I have written a few times about how patents can be an excellent way to see which technologies will soon affect / improve / disrupt the industry.

This has led a number of readers to reach out and ask more about the technology in Ben Affleck’s AI company that would make Netflix want to buy it.

Today I’m going to go through the background, the patent and what it might mean for the film business in the near future.

If you want more context on today’s topics, here are a few recent articles which may help:

• What Netflix’s patents reveal about the future of watching movies

• What Amazon’s patent filings reveal about how filmmaking may change in the future

• How AI tools are already changing the jobs of film professionals today

• Netflix owns some pretty strange trademarks, and they reveal a lot about the company

The background - Netflix bought Ben Affleck’s AI filmmaking company

Netflix announced a few days ago that it has acquired InterPositive, an artificial-intelligence filmmaking company founded by Ben Affleck.

InterPositive was founded in 2022 and operated largely in “stealth mode” under the corporate entity Fin Bone LLC. The company says it’s developing tools intended to support filmmaking rather than replace it, focusing on production workflows such as relighting scenes, removing wires from stunt shots, correcting continuity issues and enhancing backgrounds in post-production.

The company itself is fairly small, with just 16 engineers, researchers, and creative staff, but in the world of AI start-ups, that doesn’t mean it couldn’t be cooking up industry-changing tech.

Netflix hasn’t revealed how much it paid, but we do know that all the assets, etc., and staff will now be a part of Netflix, and Affleck will remain as senior advisor.

The one specific thing we do know about is a patent.

The patent - US 12,511,904 B1

In December 2025, the United States Patent Office granted US Patent 12,511,904 B1, titled:

“Method, System, and Computer-Readable Medium for Training a Captioner Model to Generate Captions for Video Content by Analyzing and Recognizing Cinematic Elements.”

The inventor listed on the filing is Benjamin Geza Affleck-Boldt, and the assignee is InterPositive, LLC (i.e. the company Netflix has just acquired).

The patent describes a system that teaches an AI model to watch film footage and understand how the shot was filmed.

Most AI vision systems are trained to recognise objects or actions. They answer questions like:

• Is there a dog in this image?

• Is someone driving a car?

• Is a person walking across the street?

InterPositive is focusing on something slightly different, namely, how the scene was actually shot.

The model learns to recognise cinematic features such as:

• Focal length. This is which lens is being used and how it affects the image's perspective. Short focal lengths create wider views with more background visible, while longer focal lengths compress space and make subjects appear closer to the background.

• Camera movement tracks how the camera changes position during a shot, such as panning, tilting, or moving forward on a dolly. These movements alter how the viewer experiences the scene and can signal emphasis or narrative focus.

• Depth of field refers to how much of the image appears in focus at once. A shallow depth of field keeps the subject sharp while blurring the background, while a deep depth of field keeps both foreground and background clear.

• Camera angle refers to the camera's position relative to the subject. For example, a high angle looks down on a subject, while a low angle looks up, often changing how powerful or vulnerable a character appears.

• Occlusion occurs when one object partially blocks another from the camera’s view. This can reveal information about the spatial arrangement of objects in a scene.

• Parallax describes how objects at different distances appear to move at different speeds as the camera moves. Foreground elements shift more quickly than distant backgrounds, helping convey depth in the image.

In other words, it attempts to convert video footage into cinematography metadata.

The way the model looks at filmed footage mirrors how filmmakers in the real world think about their shots. So InterPositive’s claim that they are looking to use AI to enhance current filmmaking rather than destroy it is borne out in this patent.

I doubt that all but the most ardent anti-AI filmmakers would have a problem with a model that looks at real footage and gives you helpful information about it (assuming the information is correct, of course!).

The model - How does it work?

Their process started with a large collection of video clips, each paired with some form of description or metadata about what is happening in the shot. The system extracts individual frames from these videos and converts them into a format that a machine-learning model can process.

Each frame is then labelled with information describing the cinematic elements present in the shot. This includes the features I already mentioned, such as focal length, camera movement, framing style, and distance from the subject.

Read more

But over the past few years, that common starting point has become ever harder to find. Even for major releases, the figure is often missing or inconsistent.

So I decided to measure how bad the problem has become. And it’s pretty bad.

What’s the issue?

I’m going to go into more detail on budgets, available data, and the root causes, but I wanted to start with a single chart that frames the topic at hand.

I collected data on 62,298 feature films released commercially since 2000, and I could only find budget figures for 10.7%.

That’s not great, but it’s the trend which is more shocking. If we narrow it to films made in the past five years, that figure drops to just 3.3%

Let’s dig deeper and see what we can learn about the topic and why this is happening.

What even is a production budget, anyway?

When people talk about a film’s budget, they usually mean its production budget. This is the amount spent to make the film, not to distribute, release, or promote it. It covers everything from cast and crew salaries to sets, equipment, and post-production.

The industry term is the “Negative Cost”, as it is the cost of producing the first version (in the pre-digital age, the first celluloid negative) of the movie.

So far, so straightforward. But complications can arise from;

1. Budgets can change dramatically once a film is underway. Some projects are re-edited or reshot after being picked up by a studio. Napoleon Dynamite was acquired by Fox Searchlight, which spent additional money preparing it for release. Paranormal Activity was famously described as “shot for $15k”, but cost over $200k in post-production to be ready for a commercial release.

2. Not every cost was paid for in cash. Cast members might take smaller upfront fees in exchange for profit participation, lowering the reported budget. Producers might own locations or equipment, reducing apparent spending. Some productions receive significant in-kind support, such as the military assistance often provided to action films, or free products through placement deals. Reusing sets or footage also complicates matters (it felt like a third of Gladiator II was footage from Gladiator).

3. Defining tax rebates as either a refund or an income stream. Many countries offer financial incentives to attract film production. If a movie spends $100 million and receives a $25 million rebate, some will record the budget cost as $100 million, while others will report it as $75 million. Both can be technically correct, and there’s no universal standard.

4. The changing nature of film production. Every element of production has shifted in cost. Location shooting and labour have become more expensive, while digital cameras, editing tools, and visual effects have become cheaper. A $10 million movie today looks very different from a $10 million movie film twenty years ago. Numbers may appear the same on paper, but what they buy in practice has changed completely.

5. Currency exchange and inflation. Currency fluctuations and inflation distort comparisons between eras and countries. A British film costing £10 million in 2005 represents a different scale of spending when converted to current US dollars.

So while it’s pleasing to have a particular numerical value for a movie’s budget, it’s often far more complicated in the real world.

Lies, damned lies, and movie budgets

Thus far, I have assumed honesty on the part of all parties. This is an oversight which has to be redressed. When it comes to film budgets, “truth” is a flexible concept. Even when numbers are shared publicly, they rarely tell the full story.

In the past, there have been a couple of times I have had access to real-world data with which to test the accuracy of publicly-delivered figures., In both cases, I found the public figures to be less than reliable.

My study of 29 Hollywood blockbusters revealed that the average $100m+ Hollywood blockbuster actually costs $19 million more than is stated publicly (i.e. 12.5% more).

Similarly, when looking at UK indie films, only 70% of claimed budget figures were even broadly accurate.

From the outside, it’s hard to be sure when numbers are wrong. Maybe Wikipedia is right and Starship Troopers 3: Marauder really DID cost $20 million, meaning it had the same resources as Blink Twice, Gold, Uncut Gems, and the four-time Oscar-winning All Quiet on the Western Front.

Producers may inflate figures to make their film seem more substantial or to justify a high price when selling it. Other times, the reverse is true, with filmmakers understating costs to appear resourceful or an underdog.

In an industry of storytellers, it is perhaps to be expected that most public budget figures are part fact, part performance.

For example, here is a chronology of how much the low-budget sci-fi Monsters (2010) is reported to have cost:

• “$15,000” – May 2010 (/Film).

• “Less than $100,000” – October 2010 (The Observer)

• “Less than $100,000” – October 2010 (TIME)

• “Way under $500,000” – October 2010 (MSN Movies)

• “Around half a million dollars” – November 2010 (The Guardian)

• “Almost no money” – November 2010 (Ain’t It Cool News)

• “£1 - < £2 m“ – BFI budget data given to me in 2014.

• “Around half a million dollars” – January 2024 (SyFy Wire)

• “Under £500,000” – March 2024 (Medium), which is equivalent to under $740,000.

So while the current Box Office Mojo figure of “$500,000” is pleasingly precise and simple, it doesn’t really look all that good when you dig into the reporting of the film.

(I’m only using Monsters as one example - over the years, I have found many films where the public assumes that what a film costs does not stand up to even a little research).

Where does movie budget data come from?

There are only a few places where production budgets can originate.

At one end are researchers and trade journalists, who track down and verify figures from reliable sources. Credible industry databases, such as The Numbers, will conduct their own research and look for figures they can confidently support.

At the other end are public forums such as Reddit or Wikipedia, where anyone can add or change a number based on what they have read elsewhere or assume.

Between the two are trade outlets, interviews, leaked memos, and the occasional press kit.

This co-mingling of sources with varying reliability can yield quite different figures for the same movie.

• Youth Without Youth (2007) is listed as costing $1 million on Wikipedia, but $5 million on Box Office Mojo.

• Wikipedia helpfully lists the budget of Bigger, Fatter Liar (2017) as “$3–7 million“, citing WhatsFilming.ca.

• Deadline reports the budget of the 2019 Downton Abbey movie as $20 million, whereas The Wrap says it cost just $13 million.

What’s happening to data coverage?

The coverage gap between reliable and unreliable data is not as wide as you might initially assume. The Numbers makes much of its information freely available online, and anyone can subscribe at a low cost for deeper access. Wikipedia, for all its flaws, is built from publicly accessible sources. But both reflect the same shrinking pool of information.

So let’s focus on studio films for a moment - i.e. those most often studied and for which we would expect the most complete data.

