Screenwriters work weeks, months, sometimes years perfecting their screenplays. Each screenplay is a labor of love, creativity, passion, and commerce. But what if, after completing your screenplay, you found out the “success” of it – as measured by potential box office – could be determined in less than two minutes by using the DNA of your screenplay – its words – using machine learning? Instead of being scared by it, you may be exhilarated by it. You SHOULD be exhilarated by it.
Now you’re asking yourself, “Why would I want my baby (your screenplay) to be put in the hands of a machine?” Story guru Robert McKee has always cautioned that writers who get so close to their work lose perspective of whether their stories and scripts are good or not. McKee would bellow during his seminars that “You need to learn how to kill your babies!”
Hollywood has become fiercely competitive for content with movie budgets regularly hovering around the $100-million mark. Netflix is spending $7.5 billion on new content in 2018 alone. It’s not only antiquated to use script reading interns as the first line of defense, but it’s also risky and time consuming on an economic level. The #1 rule in Hollywood is simple – if your script can’t make money for the studio, it’ll never be made. That’s where AI and machine learning come into play.
Good screenwriters understand the structure they need to work in, including the conventions specific to their genre. In other words, every film script has a recipe that, over time, has proven to be successful – and you need the right ingredients in your script to sell in Hollywood. Scriptonomics’ machine learning software provides the writer with unbiased coverage of their screenplay by using the story’s DNA, without the guesswork of inexperienced interns. The story-scanner helps you identify the strengths and weaknesses of your story, how well it’s is paced, how it’s driven emotionally, which characters are essential to your story and which are not, and more. It evaluates the story DNA of your screenplay. Think of it as a gym. Imagine if a computer could analyze your body in two minutes, highlight your body’s weaknesses, and teach you to strengthen them. Would you be afraid of learning where you need to grow? Probably not. You’d be thrilled that you just saved yourself a lot of time and effort by pinpointing the areas that need work.
The way the Scriptnomics system works is by projecting your story arc onto a geometric story space, where a computer can compare your story to hundreds and thousands of the world’s most profitable and successful movies, and see how your story would fit into current movie audience habits, scene by scene and character by character. Think of it as a script reader with encyclopedic knowledge of every produced film script, capable of measuring your screenplay against their collective strengths. Back to our gym analogy, imagine using the training regime used by Olympic athletes to improve your performance.
While even the most seasoned screenplay reader could not hope to take this many scripts into account at once, machine learning ensures that Scriptonomics can. It finds story comparisons across genres and across the years to account for box office trends. Even if your screenplay is ahead of the curve, the system will predict when the time is right for your screenwriting vision to be produced, and audiences have caught up to your vision.
Scriptonomics champions a new approach to the creative process that they call “Quantitative Filmmaking” that aims to inspire and educate using the story arc rather than constrain or filter out stories using commercial criteria. “We are dealing with screenwriters who are passionately interested and emotionally invested in their screenplays, and we learned that we can’t be too analytical and cold about it because they have built emotional attachments to their story and their characters” said Tammuz, “We don’t want to be a company that molds dreams into marketable products, but we’d like it to become another tool that fosters screenwriters in their filmmaking process, to help carve their stories in such a way that it fulfills both their unique intent and design, and at the same time takes into consideration practicalities that will increase the chances of the final screenplay being produced and actually seen by mass audiences.”
Its online tools give you a Quantitative Filmmaking perspective of your screenplay, and even break down which scenes seem to be most significant mathematically.
Just like Syd Field revolutionized the concept of 3-Act Structure, and Blake Snyder popularized the concept of the Beat Sheet and the Three Act Structure, Scriptonomics demonstrates that the individual words used to a structure a screenplay can be measured and analyzed. The future of screenwriting technology goes far beyond just formatting software. Rather than fearing analytics and predictive learning, screenwriters should use AI to craft the stories that producers and audiences are looking for.
Typically, one might expect the scenes in a movie to build in importance as the story progresses, becoming more and more significant until the climax, as in the following figure:
But, depending on the story, that may not be true. For example, while Beauty and the Beast received a profitable score (and made $1.2 billion off of a budget of $160 million) its significant scenes arrived in a counter-intuitive order of the top scenes being 103-115, the second being 15-22 and the third being 35-39.
There are many factors that the system evaluates to determine profitability; one such factor is the internal patterns within the story and the geometric shape of the story in comparison to a vast amount of other produced films, where the order of the significant scenes plays a considerable part in that determination. That might leave you wondering what a significant scene is, which varies between each story and even each draft. Significant scenes are sections of the movie script that spike several elements of the story such as new characters appearing, a location appearing with a new combination of characters, an action sequence, or even unique dialogue spoken by essential characters. At times a scene sequence is misidentified as significant, in such cases, the sequence might be disruptive to the story and could be reworked to improve profitability. Furthermore, there are more numerical considerations that help determine profitability. For example, Beauty and the Beast is a very action heavy story, containing 63% action and 37% dialogue, which affects the type of stories it gets compared to and has implications on the genre of the film. This is visualized alongside the story’s tempo in the following figure.
If we take a look at the American Hustle (which is also very profitable) chart, we see that it has a similar ordering of significant scenes with the climax at the top and the inciting incident second. American Hustle is close to the opposite, regarding action and dialogue, containing 30% action and 70% dialogue descriptions. It also does a great job of including and utilizing all the main characters and scene locations throughout the screenplay, as seen in this figure:
There are more considerations when writing a producible screenplay, such as the script’s use of location. For example, if you compare the $190 million horror movie World War Z to the $1.2 million Saw, you can immediately see the difference in the sheer number of locations, as visualized below:
World War Z:
Other than the sheer difference in the number of locations, World War Z uses more one-off locations (62 out of the 80 locations to be exact) than Saw visits in its entire runtime (with 10 one-off locations out of the 22 used). These considerations have a direct impact on the budget needed and affect the complexity of the production. Shooting at an interior location is known to be cheaper and more manageable, logistically, than many exterior locations that may require special permitting and scheduling. The sheer number of locations affects the producibility of the story and may increase the amount of labor and budget needed to capture all the scenes of the story. Using this, screenwriters can work to limit their locations to only those essential to the story.
These are just some of the insights that a more sophisticated machine-driven script reader can give you. Other insights include charts of your character’s journey based on their dialogue and a breakdown of driving-characters created by analyzing the relationship between characters’ appearances and the weight of their dialogue. There’s even a model of your character’s relationships crafted through network analysis of the scenes. You can see an example of this for The Watchmen below:
Curious to see how your script would be interpreted? The first 15 readers that email firstname.lastname@example.org with the subject line “Creative Screenwriting Magazine Reader” will get a free breakdown of their script – whether a TV episode or a feature – and a consultation with script analysts to help them redraft their script. These readers will be featured in the next article reflecting on how this affected their writing journey. So, what do you think? Would you use such a tool in your writing process? Do you like this shift to a data-driven and quantitative filmmaking world? Comment below to let us know!