about us - ScriptBook

Justine Harcourt de Tourville | HARCOURT DE TOURVILLE
29 JAN 2015
Museumstraat 3 | 2000 Antwerpen | Belgium
Despite a weak economy in 2013, the global box-office reached a record high of nearly $36 billion.
According to PriceWaterhouseCoopers, the global box-office will hit $45.9 billion in 2018, in part
due to expanding markets in China, and sustained growth in the US, UK and Japan.
In total, worldwide filmed entertainment revenue is projected to surpass $100 billion in 2017.
Deciding which script to greenlight carries tremendous financial responsibility—average cost to
produce, make and release a film is $74.9 million (MPAA 2011).
Box office disasters can be felt across the industry: movie studio stock prices fall, studio executives
and literary agents are fired, actors & directors lose bankability.
Gigli, a 2003 film that cost $75.6 million to produce, only earned $7.2 million:
"This is not just ordinary bad … but a hypnotic, black hole of a movie that sucks
reputations, careers and goodwill down its vortex." —Liam Lacey, The Globe and Mail of
The Lone Ranger, which came out ten years later in 2013, had to pare down the budget to be made
for $215 million. It failed miserably garnering scathing reviews (projected loss between $160-190
million). The chief analyst for Boxoffice.com, Phil Contrino, called it "the kind of bomb that
people discuss for years to come."
Studies by DeVany & Walls have shown:
Overwhelming majority of films are bad bets: 87% lose money.
Few subsidize entire industry:
80% of Hollywood's total profit
earned by a mere 6.3% of the films.
Museumstraat 3 | 2000 Antwerpen | Belgium
Many factors can contribute to films failure to earn a profit: Bad acting, bad directing; high
production, distribution and/or marketing costs. But one of the better predictive factors to a film’s
success is the screenplay.
In his book, Great Flicks: Scientific Studies of Cinematic Creativity and Aesthetics, Dean Keith Simonton
demonstrates statistically how an Oscar-winning screenplay has the highest correlated factor
(=r.63) to positive critical evaluations than any other element: directing, acting, picture, song, etc.
In other words, scripts have an inherent predictive value.
The film industry relies on outdated and error-prone methodology when considering scripts.
In 2012 in the United States, a total of 677 movies were released in theaters (MPAA 2013). That
figure compares to–using the best guess of industry insiders–the yearly “25,000 to 40,000 of scripts
that filter through the studio acquisition system.” (Scott Meyers 2012)
The predominant method of acquiring a script involves literary agencies; they in turn employ story
analysts to read scripts and make recommendations in what is known as "coverage." Scripts with
positive coverage are then forwarded to the film studio or production company for consideration.
Whether a script should be pursued starts with a subjective decision based on the story-analyst’s
personal preferences. Some high-potential scripts never make the cut, while others (which will
eventually bomb at the box office) are recommended.
“The most famous dictum about Hollywood belongs to the screenwriter William
Goldman. ‘Nobody knows anything,’ Goldman wrote in Adventures in the Screen
Trade ... ‘Not one person in the entire motion picture field knows for a certainty
what’s going to work. Every time out it’s a guess.’ One of the highest-grossing movies
in history, Raiders of the Lost Ark, was offered to every studio in Hollywood,...and
every one of them turned it down except Paramount: ‘Why did Paramount say yes?
Because nobody knows anything. And why did all the other studios say no? Because
nobody knows anything. And why did Universal, the mightiest studio of all, pass on
Star Wars? … Because nobody, nobody—not now, not ever—knows the least
goddamn thing about what is or isn’t going to work at the box office.’” — from
Malcolm Gladwell's article "The Formula" in The New Yorker 16 October 2006
Museumstraat 3 | 2000 Antwerpen | Belgium
In 2005, an executive with Leonard diCaprio’s production company asked development executives
about the best script (not in production) they read in the past year. The results were tallied; “The
Black List” was launched, creating an internal industry ranking system among those with the
expertise–and capacity–to greenlight a script and put it into production.
The Black List has grown in size and reputation (up to 500 executives participate in the tally). It has
its own official blog and script-reading service. But as Indiewire.com reports, The Black List’s track
record is the same as chance: “scripts near the top have been made into terrible movies, and those
in the lower reaches have proven to be among the best.”
Movie star Will Smith is a believer in pattern. According to a MIT paper on prediction and
customer wants, Smith methodically reviews the box office top performers to look for patterns and
identified trends that helped him choose movies. His strategy paid off. His 2007 film Hancock,
suffered terrible reviews but still earned $625 million worldwide. Smith attributed his selecting
films according to trends in high box office performance: special effects, special effects with
creatures, and special effects with creatures and a love story—and chose films with these elements
with grand success. (Davenport and Harris MIT Sloan Review Winter 2009).