When researching the topic, I noted the reliability of the budget data source. If it came from what I would regard as a “professional source”, I flagged the figure as reliable (notwithstanding all the issues we’ve already discussed on a single budget figure).

Both types of budget data have seen steep declines in recent years.

Wikipedia’s data coverage has a fascinating yet baffling pattern. Until around the mid-2010s, about a fifth of movies included budgetary information of some kind. But that cratered around 2013 and has only started to rebound in recent years.

Of the 27,766 films I found on Wikipedia (2000-25), only 16.4% listed some budget information.

Why are budget figures getting harder to find?

One of the biggest is the changing nature of the film press. The number of journalists covering the business side of the industry has dropped sharply, and those who remain are stretched thin across fewer outlets. With less time and space for investigative reporting, basic questions (like how much a film cost) often go unasked.

Even when they are asked, the resulting stories are less likely to include concrete numbers or to challenge the official line. This thinning of the trade press removes one of the main routes by which financial details used to reach the public record.

Alongside that, studios have become more guarded. In the past, budgets were routinely mentioned in press materials or interviews, whereas today such disclosures are much rarer.

Streamers, meanwhile, have set new norms around secrecy, controlling data at every stage of production and release. They have no incentive to publicise costs, particularly when their success isn’t measured at the box office.

At the other end of the market, the sharp rise in small-scale and independent productions adds to the opacity. These films are often made quickly, outside formal systems, and with little paperwork. They’re less likely to be tracked by trade journalists or data services and more prone to fuzzy or inflated figures.

Finally, the shift toward direct-to-streaming releases, which has severed the traditional chain linking studios, distributors, and the press, and it’s easy to see why reliable budget data is drying up.

The result is that, despite more films being made than ever before, we know less about what they cost.

In preparing this research, I spoke to a number of industry folk in research and the press. Most didn’t want to go on the record but Bruce Nash at The Numbers said:

The streaming services very rarely report budgets, which obviously brings the average down, but also might be giving the studios and others some cover for not reporting budgets any more. Perhaps they felt obliged in the past, but feel less pressure now?

Part of the reason for the reduction may also be that there are fewer journalists covering the industry, and those that remain work for only a handful of outlets. So producers/directors etc. may just be getting asked the question less frequently, either because they have fewer conversations with journalists, or because the journalists’ employers want to remain “chummy” with the studios by avoiding difficult questions.

Where does this leave us?

A film’s budget instinctively feels like a simple datapoint that we should be able to deduce. But we have seen that they are often inconsistent, incomplete, or deliberately shaped for effect.

More worryingly, far fewer figures are making it into the public domain.

Without reliable data, it becomes harder to track how the industry operates, what kinds of films are made efficiently, or how production practices are changing. It limits research into profitability, transparency, and trends that shape creative decisions.

For studios, secrecy protects negotiation power and brand image. For filmmakers, it shields them from awkward comparisons. For analysts, investors, and audiences, it leaves a growing blind spot.

Notes

For today’s analysis, I had a broad definition of a “studio film”, which included any feature film released by Amazon, Apple, Disney, Lionsgate, Netflix, Paramount, Sony, Universal, or Warner Bros.

There is a hard truth about cinema-going that many in the exhibition sector don’t seem to want to admit.

When you ask the public why they don’t go to the cinema more often, one reason always appears on top. But in the many conversations I’ve had with industry insiders over the years I’ve heard repeated denials or downplaying of this factor.

Today I’m going to look what the data says, detail the industry's strange resistance to reality, and share an interview I conducted with Tim Richards, Founder and CEO of Vue International, who has completely changed his mind on this topic.

What is the biggest barrier to cinema-going?

So what is this common barrier? Is it the screening technology, the seats, the on-site facilities, the cleanliness, the movies on offer?

Nope. It’s price.

To illustrate my point, I’ve gathered survey data from a range of countries that asked people directly about what stops them from going to the cinema more often.

Let’s start in Canada where a 2023 Telefilm survey of 2,000 Canadian adults found that 72% of surveyed Canadians cite the cost of tickets as the main reason they don’t go more often, far ahead of any other factor.

In France, a recent Ipsos–CESI study of 1,000 French adults found that 67% respondents said a rise in ticket prices would make them go to the cinema less often. At the same time, 70% said lower ticket prices would encourage them to go more frequently.

In March 2023, the BFI surveyed 5,000 UK adults for their “The Wider World of Film” report and found that the single biggest brake on cinema-going is cost. 34% of adults named the cost of a single ticket as the main reason they went less often.

If we zoom out to take a global view, the picture remains the same. A YouGov study carried out in August 2022 asked people whether they agreed with the statement “Going to the cinema is too expensive.” Half of respondents agreed, with only a minority disagreeing.

Modelling suggests that lowering the cost could have a big effect on attendance overall. A study of 5,000 Italians in 2022 by the Italian cinema association ANEC, modelled different post-pandemic recovery scenarios. They found that when tickets are priced at €8, attendance could rise by as much as 69%. At €12 the projected increase dropped to just 5% and two-thirds of regular cinema-goers disappeared.

What is the cinema industry’s reaction to this?

By now you have seen a pretty overwhelming case that the public say that price is a major barrier to increased cinema attendance. And yet if you were to speak to people within the exhibition sector of the film business you would get a very different view.

In 2014, Tim Richards, CEO of Vue International, told the Screen Film Summit that:

Any dialogue that helps drive attendance is a good thing but pricing is not it: absolutely not it.

We’re in an industry in which people will pay virtually anything to see a film they really want to see. If there is a film they do not want to see they will not go even if it is free.

Pricing is typically third or fourth on the scale [of concerns]. Parking is number one.

And this is a fair common view. I have heard it expressed countless times over the industry by senior people close to the exhibition sector and within major studios. They either dispute the data or just don’t agree that the cost of a cinema ticket is all that important.

I think this comes down to different mental models of why people choose to go to the cinema in the first place.

• Those who work in film place most (maybe all?) of the decision-making process on the movie itself. I.e. who’s in it, what is it about, the reviews, the marketing, etc.

• But for the public at large it’s a lot more to do with their personal situation. i.e. How much will it cost, who will come with me, can I get a babysitter, is it a hassle to get to, etc.

Therefore, when a Studio Exec sees a $50 million box office haul for their new movie they imagine that that a 10% price increase would have produced a $55 million gross, and feel that the cinemas left money on the table.

I reached out to Tim Richards to see if his view was the same. I was surprised to learn that in the past decade he has performed a 180 degree turn.

He started by laying out Vue’s perspective at the time he spoke at the Summit in 2014

For the first 15 or 20 years of the company’s life, we were premium pricers. We had the best cinemas, the youngest estate, and we weren’t shy about pricing.

Around 2017 he started to worry that the industry’s price growth was unsustainable.

I remember looking at a very simple graph of CPI inflation indexes across all our markets in Europe and then looking at ticket prices. You could see the acceleration. You can do that for a couple of years, but you can’t keep doing it every year.

Read more

Over the last few months, I have conducted a large-scale study to test whether films that attract more audience discussion of woke-adjacent themes tend to perform differently at the box office.

And the results are in.

A quick note on what this is and isn’t

When I first looked into this topic, I examined whether we could track what I call woke-adjacent themes in films. That work attracted a lot of commentary, with people across the political spectrum citing the results to support their own positions. In some cases, my analysis was co-opted into arguments I was not making.

So let me start by saying a few things:

1. This is not a personal essay nor a political statement. It is a study of measurable correlations between how people talk about films and how those films perform.

2. We’re tracking subjective thoughts in a subjective medium, centred around a subjective notion (i.e. “wokeness”). That said, the patterns I am highlighting are strong and consistent.

3. Movies are both art and business. There are many reasons why films are made and many ways they are judged, with profit being only one of them. Studios are multinational corporations with shareholder obligations, meaning that financial performance is an unavoidable part of their decision-making. Independent films sit in a different space as many investors will be motivated by impact or personal passion as much as by financial return.

What is “woke” and how do we track it?

Woke is a shifting and contested term, used as praise by some and as criticism by others.

To make it measurable, I broke it down into “woke-adjacent” themes that consistently appear in audience conversations. They largely fall into the following five categories:

1. Representation & Diversity (casting, character focus)

2. Identity Politics (gender, race, sexuality)

3. Political & Social Messaging (justice, inclusion, activism)

4. Canon & Continuity Changes (gender / race swaps, rewriting lore)

5. Cultural Tone (modern language, virtue signalling, didactic framing)

Instead of labelling films as woke or not, I tracked when these themes were present in how people talked about them.

It’s a complex spectrum of perception, across a number of intersecting themes. For example, below are twelve movies from recent years which have, to varying degrees, been brought into the woke conversation.

When we consider the accusations made against each, we can see that there is a whole landscape of topics.

If you want a more detailed breakdown of this process then head over to my previous article on the topic Can we measure if movies have become "too woke"?

To discover the effect of these themes I analysed more than 4 million user reviews across 10,524 films and tracked the correlation with film profitability.

I have put details of methodology at the end of the article so that we can get straight into what I discovered.

Is ‘Go Woke, Go Broke’ true for movies?

So is “Go Woke, Go Broke” true? The short answer is no.

Movies are too varied for such a simple phrase to capture the reality. Some films with woke-adjacent themes perform very well, while others underperform, and the same is true of films without those themes.

But it’s not without grains of truth.

What emerges from the data is an effect that depends heavily on context. Budget level and genre shape how risky certain choices are. At the very top end, where hundreds of millions are invested and returns are scrutinised worldwide, overt or clumsy political messaging tends to be punished. In other spaces, particularly horror, music, or sports, identity-driven themes can connect strongly with audiences.

The slogan makes for a neat headline, but the reality is more complicated.

Let’s dig into it.

Where woke-adjacent themes hurt profitability

Let’s start with the strongest negative correlations - i.e. if you solely care about making money with movies, then you should steer clear of these woke-adjacent ideas.

