevant TV series based on the viewing data of users. However, Netflix is able to do this because they have extremely high quality data about viewing pattern of users (including time for which they watched a series, # of episodes etc.). However, in the movie world I imagine that this data would be more dispersed – for e.g. an average user might watch only 6-10 movies in a year and it might not be easy to get this viewing history of the user [for e.g. How will Legendary know that I saw 2 movies – as in how will it connect my ticket purchase for these 2 movies]. Do you think that under this scenario, their approach to develop movies based on historical data is a good strategy?

    1. Good questions Bipul.
      To your the first question, I do think Legendary needs to make adjustments on both the source of data they collect and the algorithm they developed in order to make this work in different international markets. The data part is easy to understand. For the algorithm part, I guess the good news is that they can leverage machine learning to self adjust based on previous results. And Marolda mentioned that because of the high adaptation rate of advance online ticket purchasing among Chinese consumers, their job, in some sense, is even easier in the China market.
      The second question, I guess the answer is that they not only try to leverage historical ticket purchasing data, but also collect more up-to-date data from multiple sources like social media. But you are right, descriptive data (in this case ticket purchasing/movie watching history) should be more relevant, while referential data is an alternative. We can only count on the rich sources of referential data, and according algorithm, to improve its predictive power.
      Hope I answered your questions.

  2. Thanks for the post Yao. This is truly fascinating. I always thought that creation of movies is done by top-down i.e. by initiatives and insights of few producers and directors. My question would be that while the Legendary’s strategy seems appropriate to satisfy current demand of audiences, I wonder whether the strategy fits with creating movies which receives high recognition of arts such as Academy Award.

  3. Thanks for the great post, Yao. I think that your parting thoughts about art and science are spot-on, and that reducing the creative process to a set of metrics and models is incredibly difficult. Comic book films and established franchises aside, outsized movie successes are often novel creative works for which no comparable prior data exists. I’m curious: given this and the failure of The Great Wall, do you think Legendary’s analytics capabilities and marketing knowhow can create truly novel hits, or are they best-suited to harvest interest in pre-existing properties like Batman and Jurassic Park?

  4. Wow this is extremely interesting. I had no idea that they spent that much money after the fact to promote the movie. This program feels very similar to what Netflix is doing when they decide what types of movies to make. Thank you for the post!

  5. Yao – thank you for such an interesting post! I wonder if this use of applied analytics as it pertains to the choice of actors will drastically alter the course of some actors’ careers. Given that an actor gains popularity through the movies in which he/she stars in, I wonder if using applied analytics will ensure that only the best actors/actresses receive parts in the best movies or if there will ever be a time where applied analytics would recommend a completely unexpected actor/actresses that might either add to the movie or take away. I also think of movie stars who have faded away based on a lack of movie opportunities. I’d be curious to know if you have read anything about how this is effecting actors and their ability to land parts with Legendary. Thanks!

  6. Thanks Yao! This is very cool. In SMICI we learned that movie studios essentially follow a blockbuster strategy where they outside spending on specific ventures “blockbusters”, and those ones tend to have outsized returns in the box office. The use of data here could essentially disrupt that model and help make Legendary a stronger contender to the big six studios you named. I do wonder though if all that matters is really just marketing spend — the more a movie is marketed to us the more likely we are to believe it is a “must see”.

    1. outsize spending**

Leave a comment