Similarly, film production companies & studios need a decision support tool for scripts in the
greenlighting stage. In order to capture more of the commercial potential between what is made
and what is successful, increasing the odds of an investment return begins with better scripts.
Most academic research focuses on forecasting box-office performance after a film has been
produced and sunk costs have already been incurred. The industry needs an objective tool to assess
a script’s commercial potential prior to production, investment and sunk costs.
Museumstraat 3 | 2000 Antwerpen | Belgium
Other creative industries have already looked to predictive intelligence to deal with subjectivity in
finding commercially-viable work with success. With revenue diminishing by 50% over the past 15
years, the music industry had to explore other channels to find hit music. MusicXray created a web
application that performs multivariate analysis on a song based on elements like beat, melody,
harmony, etc. MusicXray can predict hits with 80% accuracy. For the music industry, minimizing
the subjective guessing of A&R executives in charge of locating talent, as well as the risk of signing
unknown artists, was an economic advantage in lean times.
Epagogix a U.K. company led by Nick Meaney relies on neural networks to make predictive
analyses about films that should or should not go into production by looking for script elements that
correspond with either success or failure at the box office. They try to assess risk, for example, by
looking at factors such as cast and location. Epagogix has found that A-list actors and
directors are for the most part irrelevant to a film's bottom line.
The highly successful The Pirates of the Caribbean and the ill-fated Lone Ranger films illustrate this
finding. Both films were produced by the same studio (Disney), same director (Gore Verbinkski)
and boasted the same star (Johnny Depp). While the same writing team from Pirates worked on the
Ranger film in the development at the mid-point, the Lone Ranger was subjected to numerous and
substantial rewrites. The script was considered significantly weaker.
One problem with neural network analysis, however, is the immense need for input. Outcomes are
rooted in data generated from screenplays past—problematic for cultural evolution. Data from
neural networks, for example, would predict James Bond needs to be a white male to perform
successfully at the box office; yet, the black actor Indris Elba is currently the subject of a public
petition promoting his selection as the new 007.
Concentrating exclusively on storyline, with highly accurate predictive success, Nadira Azermai
and Bart De Maesschalck founded ScriptBook in Belgium. Spearheading the effort to develop
premium decision support for filmmakers, agencies, and studios, ScriptBook uses artificial
intelligence, machine learning and natural language processing to remove subjectivity from
determining a screenplay's box office potential.
Museumstraat 3 | 2000 Antwerpen | Belgium
At its core, ScriptBook’s predictive system relies on textual information, but integrates
screenwriting expertise so it is able to distinguish characteristics between genres. A comedy will
not have the same considerations as a drama, nor an indie-cult film the same as an action thriller;
creative features are incorporated into the evaluative process. ScriptBook is a product designed to
be a tool for the numbers-centric studio heads, in addition to filmmakers.
Just as every sector has been challenged by technology, creative industries are not immune to the
shifts and change instigated by the digital world.
With funds tightening, studios either cut back or hedge bets with franchises. Either way, investors
are increasingly careful about the size and frequency of their investments. Even though formulaic
films are viewed as "safe" (the story has already been proven), film franchises steadily decrease their
box office return over time (witness: Indiana Jones, Star Trek, The Hangover, and The Pirates of the
Caribbean)—ultimately leading to "franchise fatigue," which The Hollywood Reporter attributes as the
source of 2014's poor year in film:
"The only thing the theater business needs is more movies; big four-quadrant
movies [meaning appeals to all demographics, ed.], quiet dramas, niche
comedies — you name it," says Patrick Corcoran, V.P. of the National
Association of Theatre Owners. "There were 45 percent fewer $100 millionbudgeted movies in the summer, and box office was off 15 percent. That's the
entire difference between 2013 and 2014." —from Pamela McClintock's article
"What's Behind 2014's Box-Office Slide: Franchise Fatigue, Fewer Big Movies" in The
Hollywood Reporter January 7, 2015
Demand is there, but originality and breadth are missing.
Jeffrey Katzenberg remarked "the movie marketplace is very different today than it was three or
four years ago," as he stated in Variety magazine, 'It’s much more competitive,' in terms of
playability, marketability and the availability of release dates." (Graser Variety 22 January 2015)
But with 87% of movies failing at the box office, it's easy to understand the reluctance to change—
no industry wants to lose more.
How can hard science help?
Museumstraat 3 | 2000 Antwerpen | Belgium
Already in use within the creative-based industries in the form of artificial intelligence and behavior
prediction, content-filtering is used by Amazon and Netflix to suggest products, books, and movies
based on customers' previous purchases. The music industry has adapted to using predictive
intelligence by parsing song components in order to scout artists. Both Will Smith and some major
film studios already look at data-driven analyses to forecast box office performance. Data science in
film would be neither novel nor unwarranted in this financial climate.