1. Heavy-handed moral lecturing in big-budget films

The larger the budget, the greater the financial risk. Audiences tend to punish heavy-handed political or cultural messaging in expensive blockbusters, where broad appeal is critical.

When messaging feels bolted on or too explicit, audiences push back and profitability suffers. Don’t Look Up (2021) was criticised as “without trace of subtlety” while The Woman King (2022) was accused in reviews of trampling historical accuracy for the sake of its empowerment message. Whether or not those views are fair, once audiences feel lectured, the film pays a price.

2. Breaking continuity in established franchises

Changing the perceived “contract” of a long-running character or series raises the risk of backlash. Small adjustments may be accepted if they feel authentic, but multiple or fundamental changes multiply the danger.

Franchise audiences expect consistency. Changing a defining element can feel like betrayal, especially if multiple changes stack up at once. The Last Jedi (2017) provoked backlash from fans who said characters were “unrecognisable,” and Ghostbusters (2016) faced fierce criticism for its all-female team.

By contrast, spinoffs like Bumblebee or reboots like Scream (2022) worked because they refreshed without discarding the essence. The data shows continuity breaks are one of the riskiest moves for established IP.

This is probably most acute now when it comes to the conversation about who should be the next James Bond.

We all share a Platonic idea of what the character of Bond is. He’s British, suave, authoritative, with a charming masculinity. The actors who have portrayed him thus far have also shared many qualities, namely they have been white, middle-aged, conventionally attractive men with athletic builds in a sharp suit, and with a polished physical presence signalling both sophistication and danger.

The data suggests that the further a new Bond strays from the qualities audiences perceive as canonical, the more profitability risks increase. Even Craig’s “Bond who bleeds” was framed as a major reinvention despite being a subtle shift, showing how sensitive the role is to change.

3. Modern terminology in immersive worlds

Audiences expect dialogue to feel consistent with the world on screen. Contemporary slang in a historical or fantastical setting often breaks immersion and gets called out as pandering.

Persuasion (2022) set in Regency England drew ridicule for lines like “the thing about me is, I’m an empath” while The Last Duel (2021) placed phrases like “this is just science” in 13th-century France. Viewers felt these choices clashed with the setting and undermined authenticity.

4. Perceived inauthenticity in historical or biographical films

Films rooted in history carry an expectation of accuracy. Even small liberties can be seen as ideological rewriting, damaging trust. Mary Queen of Scots (2018) faced pushback over inaccuracies, and Selma (2014) was attacked for downplaying Lyndon B. Johnson’s role in the Voting Rights Act.

The key issue is not always what is academically true, but whether audiences feel they are being handed a version of events that lacks authenticity.

(This connects to what linguists call the ‘Tiffany problem’ which is when something that really did exist in history feels false to modern audiences. The name comes from the fact that Tiffany was a common medieval girl’s name, but today people see it as anachronistic, so it gets rejected as inauthentic even when accurate).

Where woke-adjacent themes help profitability

So far we’ve seen four ways woke-adjacent themes are negatively correlated with profitability. But there are many cases of the inverse - where progressive themes were positively linked to commercial outcomes.

1. Horror films thrive on reinvention

Horror has the youngest adult audience of any genre, and those viewers are generally more receptive to contemporary themes.

Also, the genre itself is built on confronting taboos and reframing cultural fears, so layering in identity-driven stories can feel natural. Candyman (2021) used gentrification and Black identity to refresh the mythology, while Evil Dead Rise (2023) shifted the setting to a city high-rise.

These choices strengthened rather than weakened audience engagement, and across low-budget horror overall, woke-adjacent themes correlated positively with performance.

2. Thrillers benefit from shifting perspective or canon

While continuity risks are high in major franchises, mystery and thriller films often gain from a twist in framing.

• The Invisible Man (2020) retold a classic story from the victim’s perspective

• Glass Onion (2022) transplanted its detective into the world of tech elites

• Enola Holmes (2020) centred the tale on Sherlock’s younger sister.

These changes gave audiences something fresh while still delivering the core genre experience. Reinvention in thrillers is rewarded rather than punished, particularly with younger audiences.

3. Sports films welcome identity narratives

The structure of sports stories (i.e. struggle, perseverance, and acceptance) aligns well with themes of identity.

Characters trying to join a team or prove themselves on the field mirror broader stories of belonging. Gender-swapped or underdog narratives work here because the genre expects uphill battles and celebrates resilience.

A player lifted by the crowd at the end is both winning the match and being accepted for who they are. This makes sports films unusually fertile ground for identity-driven storytelling.

4. Music films support moral and identity themes

Stories built around music often blur realism and fantasy, allowing overt themes to feel more authentic. Songs can carry emotional or moral weight more directly than dialogue, which helps audiences accept identity-driven stories.

• CODA (2021) explored both deaf family dynamics and a young woman’s musical ambition

• Rocketman (2019) wove Elton John’s sexuality and struggles with addiction into a musical biopic

• A Star Is Born (2018) used music to heighten themes of fame and self-destruction.

In this space, moral or identity themes enhance the experience rather than distracting from it.

The overall conclusion

“Go Woke, Go Broke” does not hold up as a general rule. Films are too diverse, and audience responses too varied, for a cute slogan to explain outcomes across the whole industry.

What the data shows is that context matters, with the same theme that strengthens one film can weaken another, depending on budget, genre, and execution.

Risks multiply when several sensitive factors are combined, and rewards appear when themes align naturally with the purpose of the genre.

And a good story well told will overcome most reflexive objections. For example, when considering how much Amazon can stray from the Bond we know and love, we should consider the casting to Renée Zellweger as British icon Bridget Jones in 2000.

Initial news of the Texan’s casting caused outrage among the Bridget Jones fanbase. But thanks to Zellweger’s dedication to the role (i.e. gaining weight for the role, changing her look and working undercover in the UK as research), along with her talent (the spot-on British accent, and her on-screen performance) the audience were won over.

The takeaway is not that filmmakers should avoid or embrace particular ideas, but that success comes from handling them with authenticity and awareness of context. Big-budget projects carry the greatest scrutiny, while smaller genres can often integrate identity-driven stories more effectively.

Notes

I built a dataset of over four million user reviews and comments across more than 50,000 feature films. Using language similarity rather than keyword matching, I tracked how closely each review aligned with a set of dummy texts I had written to capture different cultural or political themes. These included ideas such as representation, inclusivity, political messaging, canon changes, and overt signalling. If a film’s reviews clustered near a dummy text, that film was counted as having audience conversation about that theme.

Box office figures and budget estimates were drawn from publicly available sources, then used to calculate broad profitability measures and sense-checked by inside figures I have on real projects.

The analysis focuses on correlation. It asks whether films that audiences discussed in connection with certain themes tended to perform differently at the box office. It does not claim that those themes caused the outcome, only that they are associated with it.

The categories used here are not fixed definitions of “woke.” They are recurring ideas that audiences themselves bring up when talking about films. The purpose is not to decide which films are woke or not, but to measure patterns in audience response at scale.

It is also important to note that “woke” itself is a contested and shifting term. What counts as woke-adjacent in one person’s eyes may not in another’s. The categories here reflect recurring themes that audiences themselves bring up, not a fixed label imposed from above.

The data highlights patterns worth paying attention to, but it cannot tell us what films should or should not be. That remains a creative choice.

When I have highlighted an example I am not saying they were unprofitable, nor that they were the worst offender of the topic. It’s just a handy way of bringing to life the ideas. The findings are about overall correlations, not pointing to the faults of any one film.

This prompted a few readers to get in touch and ask if the same domination (i.e. four companies representing half the entire field) is present for directors.

To answer that, I gathered data on the top 1,000 directors and linked them to their current agencies. There is more detail on my methodology in the Notes section at the end of the article.

It’s lonely at the top

For actors, Creative Artists Agency (CAA) was at the top, with almost a quarter of top actors on their books. The effect is even more pronounced among the most elite directors, as CAA represents 19 of the top 50 directing names (38%).

Coming in behind are William Morris Endeavor (WME) at 12%, United Talent Agency (UTA) at 10%, and Gersh with 8%.

It won’t come as a shock to any industry-watchers that CAA is the top dog. CAA was founded in 1975 by five agents who left William Morris, and went on to restructure how Hollywood works. It introduced packaging, whereby the agency brings together a director, writer, and lead actor into a fully formed project and sells it to a studio.

This gave CAA enormous influence, allowing them to control not just talent deals but entire films. For directors, this means that CAA became a development engine, attaching financing, talent, and distribution from inside its own ecosystem. It remains the most powerful agency in the business, especially for top-tier directors.

Second placer WME, while also being a major player, has a slightly different profile to CAA. It was formed in 2009 when Endeavor acquired the legacy William Morris Agency, and since then has pursued a strategy of vertical integration. Through Endeavor, it owns major global properties like UFC, WWE, and IMG. The agency is structured more like a global entertainment company than a traditional rep firm.

When we widen the field, more come out to play

Zooming out from the top 50 to the top 1,000 gives us a fuller picture of how agencies operate across the spectrum of director visibility.

CAA is still top, but their share drops to 16.2%. The next two spots are taken up by WME (10.3%) and UTA (9.8%).

Gersh is smaller than the ‘Big Three’, but punches above its weight in director representation, especially in the independent film space. Known for its presence at Sundance, TIFF, and Berlinale, it represents many breakout voices and first-time filmmakers.

Independent Talent is the largest agency in the UK and a key player in British film and television. It represents a wide range of directors, from commercial studio names to public-service auteurs. Unlike US agencies, it doesn’t package aggressively or operate across media verticals. Instead, it offers deep ties to UK broadcasters, funders like BFI and Film4, and international co-production partners.

Casarotto Ramsay is probably the most unique of the agencies I’ve name-checked so far. It’s selective, independent, partner-owned, and focused solely on creative representation. It doesn’t represent actors, doesn’t take packaging fees, and rarely chases commercial scale. Instead, its directors are positioned for prestige work, long-term development deals, and international recognition.

