26 Mar 2025
Dear Analyst #134: How Stephen Follows used AI to decode the Oscars and film data to debunk the film industry
Ever wondered if Hollywood's magic is just smoke and mirrors, or if there's actually some cold, hard data behind those blockbuster hits? I sat down with Stephen Follows (https://www.linkedin.com/in/stephenfollows/), the film data analyst who's basically Indiana Jones, but instead of chasing ancient artifacts, he's digging up the buried truths of the movie industry. We're talking about everything from Oscar speech patterns to why producer experience might be as useful as a screen door on a submarine. If you've ever wondered if the film industry runs on spreadsheets or gut feelings, you're in for a treat. Stephen's journey is a fascinating blend of film school ambitions, Guinness World Records projects, and a relentless pursuit of the hidden patterns behind Hollywood's stories.
From Film School to Data Nerd: A Journey of Discovery
Stephen's career path took an unexpected turn from aspiring filmmaker to data aficionado. He candidly shared how he felt a lack of intellectual engagement during his early film school days, a stark contrast to the stimulating world of data analysis.
I was just not using my brain.
This started his journey into more analytical pursuits. While others suggested a path in Philosophy, Politics, and Economics (PPE), with film as a side hobby, Stephen plunged directly into the film industry, establishing a production company and crafting advertisements, including work for non-profits. However, the allure of numbers and logical inquiry was undeniable.
He recounted how, two decades into his film career, he realized he was missing the critical thinking aspect. "Twenty years ago, as I got into film, I realized I wasn’t doing any thinking," he explained. "So, I started doing the numbers for films for my friends." This move eventually led to a collaboration with Guinness World Records. Stephen's initial expectation was that film studios would be heavily reliant on data and spreadsheets, but he quickly discovered that most industry decisions were surprisingly intuitive, rather than data-driven.
In the first 10 years of doing numbers work, I thought I’d see people at studios using spreadsheets and doing data. But most decisions are not data-driven.
Excel, AI, and the Oscars: A Technological Toolkit
Stephen's approach to data analysis is a blend of Excel's robust functionality and the burgeoning power of AI, a skill set he's honed through self-learning. Despite identifying as dyslexic, which posed challenges with traditional coding, he's leveraged his strong logical abilities to navigate these tools effectively. "I’m good at logic but dyslexic, so I can’t code," he clarified. Before AI became mainstream, he relied on online tutorials to enhance his Excel proficiency. His early exposure to AI, through a ChatGPT-assisted scriptwriting project, revealed its potential to bridge gaps in his data knowledge.
I realized I could use AI to fill in the gaps in my knowledge about data. I’ve increased my output 20 times. I can study all films and not just a subset of films.
His Oscars blog post titled Was Harvey Weinstein thanked more often than God at the Oscars? (https://stephenfollows.com/p/harvey-weinstein-thanked-more-than-god) analyzed a dataset we don't normally come across every day: Oscar speeches. It involved analyzing 2,000 acceptance speeches and highlighted the industry's focus on storytelling, even in its own narratives.
The film industry is all about telling stories about the stories they are telling.
