How Data Can Save Hollywood
Prerna Gupta
@prernagupta
Can art and data be friends?
A good story is worth billions
Twilight
$829,685,377
The Hunger Games
$864,912,963
Alice in Wonderland
$1,025,467,110
Skyfall
$1,108,561,013
Transformers
$836,303,693
Spiderman
$890,871,626
Hollywood’s innovation crisis
• Blockbusters require $100M+ budgets
• Project greenlighting driven by past & politics
• Difficult to get actionable data during production
Case study: Jupiter Ascending
• Made by Wachowskis (of Matrix fame)
• Astronomical budget (300M+)
• Nerd rapture (space travel, genetics, Mila Kunis)
Result: box office bomb
Loss of $150M+ at box office
“Jupiter Ascending is a gigantic, soulless misfire from the
Wachowskis...”
“Aggressively dull and terribly paced.”
“Your movie’s descended into Uranus”
How can Hollywood increase hit rate?
Hint: 50% of Top Grossing Movies
are based on successful books
Case study: iterative writing
• Launched MVP, one chapter at a time
• Got feedback from early adopters
• Early evangelists spread the word
Result: from blog to blockbuster
• Production budget: $108M
• Worldwide gross: $440M
…and counting!
But What About Data?
How do we know a story is good?
Data can identify a good story
20 40 60 80 1000
50
25
100
75
% Read
%UsersRetained
Uglies
Avalon
Graceling
Grisha
Delirium
Fiction has a UX problem
Solution: lean fiction
Completion rate skyrocketed: 85%
Case study: power of data
20 40 60 80 1000
50
25
100
75
% Read
%UsersRetained
Result: completion rate doubled
20 40 60 80 1000
50
25
100
75
% Read
%UsersRetained
Does data kill creativity?
Prerna Gupta
@prernagupta

How Data Can Save Hollywood

  • 1.
    How Data CanSave Hollywood Prerna Gupta @prernagupta
  • 2.
    Can art anddata be friends?
  • 3.
    A good storyis worth billions Twilight $829,685,377 The Hunger Games $864,912,963 Alice in Wonderland $1,025,467,110 Skyfall $1,108,561,013 Transformers $836,303,693 Spiderman $890,871,626
  • 4.
    Hollywood’s innovation crisis •Blockbusters require $100M+ budgets • Project greenlighting driven by past & politics • Difficult to get actionable data during production
  • 5.
    Case study: JupiterAscending • Made by Wachowskis (of Matrix fame) • Astronomical budget (300M+) • Nerd rapture (space travel, genetics, Mila Kunis)
  • 6.
    Result: box officebomb Loss of $150M+ at box office “Jupiter Ascending is a gigantic, soulless misfire from the Wachowskis...” “Aggressively dull and terribly paced.” “Your movie’s descended into Uranus”
  • 7.
    How can Hollywoodincrease hit rate? Hint: 50% of Top Grossing Movies are based on successful books
  • 8.
    Case study: iterativewriting • Launched MVP, one chapter at a time • Got feedback from early adopters • Early evangelists spread the word
  • 9.
    Result: from blogto blockbuster • Production budget: $108M • Worldwide gross: $440M …and counting!
  • 10.
  • 11.
    How do weknow a story is good?
  • 12.
    Data can identifya good story 20 40 60 80 1000 50 25 100 75 % Read %UsersRetained Uglies Avalon Graceling Grisha Delirium
  • 13.
    Fiction has aUX problem
  • 14.
    Solution: lean fiction Completionrate skyrocketed: 85%
  • 15.
    Case study: powerof data 20 40 60 80 1000 50 25 100 75 % Read %UsersRetained
  • 16.
    Result: completion ratedoubled 20 40 60 80 1000 50 25 100 75 % Read %UsersRetained
  • 17.
    Does data killcreativity?
  • 18.