The ‘Big Three’ are top heavy

We’ve already seen how CAA dominates the very top tier but loses market share as we widen the pool of directors. This sets it apart from its neighbours, who have the opposite pattern.

WME and UTA may be less concentrated at the very top, but hold their ground more consistently as the dataset expands.

The result is that director representation looks more top-heavy than actor representation. The biggest agencies secure the most visible names, but as you move beyond the top tier the field opens up to a broader set of players.

How director-agent relationships differ from actors

Before we go, I wanted to quickly note that directors want slightly different things from their agent than actors do.

Representation for directors tends to be a slower, more strategic relationship. Instead of job-getting, the agent’s role is closer to career architecture. This difference shapes how agencies build their rosters and how power functions inside the system.

Where an actor might bounce between studio, streaming, or indie films year to year, directors often wait years between projects. In that time, it’s the agent who keeps momentum going, by positioning scripts, attaching cast, building financing, or shaping festival strategy.

Consequently, while actors often switch agencies after a single breakout role, directors rarely do.

Notes

The cohorts of actors I studied were based on IMDb’s STARmeter, which ranks actors by a combination of public interest signals, including page views and media coverage. It changes every week and I looked at it during the week commencing 25th August 2025. I’m sure if I re-ran this another time, the order of the people would shift around. That said, this is unlikely to change the overall picture in any meaningful way.

Actors move agents and agencies. This means both that the picture may change in the future, and also that just because a successful actor is currently with a certain agency it doesn’t mean that it was that agency that got them there.

Each director’s agency was taken from their IMDb Pro profile, focusing only on the first-listed agent. This is usually their main representative for film and television work. Some actors have multiple agents across different departments, territories or media. Some may also have managers, lawyers or publicists listed. These were excluded.

In just a few cases, no agent was listed. Percentages are calculated based only on those actors who actually had a talent agent named at the time I looked at it.

Although UTA acquired a controlling stake in Curtis Brown in 2022, I have kept Curtis Brown listed separately in this analysis. This reflects how directors themselves present their representation on IMDbPro, where Curtis Brown is typically listed as a distinct agency rather than as part of UTA. While the two agencies may collaborate internally, this approach preserves the way directors identify their primary representation.

It’s worth emphasising that this is not a measure of quality. An agency may have fewer clients because it is selective, or because it is smaller. An agent with many famous clients is not automatically more effective than one with a small, focused roster. Today’s research is tracking the representation distribution, not an endorsement of any one approach.

Finally, while directors within the same agency benefit from shared infrastructure and branding, they may have different personal agents, priorities or strategies. This analysis treats agency affiliation as a meaningful organisational link, but not as evidence of uniform treatment or outcomes.

And this is never more true than among Hollywood agents.

I wanted to check in with the situation as it is today. Which agencies represent the greatest number of high-profile actors?

Which Hollywood agencies represent the biggest actors?

In most cases, when actors talk about "my agent", they mean the person who helps them find work, negotiates deals, and manages the business side of their acting career.

Those agents work for “an agency”, which might be just them or anywhere up to a massive corporation. According to LeadIQ, WME employs around 5,000 people, CAA has 4,400 employees, and UTA 2,300.

Today I’m looking at agencies, not individual agents (research for another time, maybe!).

To get a sense of overall agency dominance, I researched the top 1,000 male actors and top 1,000 female actors on IMDb’s STARmeter in August 2025. There’s more detail on methodology in the Notes section at the end.

The chart below shows which agencies are credited with representing the largest share of these 2,000 actors.

Creative Artists Agency (CAA) is out in front, representing 23.7% of top actors who publicly list their agency. They are followed by United Talent Agency (UTA) at 11.8%, Gersh at 10.4%, and William Morris Endeavor (WME) on 10.0%.

So although I tracked 156 agencies across the dataset, 56.0% of the market share went to just four companies.

Some agencies prioritise the most famous actors

If you regularly read the film trade press then you may be surprised to see WME in fourth place, given that it’s often regarded as part of the “Big Three”.

This is, in part, a consequence of my decision to look at the top 2,000 actors. If we narrow it down to the top 100 (i.e. the 50 most famous male actors and the 50 most famous female actors), then the triumvirate is restored.

This reveals something interesting about the differing business strategies of each agency (well, the strategies that the market and their wallets will allow them to execute).

Some agencies seem to prioritise elite visibility, while others focus more on volume or emerging talent.

We can see this in action when we split each agency’s share across broader tiers: Top 50, Top 100, Top 250, Top 500, and Top 1,000 actors.

Read more

What’s very frustrating is that filmmakers keep making more seminal films, and film scholars keep adding to ‘the canon’. It’s quite unfair as I’m still playing catch up.

By ‘the canon’ I mean the informal but widely acknowledged body of films considered essential viewing by critics, historians, educators, and institutions.

With new films added every few years, it stands to reason that some older ones might lose ground. This isn’t a strict one-in-one-out system, but certain forces do make films fade from prominence:

1. Fashion show. Canons reflect the aesthetic priorities of their time. As collective taste evolves (especially around pacing, tone, or subject matter) older styles can fall out of favour.

2. Context matters, and shifts. We’re always learning more about the circumstances in which films were made. That includes the politics of their production, the personal behaviour of filmmakers, and the broader ideologies embedded in the work (e.g. Birth of a Nation and Triumph of the Will, plus the works of Woody Allen and Roman Polanski).

3. Who’s asking? As more critics and curators from outside the historically dominant regions contribute to these lists, the canon expands. That expansion means more films from underrepresented regions and communities, and less space for overrepresented ones.

4. Gotta see it to appreciate it. Some films stay in conversation because they’re visible, while others fall away due to limited availability. Restorations, repertory screenings, and streaming revivals have pushed films like Yi Yi and Sátántangó, into wider awareness.

5. It’s all academic. Academic syllabi shape generations of critics and voters, and so when their courses shift their case studies, that affects what’s people see and appreciate. For example, In the Mood for Love has become a staple in courses on romantic drama and visual design.

6. That’s not ok (now). Some films age poorly due to outdated or offensive content. The dog in The Dam Busters being named the N-word, or Mary Poppins having its rating raised over the use of “Hottentot,” are reminders that acceptability shifts.

7. Lists are literally zero-sum. Top 10s and 100s are, by definition, zero-sum. There’s limited space, so even a small shift in attention (such as a retrospective, anniversary, or critical reappraisal) can nudge films on or off the list.

This got me thinking - which films are falling out of fashion within the debate of “the best films ever made”?

Seeing and hearing Sight & Sound’s poll

I decided to focus on Sight & Sound’s once-in-a-decade poll which has been running since 1952.

It is the longest running, regularly repeated poll of critics and filmmakers, and offers consistent ranks and vote counts across eight cycles from 1952 to 2022. This means while it’s not the official arbiter of ‘the canon’, it is a reliable proxy that shows change over time.

The poll’s electorate has grown and diversified over the years, with just 63 critics being quizzed in 1952, to 122 in 1982, 101 in 1992, 145 in 2002, then 846 in 2012 and 1,639 in 2022. The most recent cohorts include programmers, curators, archivists and academics.

The most recent poll lists the ‘Greatest Films of All Time’ as (in order):

1. Jeanne Dielman, 23, quai du Commerce, 1080 Bruxelles (1975, Chantal Akerman, Belgium/France).

2. Vertigo (1958, Alfred Hitchcock, USA).

3. Citizen Kane (1941, Orson Welles, USA).

4. Tokyo Story (1953, Yasujirō Ozu, Japan).

5. In the Mood for Love (2000, Wong Kar Wai, Hong Kong/France).

6. 2001: A Space Odyssey (1968, Stanley Kubrick, USA/United Kingdom).

7. Beau travail (1998, Claire Denis, France).

8. Mulholland Drive (2001, David Lynch, France/USA).

9. Man with a Movie Camera (1929, Dziga Vertov, Ukrainian SSR/USSR).

10. Singin’ in the Rain (1951, Gene Kelly, Stanley Donen, USA).

What are some recent additions to ‘the canon’ of great cinema?

So many films are produced each year, and so much of the public noise around movies is on the popcorn fare, that it can be hard to remember which new works of great cinematic art have also been released.

Here’s a rundown of a few films released in the past few decades which have entered Sight & Sound’s top 100 list:

• Parasite (2019, Bong Joon Ho, South Korea). Best rank 90.

• Portrait of a Lady on Fire (2019, Céline Sciamma, France). Best rank 30.

• Get Out (2017, Jordan Peele, United States). Best rank 95.

• Moonlight (2016, Barry Jenkins, United States). Best rank 60.

• Tropical Malady (2004, Apichatpong Weerasethakul, Thailand). Best rank 95.

• Mulholland Drive (2001, David Lynch, United States). Best rank 8.

• Spirited Away (2001, Hayao Miyazaki, Japan). Best rank 75.

• The Gleaners and I (2000, Agnès Varda, France). Best rank 67.

• In the Mood for Love (2000, Wong Kar Wai, Hong Kong). Best rank 5.

• Yi Yi (2000, Edward Yang, Taiwan). Best rank 62.

Recent additions have much more diverse origins, both the countries they were made in (for example, there are many more from Asia) and the profile of the filmmakers behind them (from Varda and Campion to Denis, Sciamma and Dash).

The genre and form of the films have also expanded. Animation (Spirited Away) sits next to essay‑documentary (The Gleaners and I), queer cinema (Beau travail, Portrait of a Lady on Fire), slow cinema (Sátántangó, Tropical Malady), horror (Get Out) and contemporary US indies centred on identity and class (Moonlight, Daughters of the Dust).

In with the old

It’s not just new films which have been recently added. Some older films which made their way into the list for the first time in 2022 included:

• News from Home (1976, Chantal Akerman). 2022 rank 52.

• Daughters of the Dust (1991, Julie Dash). 2022 rank 60.