Implementation remains tricky, however. Capitalizing on his background in actuarial science, Nick
Meaney has positioned Epagogix as a process that offers film investment risk mitigation with
predictions extracted from massive data inputs that can save studios money. Epagogix still has two
1) The inability to forecast future shifts in cultural mores, i.e. the outdated tendency to
cast villains who are dark-haired or foreign, for example, will automatically suggest that future
films continue to typecast villains along these lines.
2) Technological tools feel threatening to Hollywood insiders because it renders hardearned human expertise and insight obsolete. Bankability, box office points, and perks (like invites
to the important parties) stem from having power and knowledge. A tool perceived to eliminate
creative–human–input, such as pointing out that movies, where the hero-wearing-a-blue shirt has
greater box office potential, is going to ruffle creatives, who envision the hero in a red shirt.
Still, studio heads are increasingly engaging Epagogix to find ways to save—often discovering that
they will make more money without a certain mega-star attached, which dispels conventional wisdom
that stars are box office draws and essential to movie success.
More friendly to stars and creatives, their teams, and ultimately the studios, ScriptBook focuses on
the storytelling metric. Stories–the backbone of a script–are time-tested; the three-act structure is
found in the writings of Ancient Greeks and Shakespeare.
Not only are agents given a reprieve from reading scripts late into the night because there is a
subjective tool to separate the good from the bad, ScriptBook has the ability to position itself as a
support tool to creatives. Since outcomes are derived from creative elements like story structure,
language, and genre, ScriptBook generates data unaffected by social mores—and well-before the
investment and greenlighting phase, too. Using story alone, ScriptBook can predict that a
Hitchcock thriller ranks high, Gigli and The Lone Ranger do not; an analysis that could have saved
hundreds of millions of dollars—and Jennifer Lopez and Johnny Depp some credibility.
Museumstraat 3 | 2000 Antwerpen | Belgium
A story-based predictive system benefits actors, directors and agents, in addition to film studios,
because not only investment loss is averted, but stars' bankability remains intact. ScriptBook steers
users away from a bad script allowing an actor or director (agents and production companies) to
avoid being attached to box office bombs, or worse seen as overpaid or lackluster. Investors win
because bad stories, the best predictor of success, go unmade.
Like Epagogix, ScriptBook's predictive prowess improves with more data, though ScriptBook only
requires a simple input–screenplays. ScriptBook's biggest challenge will be developing product
awareness and connecting with Hollywood players from their tiny offices in Belgium. Epagogix has
done fine from London, proving that there is both a demand and need for big data in the film
industry, irrespective of headquarters. Whether creatives embrace either hard-science product,
however, remains to be seen.
Museumstraat 3 | 2000 Antwerpen | Belgium
As Studios Cut Back, Investors See Opening
By Michael Cieply, The New York Times November 14, 2010
The Cruel Math Behind Why Streaming Will Never Save The Music Industry
By Mark Rogowsky Forbes.com 3/20/2014
The Formula - What if you built a machine to predict hit movies?
By Malcolm Gladwell The New Yorker October 16, 2006
20 Best & Worst Films Made From Black List Scripts
By The Playlist Staff, indiewire.com April 10, 2014 http://blogs.indiewire.com/theplaylist/the-10best-10-worst-movies-made-from-black-list-scripts-20140410
The Definitive Spec Script Sales List (1991-2012): 2012
by Scott Myers Gointothestory.blcklst.com - Official Screenwriting Blog of The Black List
July 4, 2012
Odds of Selling a Screenplay
by Unk screenwriterunknown.com July 14, 2011
Theatrical Market Statistics 2013
Motion Picture Association of America March 2014
Global and entertainment media outlook 2014 - 2018
Filmed Entertainment
Paul Orchard & Alayna Francis, PriceWaterhouseCoopers
When Number Crunching Met the Creative Industry - (video)
Nick Meaney, Keynote Speaker at Casualty Actuarial Society Seminar, March 13, 2013
Great Flicks: Scientific Studies of Cinematic Creativity and Aesthetics
By Dean Keith Simonton, Oxford University Press, January 26, 2011 p.182/240 pp.
Masked Lawmen Stumbles at the Gate
by Brooks Barnes, The New York Times, July 7, 2013.
What's Behind Hollywood's Big Slide: Franchise Fatigue, Fewer Big Movies
by Pamela McClintock, The Hollywood Reporter, January 7, 2015
Museumstraat 3 | 2000 Antwerpen | Belgium
for more information, please contact: [email protected]
Museumstraat 3 | 2000 Antwerpen | Belgium