Editor's Notes

  • #2 Think about the last movie you saw, or book you read, that moved you? Took you on a journey. And maybe even changed the way you look at the world. What was it that made this story so special? It was likely something about the characters, the world and the plot that worked together to get you to suspend your disbelief and keep you engaged through the course of the story.
  • #3 We tend to think of storytelling as an intuitive artform. A product of raw talent / creative genius and moments of inspiration. What place does data have in this sort of creative endeavor? [This is the question I kept asking myself when I began writing my first novel a couple of years ago. It’s a sci-fi fantasy trilogy set in Silicon Valley 100 years in the future. I’m a tech entrepreneur by training. And in my last startup I…] But what I want to show you today, is that great stories are not really that different from great tech products. They are developed through a process of iteration and benefit greatly from data.
  • #4 And in fact, there is good reason to incorporate Lean methodologies into story development Because if you mange to develop a good story, it can literally be worth billions Hollywood is a massive, multi-billion dollar industry. And when their products succeed, they are hugely profitable Hollywood has already grossed over $8B this year domestically. And what many believe will be the biggest of all time is scheduled for release in Dec (star wars) Did $10B last year domestic, $39B worldwide; $88B total “filmed entertainment revenue” Ticket revenue was down more than 5 percent in 2014, to an estimated $10.35 billion, compared with $10.92 billion last year.
  • #5 But for every profitable blockbuster, there are many more flops And there is a general sense of crisis these days in Hollywood The movie production model is broken Budgets are skyrocketing. To give yourself a shot at making a billion dollars, you need to spend several hundred million There is no objective method for greenlighting one project over another And, there’s now way to get feedback during development They do focus groups AFTER production, but very little can be changed at this stage. And most of the $100M has already been spent. - But big revenue requires increasingly big production budgets - bc blockbusters make all the money - but creating a blockbuster takes huge capital investments - and returns are unreliable, at best - Hollywood’s average return rate last year was XX% (negative) - software analogy - can you imagine a company pouring $100M+ into developing a product, without testing a beta first?
  • #6 - To give you an example of this, let’s look at a big-budget movie that was released earlier this year, called Jupiter Ascending directed, written and produced by two of the most beloved directors of our time: the Wachowskis, who created the Matrix Had one of the biggest budgets of all time And addressed classic sci-fi themes, which on the face of it, might have been the building blocks of a great story
  • #7 So how’d it do? it was a huge flop at the box office And it was slammed by critics What was so bad about it? Well remember all those things I mentioned at the beginning that make a story great? Namely, the world, the characters and the plot? They basically got it all wrong. - so what’s the solution? - how can hollywood introduce lean startup into their process? - after all, you can’t a/b test a movie…can you?
  • #8 - well, the market is starting to show a path - if you look back at that list of Top 50 grossing movies of all time, you’ll notice an interesting pattern: 50% of them are based on successful books A book is increasingly become a place for market validation of the most important parameters in a story: namely, the world, the character and the plot And guess what? Books are insanely cheap to produce!
  • #9 And books are much easier to test during the creative process Andy Weir, a software engineer based in Mt View, started posting his story The Martian on his blog a few years ago One chapter at a time He began attracting an audience of early adopters, and this audience gave him feedback, which he incorporated into the story As he iterated, his audience of early adopters grew, and they began encouraging him to release the book on Kindle He released it for $.99 and it shot to the top of the charts And the rest is history andy weir / the martian backstory Trend is toward iteration and market validation Much cheaper to produce stories in text first
  • #10 andy weir / the martian backstory Trend is toward iteration and market validation Much cheaper to produce stories in text first! Worldwide gross: $440M Production Budget: $108M
  • #11 Explicit customer feedback is not the same thing as data Is there a way to use statistics to determine when a story is resonating with your audience? Sort of like how we use engagement metrics, and other KPIs, when we develop software products? I was very interested in answering this question a few months ago My background is in mobile app development. I previously developed music apps, like AutoRap and MagicPiano, that have over 150M downloads And we used Lean Startup methodologies extensively to improve these products Last year, I began writing a novel – a sci-fi fantasy trilogy – and as I was writing, I found myself wishing I could a/b test some parts of my book. I wanted objective data that told me whether people actually liked what I was writing
  • #12 So I set out to answer this question: can data tell me when a story is good? We tested 50 best-selling young adult novels
  • #14 And our hypothesis was that this 35% ceiling was primarily driven by a UI/UX problem Reading long-form fiction on your phone kind of sucks - We wanted to see if we could come up with a format innovation, that made reading fiction a more native experience on mobile - And therefore, make reading fiction more engaging
  • #16 I’ll discuss how we tested a sci-fi story named Discovery, about a guy in space. First version had 39% completion I noticed huge drop-off in first 20 messages, so I edited beginning of story to make it clearer, punchier I re-tested the story, and completion jumped to 80%
  • #18  Here I’ll make the point that, isn’t art ultimately made to get a particular reaction out of your audience / deliver a particular experience to your audience? and wouldn’t it be helpful to get objective metrics on whether your art is actually accomplishing that? as an artist myself (i’m a singer and an author), I believe data frees to be more creative if Hollywood used data, maybe it would break them of their increasing reliance on awful sequels, and get them back to creating art that truly has the potential to move the masses, and shape how generations view the future