• The Watermelon Woman (1996, Cheryl Dunye). 2022 rank 146.

• Le bonheur (1965, Agnès Varda). 2022 rank 153.

• Symbiopsychotaxiplasm: Take One (1968, William Greaves). 2022 rank 169.

• One Way or Another (1977, Sara Gómez). 2022 rank 196.

• Je tu il elle (1974, Chantal Akerman). 2022 rank 226.

• Oh, Sun (1970, Med Hondo). 2022 rank 244.

First‑time entries seem to arrive in waves. The 1962 and 1972 lists consolidated mid‑century European art cinema and post‑war classics. The 2002 poll pulled in a set of late‑20th‑century titles that had matured into consensus. The 2012 and 2022 editions added contemporary works that crossed the threshold surprisingly quickly (Parasite, Get Out, Moonlight) alongside slow-burn classics that had built a case over decades (In the Mood for Love, Mulholland Drive, Sátántangó).

Also, I spotted that Paul Verhoeven’s 1995 “classic” Showgirls received seven votes in 2022. This ranks it joint 405th, so outside of the official ‘Top 250’ but not absent.

Which classic films are on the wane?

To try to find the films falling out of fashion I created a ‘Forgotten Score’, based on the Sight and Sound data.

It took in the best ever rank the film has scored and its most recent standing. I narrowed down to films that had ever been in the top 100 but which were outside the top in the 2022 poll.

Below are the films, in order, which seem to be at most risk of being fired from the canon:

1. 2 or 3 Things I Know About Her (1967, Jean-Luc Godard, France). Best rank 41. 2022 rank 462.

2. Gertrud (1964, Carl Theodor Dreyer, Denmark). Best rank 25. 2022 rank 136.

3. L’Âge d’or (1930, Luis Buñuel, France). Best rank 24. 2022 rank 553.

4. The Travelling Players (1975, Theodoros Angelopoulos, Greece). Best rank 46. 2022 rank 405.

5. The Godfather: Part II (1974, Francis Ford Coppola, United States). Best rank 11. 2022 rank 104.

6. La Terra Trema (1948, Luchino Visconti, Italy). Best rank 9. 2022 rank 553.

7. Ivan the Terrible (1944, Sergei Eisenstein, Soviet Union). Best rank 6. 2022 rank 405.

8. Viridiana (1961, Luis Buñuel, Spain/Mexico). Best rank 7. 2022 rank 405.

9. Three Colors: Blue (1993, Krzysztof Kieślowski, France). Best rank 89. 2022 rank 462.

10. The Eclipse (1962, Michelangelo Antonioni, Italy). Best rank 73. 2022 rank 196.

11. Belle de jour (1967, Luis Buñuel, France). Best rank 54. 2022 rank 1652.

12. Diary of a Country Priest (1951, Robert Bresson, France). Best rank 47. 2022 rank 462.

13. Raging Bull (1980, Martin Scorsese, United States). Best rank 18. 2022 rank 129.

14. The Passenger (1975, Michelangelo Antonioni, Italy). Best rank 44. 2022 rank 292.

15. Nazarin (1959, Luis Buñuel, Mexico). Best rank 20. 2022 rank 1105.

16. The Grapes of Wrath (1940, John Ford, United States). Best rank 13. 2022 rank 668.

To be clear, I’m not saying anything negative about these films. They have not got worse, nor will they ever been forgotten. They are just the ones whose star seem to shine a little less brightly than in previous years.

Notes

The most recent data is nicely available on Sight & Sound’s website but the 20th century polls are slightly more opaque as they were only in the printed editions. However, Katie Donia (CUNY) has done God’s work in creating the “Complete Sight & Sound Greatest Films of All Time Database” by reading through those old editions.

Thank you to Isa for our conversation which led to this and to tipping me off about Katie’s work.

Namely, that there are some genres are best suited to smaller budgets and are rarely improved by scale.

I looked at the financial and critical performance of over 10,000 movies over the past half century to understand which types of movies both thrive on lower budgets, AND seem to struggle when more money is spent.

The four types which jumped out to me were:

1. Offbeat character comedy feature distinct voices and quirks but suffer of forced to soften or broaden their appeal.

2. Brainy science fiction uses tight resources to focus on clever concepts and bold ideas, leading to the intimacy and inventiveness that the audience crave.

3. Terrifying horror relies on the limitations low budgets forces upon filmmakers.

4. Intimate drama and romance find their strength in stripped-down settings and grounded performances.

Let’s look at each in turn and then draw wider lessons from the journey.

1. Offbeat character comedy

Offbeat comedy thrives on intimacy, specificity, and (perhaps seemingly ironically) a sense of realism.

The jokes that land hardest are usually rooted in idiosyncratic characters or sharply observed detail, both of which risk dilution with a bigger budget or star-heavy ensemble.

The more contained and peculiar the premise, the more a modest production value supports the comedy’s tone. In short, these films succeed when they resist the urge to broaden their appeal or add unnecessary gloss.

Because they don’t have to justify a big budget to investors, the filmmakers have less pressure to attract a mainstream audience. This frees them up to shoot for riskier, more personal humour, and let their artistic voice run riot.

Offbeat character comedy has proven itself at both micro and lower budgets. Napoleon Dynamite (reported budget of $400k) Clerks ($27k) both brokeout and launched careers. The approach works just as well at a slightly larger scale: The Big Sick ($5m) grossed over $56 million globally, and Booksmart (budgeted at $6m, grossed $25m worldwide), and a few years ago Palm Springs (shot for $5m) became one of Sundance’s prized acquisitions for a reported $22m.

Attempts to upscale these comedies often flatten what made them appealing. Studio-backed character comedies force in bigger effects, expensive locations, and cameos, but these elements rarely serve the core material. The sense of reality and the oddball social dynamics disappear in favour of production design and set-pieces.

When Hollywood spent big on comedies like Son of the Mask or The Adventures of Pluto Nash, the creative identity vanished and audiences stayed away.

Offbeat character comedies solve their biggest creative problem not with money, but with the freedom to stay strange. The lower the risk, the more producers and writers can let unusual voices and perspectives come through.

2. Brainy science fiction

Brainy science fiction is all about resourcefulness and intriguing concepts. Unlike effects-driven blockbusters, these films attract audiences who value ideas and narrative complexity above large-scale spectacle.

The smaller the budget, the more directors focus on contained settings, minimal casts, and clever writing to deliver their story. This approach often heightens the sense of possibility and makes even the most outlandish concepts feel credible.

Films like Primer ($7k), Coherence ($50k), and Moon ($5m) highlight why this works. With few locations and a focus on characters, they achieved cult status and critical acclaim despite minimal spending. District 9, while larger at $30m, still counts as modest for its genre and used its budget to create gritty realism and political allegory, showing that a strong concept can stretch every production penny.

Bigger budgets don’t necessarily help upscale brainy science fiction. Cloud Atlas and Annihilation struggled to recoup their investment despite ambitious scope and visual effects. Blade Runner 2049, though respected by critics, failed to match its towering budget with box office receipts.

Read more

While the core business is streaming, the trademark register shows Netflix experimenting with everything from memes to restaurants, live events, and product design.

Each trademark sits as a data point in Netflix’s strategy for growth and risk. The legal paperwork can be the earliest hint of where the company is heading.

I found over 300 active trademarks Netflix maintains, plus hundreds more that are dead, rejected, abandoned or tangentially-related. I have picked out eight which give us an insight in the streaming giant:

1. “Strong black lead”

2. “Qwikster“

3. “What a Joke”

4. “Netflix bites”

5. “Queue”

6. “Cheat”

7. “Easy”

8. “Netflix and chill”

1. “Strong Black Lead” - Trademarking diversity

Serial No.: 90875991

Status: Registered (as of April 2025)

Summary: Covers a wide array of goods and services (Classes 16, 18, 21, 25, 35, 38, and 41), including publishing, apparel, housewares, broadcasting, and event production. Netflix Studios, LLC owns the mark.

That’s right, Netflix has exclusive control over the phrase “strong black lead” for everything from print and publishing to digital media, apparel, events and product tie-ins.

At first that may feel like the worst kind of corporate land-grab of a social issue. But under the surface, Netflix has done the work.

Their Strong Black Lead initiative started with Black executives flagging the lack of targeted promotion for Netflix’s significant library of Black-led stories. From this, the company built a social media presence, original podcasts, and special campaigns. The public launch, timed with the 2018 BET Awards, featured a recreation of the historic “A Great Day in Harlem” photo, positioning Netflix’s Black creators as key to its identity.

Strong Black Lead became a core business strategy, supporting US subscriber growth at a time when domestic numbers were under pressure. Netflix gave the brand its own budget, developed original video and podcast content, and tied it to a pipeline for Black writers, directors and creative talent.

Netflix published internal research showing a correlation between Black creators involved behind the camera and improved representation on screen. The company spent over $100 million on creative equity funds and talent pipelines, including specialised bootcamps and partnerships with Black media educators.

2. “Qwikster” - Netflix’s biggest mistake

Serial No.: 85425822

Status: Abandoned/Dead

Summary: Filed in 2011 for computerised retail, streaming, video rental, entertainment, etc. (Classes 35, 38, 41). Application abandoned due to failure to respond to an Office Action.

In 2011, Netflix tried to spin off its DVD-by-mail service as a separate company called Qwikster. The plan was framed as forward-thinking self-disruption: give each business its own focus before streaming eclipsed DVDs entirely.

Internally, two divisions had already split and key DVD executives were leaving. Netflix’s leadership believed unbundling the services would allow streaming to grow faster, and the DVD side to run out its model on its own terms.

But there was a lot wrong with the plan, such as:

• Not enough brand planning. Netflix had failed to secure both the trademark and @qwikster Twitter handle.

• No user focus. Subscribers faced dual websites, queues, bills and lost Netflix’s single, unified recommendation engine.

• Bad timing. The announcement followed a price hike that had already angered customers.

• Poor communication, with the messaging being focused on the strategy, not benefits.

The result of this misteps was felt almost immediately. Netflix lost at least 800,000 subscribers in a single quarter. and the stock price dropped by more than 75% in four months.

But almost as quickly, the company responded - reversing their plans just 22 days after launch and adjusting to the overwhelming feedback. The stock price rebounded, customers returned, and renewed focus on product and content led to Netflix Originals.

So, somewhat ironically, Netflix didn’t have the Qwikster trademark when they actually needed it, but now they do, I very much doubt they’ll be doing anything with it in the future!

3. “What a Joke” - Making fun of yourself for views

Serial Nos: 88041782 (Reg. No. 5875454 — radio), 90818362 (entertainment expansion), and 88482751 (Reg. No. 6137493 — "What a Joke")

Status: "Netflix Is a Joke" is registered for radio entertainment (Class 38; since Oct 2019). Applications expanding coverage to Class 41 (streaming, live comedy events, festival production) are live and in progress. "What a Joke" (the Netflix comedy radio show) is separately registered (Class 41; Aug 2020).

The first the world heard about these trademarks was in 2017 when a series of mysterious billboards appeared across LA and New York, proclaiming that “NETFLIX IS A JOKE”.

Read more

Studios have always cancelled projects. In the past, the reasons were usually creative or logistical. A film might collapse under its own weight, or be sunk by clashing egos. That still happens, but something else is happening more often now.

Films are being cancelled on purpose, because finishing them no longer fits the plan.

I thought it would be helpful to go over why these cancellations happen, what they cost, and who is left carrying the consequences.

When films collapse

Studios have been cancelling films for almost as long as they’ve been making them. These decisions can happen at any stage, from deep development to right through to after post-production is complete.

The causes fall into a few sub-categories:

1. Development dead ends

2. Sabotage from within

3. Chaos behind the camera

4. Tragedy on set

5. External fallout

6. Franchise failures

7. The strategic cancellation

1. Development dead ends

Some projects stall before cameras roll. Kubrick’s Napoleon was in deep pre-production by 1970. Costumes were made, locations were scouted and Jack Nicholson had been cast. But after Waterloo flopped and MGM changed leadership, the project was cancelled. The budget felt suddenly too risky and the studio lost its nerve.

Disney’s Gigantic, an animated take on Jack and the Beanstalk, went through multiple rewrites and delays before it was shelved in 2017. By that point, years of development work had already been paid for.

2. Sabotage from within

Starting filming is no guarantee of finishing it, as some projects collapse from inside the production. Broadway Brawler began filming in 1997 but fell apart after Bruce Willis fired the director, cinematographer and several producers. Twenty days into shooting, Disney shut it down. Willis was later forced to make three further films under contract to cover the losses. One of them was The Sixth Sense (neatly bringing his career back from the dead).

In 1988, Atuk made it as far as a single scene before Sam Kinison demanded major script changes, then walked off. The production folded. Later attempts to restart the film failed. Several actors linked to the lead role (including John Belushi, John Candy and Chris Farley) died before any version was made.

3. Chaos behind the camera

Sometimes the dysfunction is not in front of the lens. Empires of the Deep, a fantasy epic funded by a Chinese property tycoon, ran through multiple directors and was shot in chaotic conditions. Cast members reported unpaid wages and poor treatment. One actress said her passport was withheld when she tried to leave. A trailer leaked, but no finished film was ever released.

Terry Gilliam’s first attempt at The Man Who Killed Don Quixote collapsed after a few days of shooting. A flash flood damaged the set, and lead actor Jean Rochefort suffered a serious back injury. The production was suspended and later abandoned. Insurance covered part of the loss. Gilliam eventually remade the film nearly twenty years later. The original collapse is documented in Lost in La Mancha, which should be required viewing for anyone heading into production.

4. Tragedy on set

Midnight Rider, a biopic about Gregg Allman, was cancelled on its first day. A train struck and killed camera assistant Sarah Jones during a scene that was being shot without a permit. Seven others were injured. The director was jailed. The film was never restarted. The only lasting result was a safety movement in her name.

5. External fallout

Sometimes the problem isn’t in the production at all. Netflix completed Gore, a biopic starring Kevin Spacey, but shelved it after sexual misconduct allegations emerged. The edit was finished but the film never went out.

Read more

Dig through the data and tell us what’s most interesting about the German film industry.

So I spent a few months exploring all aspects and ended up with five different insights, one of which I’m going to share today.

Namely, that German cinema has two distinct identities depending on who is watching.

1. Inside Germany, the box office is dominated by light-hearted comedies and family films. These films are popular and profitable in Germany but barely seen beyond German-speaking countries.

2. Outside Germany, the story looks very different. The films that succeed internationally are typically serious dramas centring on deep moral conflict, psychological tension or historical themes. They still do well within Germany but account for a much smaller share of domestic box office.

It’s a great example of how national cinema can be shaped by different expectations in different markets. What Germans see as representative of their culture is not always what the world recognises as German film.

Let’s dig into this….

The types of topics German filmmaker tackle

To understand why German cinema looks so different at home and abroad, it helps to start with the stories themselves. Over the past three decades, certain themes have consistently dominated German feature films.

The most common theme, by some margin, is family and relationships, present in nearly 60% of German films. Close behind are stories focused on society and social critique, psychological struggle, history and trauma, and personal identity.

Things are shifting slowly over time, with more films tackling external power over internal struggles.

So across the sector, we are dealing with a storytelling tradition rooted in the personal and the reflective. German cinema tends to ask questions about how people live, how society functions, and how individuals confront the weight of their past and present.

But the way these themes are expressed differs significantly depending on the intended audience.

German language films in Germany

In order to highlight the divide we’re going to focus in on movies made in Germany, which are in the German language, and also not in English.

Even before we get to what audiences think, we can see that German filmmakers in this category are much more likely to make comedies and/or family films.

The chart below shows the genre of German movies, split by language. Comedies and family films over-index among the German language production.

And this local appreciation for lighter fare is matched at the German box office, accounting for almost three-quarters of admissions since 2000.

A few examples include:

• Chantal im Märchenland (2024): A fantasy-comedy that led the German box office for several weeks in March and April 2024..

• Die drei ??? – Erbe des Drachen (2023): Based on the popular youth detective series, this family adventure film sold over 1.5 million tickets.

• Rehragout-Rendezvous (2023): Part of a Bavarian crime-comedy series, this film attracted over 1.5 million admissions.

• Manta, Manta – Zwoter Teil (2023): A sequel to the 1991 cult classic, this action-comedy directed by Til Schweiger attracted over 1 million viewers within its first three weeks, making it one of the most successful German films of the year, despite some pretty harsh audience feedback.

• Wochenendrebellen (2023): A heartwarming comedy-drama about a father and his autistic son embarking on a journey to find the son's favourite football team, grossing over $8 million domestically.

But when we look at how these same films perform across the Atlantic, it’s negligible.

Less than 1% of all the money grossed by German language German movies in the North America goes to comedies or family films.

I’ll repeat that for emphasis - in Germany they take 73% of the box office share and in the US and Canada it’s 0.9%.

So what are North Americans watching from Germany?

The types of German language movies that travel

When German films succeed internationally, they look very different from the ones that dominate the local box office.

Read more

Today, I thought I'd take a look at these two job titles and try to make sense of where each is used.

What is a cinematographer / director of photography?

In most real-world situations, the two job titles are interchangeable. Simply put, this person is responsible for crafting the film's visual style, or 'look'.  They report directly to the director and have a large number of people answering to them.  They run the camera department and also manage the gaffer (who runs the lighting department) and the grips (who are responsible for moving the camera on dollies, cranes, rigs, gibs, etc).

Whilst their work may be most apparent during the shoot, it actually starts in pre-production, when they will be working with the director to plan the look, hiring their team and managing decisions over what equipment to use.  And after the shoot is over they will often be included in the grade (the post-production process which tweaks the look and colour of the final film).

So, you may ask, why have both terms?  The historical reason is that many national film industries evolved independently from each other and so different terms developed for similar roles. The modern-day reason is that some of the people doing the job want to be credited as cinematographer while others prefer to be called the director of photography. It's this that I'm studying today.  It's mostly a matter of fashion and convention, which makes it a fascinating thing to measure, as we can see how states change in different parts of the industry.

A few years ago I used my dataset of the top 200 US-grossing films of each year of the past two decades (i.e. 4,000 films released between 1997 and 2016) and looked at the credits given to the cinematographer / DoP. Over this period, 1,092 different people received a cinematography credit on the films in my dataset.

Which term is used more widely?

The short answer is that director of photography is more commonly used.  58.4% of films released over the past two decades use the term, compared with 40.8% opting for cinematographer.  0.8% used other terms, including "shot by" (Sin City), "lighting cameraman" (Eyes Wide Shut), "director of cinematography" (Tees Maar Khan) and "photographed by" (Enemy).

The fashion for one term over another does seem to be cyclical, with cinematographer overtaking director of photography in the late 1990s and again in the early 2010s.

When we look at the genre of movies using each term, it seems that movies which rely on big, expensive visuals (i.e. fantasy, adventure and sci-fi movies) are much more likely to favour 'director of photography', while smaller, more human-interest genres (i.e. romance, drama and music-based movies) opt more often for 'cinematographer'.

In the chart above, you may notice that animation is the genre with the largest percentage of 'other' terms.  That's because the traditional role of a cinematographer is more varied on animation-based movies.  The process of creating the visual look of a live-action film differs significantly from that of a 2D hand-drawn animated movie, which again differs from a 3D computer-generated production.  In fact, only 52% of animated movies in my dataset even listed a cinematographer of any type.

An oft-repeated line about the difference between the two roles is that it has to do with the nationality of the crews.  I looked at the countries of origin for the movies and, as shown below, the UK seems to be where director of photography is most commonly used.

It should be noted that this study is only of the top 200 US grossing movies, and so if we wanted to do a thorough investigation of how the terms change between countries we'd need to study all movies made in each country, not just the ones which performed well in US cinemas.

Note: For readability, in the rest of the article I'm going to use the phrase 'cinematographer' to refer to people credited both with cinematographers and director of photography.

How many cinematographers shoot multiple movies?

A few weeks ago, I looked at the most prolific crew members on top-grossing movies, including cinematographers.  Roger Deakins came top, with 29 movies credits as cinematographer over the two decades studied, followed by Dean Semler (26 movies in the study), Robert Elswit (26), Andrew Dunn (26), Oliver Wood (24) and Don Burgess (24). But how typical are these prolific cinematographers?

Only 1.3% of cinematographers in my study were credited: an average of one movie a year (i.e. at least twenty credits over the twenty-year period).  Fewer than one in five cinematographers have shot more than five movies over these two decades.

Do cinematographers play well with other cinematographers?

I've previously shown how the role of cinematographer is the least likely of all major creative roles on a movie to be shared among two or more people - even less so than the role of director.

Only 2% of movies in my dataset had two cinematographers (such as Bollywood movie Ra.One which credits both a director of photography and a co-director of photography) and only 12 movies (that's 0.3% of the whole dataset) listed three or more.

So what are these rare movies which list multiple cinematographers?  Mostly animations (such as Chicken Run, which has two directors of photography and three lighting cameramen) and portmanteau movies (such as Paris, je t'aime and Movie 43, which have 17 and 9 respectively).

It's a man's world

Today's article focuses on the different terms used to describe the job, but it's not possible to cover cinematography and ignore the gender inequity.  Of all the major creative roles, cinematographers have the second worse female representation, only faring better than composers.

This has been a common thread for as long as the role has existed.   It wasn't until 1980 that the first woman was accepted into the American Society of Cinematographers (Brianne Murphy) and the first woman to receive an Oscar nomination for Best Cinematography did so this year!  Rachel Morrison earned her nomination for her work shooting Mudbound. Over the twenty years, only 45 of the 1,091 people I studied were female - that's just 4%.

Notes

The data for today’s research came from IMDb, Wikipedia and The Numbers. I have used the US box office as the criteria because the data is more reliable (and complete) than that of the worldwide box office. I focused on the top 200 grossing films, rather than all films made or all films released, in order to be able to compare like with like.

I used biographical information to determine the gender of cinematographers, with only one person out of 1,092 people eluding me. It’s worth reiterating that I was just looking at movies. This means that a cinematographer could have a very prolific career within TV, music videos, adverts, etc and yet not rank too highly within today’s charts.

Epilogue

The aim of today's research is not to decide which term is correct, but rather just to observe what has been happening in recent decades.

In researching this piece, I reached out to a number of cinematographers in the hope of finding a common consensus on the use of the terms.  After all my conversations, I concluded that there isn't any industry-wide agreement.   I am sure many people will have thoughts on the debate between the two titles, so I'd welcome these in the comments below.

Everyone seems to have a hot take on whether AI is going to radically transform the film industry, or is just flashy vapourware.

What is certainly true is that there are already a bunch of tools on offer to film professionals which change many of the tasks they have to conduct on a daily basis.

So I thought it would be useful to go through them and share with you the breadth and depth of what’s out there.

Note: Nothing in this article is an endorsement or any production, service or company. I have not been paid by any, and I am not recommending any. Examples cited are just illustrative of being offered, not what works, what’s good value, or what’s best in class. Just because someone is willing to sell you something, doesn’t mean it works!

We’ll go through seven different areas, including:

1. Screenwriting. AI can generate story ideas, produce outlines, draft dialogue, suggest rewrites, provide feedback on structure and market fit, and automatically tag scripts for production breakdowns. This means screenwriting may becomes more about curating and editing machine output, as well as managing floods of new iterations.

2. Packaging. AI can help turn a half-baked concept into a fully fleshed-out proposition overnight, identifying your audience and generating visuals, decks, and materials to make a project feel inevitable.

3. Financing. AI-driven platforms claim to model budgets, simulate box office, and rank casting choices by market value. Whether it’s real insight or risk cosplay, producers are being offered more data than ever to back up the numbers.

4. Legal. Producers with no legal background are leaning on AI to summarise contracts, generate agreements, and flag missing clauses. While risky, it’s easy to see the appeal when legal professionals aren’t exactly queuing up to work for back-end points.

5. Pre-production. Automated tools extract script elements, generate shooting schedules, match crew and cast to projects, scout locations using databases, and help directors visualise shot lists. Planning becomes faster and more integrated, but demands greater clarity and earlier creative decisions.

6. Production. Ironically, the bit that still needs people in a room with a camera may be the least affected in the short term. But streamers are already using patented tech to tag, track, and pre-process footage in real time.

7. Post-production. From editing and VFX to dialogue clean-up and synthetic voice work, post is where AI is having the fastest impact. Some of the biggest shakeups may come in dubbing and localisation, where traditional workflows are looking increasingly outdated.

Let’s start with where every good ideas does - in the script…

1. Screenwriting

The first category we’ll look at is perhaps the furthest along, since AI-generated text is currently more developed than its image, video, or audio counterparts.

It’s also the area likely to impact the greatest number of people. Screenwriting has the lowest barrier to entry of any filmmaking role as all it takes is an idea and somewhere to type it.

Generating story ideas or loglines

A number of platforms now offer AI-powered tools that generate basic concepts or loglines based on a few words, themes, or genre inputs. These can be used to explore different angles, test combinations of setting and character, or tailor stories to specific production resources such as locations or cast.

Writers can use custom-trained models in ChatGPT or Claude, loading their project bibles, beat sheets, or tone guidelines and then asking the AI to suggest new loglines, alternative scenarios, or setting variations that stay true to their worldbuilding.

It’s also possible to bring your characters to life, conversing with them or asking them to respond to prompts or scenarios. Character.AI lets users create and converse with AI-powered characters who respond in real time and in personality. Writers use this to roleplay scenes, stress-test character voices, or explore how two personalities might interact.

Tools not designed specifically for film can still be of use to screenwriters. On the surface, NovelAI is all about “anime-inspired” stories but it’s writing platform generates story ideas, dialogue, and even full scene descriptions that could apply in any genre. Likewise, while Jasper is a general-purpose writing assistant that can output story summaries or brainstorm premises tailored to custom instructions.

Building outlines and beat sheets

Several tools move beyond loglines into full structural outlines, using three-act or genre-specific templates. These systems can suggest a series of events or turning points from beginning to end, often editable or regeneratable with a single click.

StoryLab.ai generates outlines and hooks tailored to genre or audience type, while Plottr AI offers visual timeline tools with AI-assisted suggestions for beats and character arcs. Saga uses generative AI to build full plot structures that can be exported into screenplay format.

Writing scenes and dialogue

Some platforms will draft scenes directly, generating formatted screenplay text from a brief prompt or continuing from user input. These are typically aimed at breaking writer’s block or creating quick drafts for revision.

DeepStory can autofill scenes using its story model, aiming for logical consistency, and Sudowrite includes tools like “Continue Scene” or “Rewrite” to expand or rework text in context.

Rewriting or refining scripts

Other tools are designed to help polish or revise existing writing. This includes rewriting for tone or clarity, adjusting pacing, or aligning with specific genre expectations.

WriterDuet’s ScreenplayProof is a screenplay-specific proofreading assistant that says it “helps you refine your script without compromising your unique voice”, focusing on spelling, grammar, and formatting mistakes while preserving character dialogue and style. LanguageTool AI Rewrite rephrases sentences while offering multiple options per paragraph.

GrammarlyGO provides sentence-level rewrites for tone, simplicity, or length and I’m not sure I could personally cope if Grammarly was taken away from me!

Analysing scripts

AI is also being used to evaluate completed screenplays, providing feedback on story structure, emotional tone, character journeys, and even commercial potential. This is pitched as a tool for both development and investor decision-making.

ScriptBook uses machine learning trained on thousands of scripts to predict audience appeal, genre fit, and potential box office. While both Largo.ai and Cinelytic offer analysis and market insights based on script features.

Where this goes from here

These types of tools could conceivably shift the role of the screenwriter from a purely generative one to something more akin to an editor, curator, or systems designer. The “writer” might be shaping and refining content from machine outputs rather than always starting with a blank page.

And, rather terrifyingly, these tools may just be a brief intermediary steps to the point when AI models can “write” entire screenplays tuned to tone, structure, and cast without a human insight. Whether we want that or not is different topic entirely.

2. Packaging

Packaging is part business, part sales.

Producers need to make a series of smart decisions and compromises to turn an idea or script into an investable business proposition. AI tools can help by identifying a potential film’s audience, both in terms of overall scale and precisely who to target. Claritas and Parrot Analytics provide insight into audience behaviour and demand signals across regions.

They also need to create a sales pitch so compelling as to make the project feel inevitable to gatekeepers, investors, or collaborators. This often involves creating posters, key artwork and sales decks.

In my recent round up of the Cannes 2025 Marche du Film I noted how the early-stage sales posters have seen a dramatic rise in quality over recent years. A large part of that has been the proliferation of powerful image generation models.

While this may be acceptable at a private trade fair such as Cannes, the wider public are less forgiving. Disney got into a bit of hot water recently with a Fantastic Four poster. My favourite headline on the scandal comes from Futurism.com who said “Disney Says Its "Fantastic Four" Posters Aren't AI, They Actually Just Look Like Absolute Garbage“

3. Financing

Money makes the world go round, so it’s not surprisingly that AI-powered tools are being pitched to film professionals.

Although this area is still developing, a number of tools now offer data-led support. These systems claim to provide producers, financiers, and sales agents faster access to risk assessments, market intelligence and help with financial forecasting.

• Budget prediction and modelling. AI platforms can now scan a script or project brief and generate budget ranges. They do this by tagging elements such as cast size, locations, effects, or specific production requirements and mapping them against data from previous projects. This support can speed up early-stage planning and give a stronger foundation for conversations with investors or co-producers.

• Forecasting revenue and audience reach. Increasingly, AI models are being brought in to estimate how much income a project could generate. They use a combination of script analysis, cast information, genre, market trends, and data from similar titles. My personal opinion is that this area has a lot more hype than heft, as many services seem to address the perception of risk rather than actually minimising the actual risk. That said, maybe that’s what all film producers have done time immemorial.

• Assessing casting impact on financing. Casting is often the single biggest driver of project value in a financing context. Some tools now quantify the impact of attaching certain actors to a project by simulating how cast changes affect distributor and investor interest. This can influence early casting decisions, sales planning, and even negotiations.

• Creating pitch decks and finance packs. Building pitch materials and business cases is a notorious time sink. AI-driven tools now automate much of this by pulling market comparables, audience insights, and project-specific data into professionally formatted decks. These can be customised for different markets, investor types, or festival submissions.

Some of the products and services being offered in this area include:

• Largo.ai analyses screenplays and production data to generate budget ranges, forecast box office and VOD revenue, and simulate how different cast choices affect a project’s financial outlook. Also provides exportable reports for investors.

• Cinelytic ingests scripts or pitch decks to estimate production budgets, models potential revenue based on cast, genre, and release strategy, and offers star ranking models to show how casting impacts market value.

• Filmustage breaks down scripts to highlight elements that drive costs, such as cast days, location changes, or special requirements, and presents scene-by-scene financial implications.

• Vault AI uses market data and behavioural analysis to project audience demand, global reach, and income potential based on a project's creative details.

• Slated generates a “project score” by combining actor value, script quality, and package strength to support investment and distribution decisions.

4. Legal

Despite the obvious reasons not to, huge number of producers are turning to off-the-shelf generative text models like ChatGPT to produce or adapt their contracts.

On the surface this makes sense. AI can summarise long contracts, extract key clauses, and flag missing elements, in a heartbeat, all while telling you that you’re doing a great job.

Read more

National film cultures, policy choices, and industry structures have all shaped very different paths back to the big screen.

Across the 30+ countries I looked at, the following factors seemed most relevant when considering the level to which the cinema admissions bounced back from Covid:

1. Local films helped some countries fill the gap left by Hollywood delays. France bounced back faster than most thanks to a strong local film industry. With nearly 40% of 2023 admissions going to French films, including major hits like Astérix & Obélix, it kept audiences coming even when US titles were thin.

2. Streaming took a bigger bite in countries where it's already part of daily life. Sweden has high-speed internet and early adoption of streaming services, which meant many people grew accustomed to watching films at home. As cinemas reopened, fewer people felt the need to return.

3. People returned faster where cinema-going was already a strong habit. In Georgia, cinema-going has become a valued social activity in recent years. That enthusiasm helped push admissions 7% higher than 2019 levels, despite a full year of closures.

4. Extended lockdowns didn’t always hurt if reopening was done well. Spain had strict and extended restrictions on cinemas, and once reopened, many audiences didn’t return. The loss of habit, combined with economic strain, left Spain stuck at just 75% of pre-COVID levels.

5. Government funding helped keep cinemas open and attract audiences. The French authorities provided cinemas with direct financial support and ran national campaigns, such as discounted ticket days.

6. When money is tight, people stop going to the cinema. Türkiye saw inflation climb to 65% in 2023, making cinema tickets unaffordable for many. Even with popular local films, admissions were less than half their pre-pandemic levels.

Which countries have fully recovered or grown

Across all European nations, only Georgia saw an increase in admissions between 2019 and 2024, from 1.4 million tickets to 1.5 million, respectively. On the surface, this might seem like a slight increase, but in the context of a global downturn in theatrical attendance, any growth is notable.

Even among countries that haven't quite passed their 2019 numbers, a few are getting close. Bulgaria (92%), Lithuania (86%), and Serbia (85%) are among the top performers.

The standout performances challenge the idea that only large or wealthy markets can sustain strong cinema attendance. In this case, size has not equalled speed of recovery.

The countries that have mostly recovered but are not quite there yet

Several European countries have experienced strong recoveries, reaching between 80% and 90% of their pre-pandemic admissions levels. These include France (85%), Romania (85%), Serbia (85%), Lithuania (86%), and Bulgaria (92%).

Read more

But something has shifted.

I believe that we will look back at 2025 as an inflection point. The most negative reading might be that this was the year that Wylie Coyote realised there was no ground underfoot. A more positive reading might just be that it’s another bend in the wobbly journey of the evolution of the film industry.

The usual frenzy of handshakes, hype and headlines gave way to a slower, flatter market. Deals happened, screenings rolled, and the rosé flowed, but the energy was cautious and the big swings were few.

If last year’s Marché was about bouncing back, this year felt more like bracing for what comes next.

The mood across the market was one of quiet calculation. Sellers came armed with high-profile packages, but many buyers were wary, sitting on their hands unless the film came with a proven name, a festival slot, or a clear global angle. Conversations revolved around political risk, production costs, and the hard maths of international distribution. Everyone had something to show, but fewer people were buying, and even fewer were betting big.

And yet, amid the uncertainty, the market is clearly evolving. Cannes is trying to stay relevant, to reframe what the Marché is for and evolve how it delivers value.

Here are eight trends I spotted:

1. Cannes felt quieter and flatter this year. Cannes 2025 lacked its usual buzz, with a subdued atmosphere and restrained energy throughout the Marché.

2. Everyone is playing it safe with their money. Buyers and sellers were cautious, avoiding risk and focusing only on projects with proven commercial appeal.

3. The middle tier of films is being squeezed out. Mid-budget, non-star-driven films struggled to attract buyers in a market focused on extremes.

4. Microsoft vanished, but the AI conversation stayed. AI remained a major theme despite Microsoft's absence, with growing focus on ethical, legal and creative concerns.

5. Cannes is turning into a conference as much as a market. The Marché featured more panels and talks, reflecting a shift towards thought leadership alongside deal-making.

6. Poster design is finally catching up. Sales artwork quality improved significantly, signalling a more professional and visually strategic approach to marketing.

7. The market space is changing to make room for seating not sales. Informal seating areas replaced some booths, reflecting changing behaviours and networking styles.

8. Sustainability is still a brand, not a behaviour. Despite green messaging, Cannes practices remained resource-heavy, revealing a gap between values and actions.

Let’s go through each in turn.

1. Cannes felt quieter and flatter this year

Across the board, professionals commented on how subdued the market felt this year. There was no single high point, no day when everything seemed to come alive.

This was partly structural as the Marché started on a Tuesday, breaking with its usual rhythm and thinning out the natural build-up of energy. But it was more than just the calendar. The buzz that usually defines Cannes felt muted, replaced by a steadier, more restrained tempo that never quite spiked.

Despite a reported 15,000 attendees and a packed programme of screenings and talks, the atmosphere inside the Palais and around the pavilions was noticeably less intense.

For many, the week felt more like a collection of parallel meetings than a unified market. Some sales agents described buyers as “very focused” and “engaged,” but only around very specific projects. Most others felt that the pace was slower and the stakes were lower.

This lack of urgency may point to a more profound shift. Cannes is no longer guaranteed to be the epicentre of deal-making. With digital tools, year-round access to buyers, and growing regional markets, Cannes must now prove its value each year.

2. Everyone is playing it safe with their money

Risk aversion was the defining financial mood of this year’s Marché. Buyers were highly selective, sellers were under pressure, and nobody wanted to be the one caught overpaying in a volatile market.

The main trigger was political: Donald Trump’s proposal of a 100% tariff on foreign-produced films sent a wave of uncertainty through the Croisette. While many dismissed it as posturing, it immediately clouded negotiations, especially for projects relying on North American pre-sales.

This sense of caution filtered through every level of the market. Streamers were present but subdued, favouring global rights on big, recognisable titles over speculative buys. U.S. distributors were particularly hesitant, forcing international sellers to adjust expectations downward.

Some buyers openly stated they were holding back to see where the major players landed before making offers. Even those still active were operating with “tighter filters,” carefully weighing commercial viability, star power and genre before committing.

The result was a fragmented market, where a few packages generated competitive bidding, but much of the inventory sat idle. Films without A-list attachments or strong festival credentials struggled to move.

3. The middle tier of films is being squeezed out

One of the most apparent structural shifts at Cannes 2025 was the near disappearance of the market’s middle ground. Films that would have comfortably sold five or ten years ago, such as modestly budgeted dramas or festival-friendly titles without big stars, are now struggling to find homes. This middle tier has long been the backbone of the independent sector but is increasingly being edged out by a market that only rewards the extremes.

At the top end, there was no shortage of money for star-driven or award-contending packages. Mubi’s reported $24 million multi-territory deal for Die My Love set the tone early. The project had all the right ingredients, with Jennifer Lawrence, Robert Pattinson, and a director with a Cannes pedigree.

Everything else, especially films without marquee talent, faced an uphill battle. Sales agents described a growing reluctance among buyers even to consider projects that lacked clear marketing hooks or theatrical potential.

Films that might once have been seen as safe bets are now viewed as liabilities. They are squeezed between rising production costs and shrinking downstream revenues. The message from buyers was clear: If your film cannot stand out at a glance, it might not stand a chance.

4. Microsoft vanished, but the AI conversation stayed

Last year, Microsoft was a key sponsor of the AI Pavilion at Cannes. This year, they were nowhere to be seen.

(While it may not make corporate sense to promote technology to filmmakers, 2025 saw the return of ever-present stimulants sponsors, hawking caffeine, alcohol and sugar).

Microsoft’s decision not to return was noticeable, mainly because artificial intelligence remained one of the most talked-about themes across panels, tech showcases, and side events. The withdrawal of one of tech’s most prominent players suggests that the corporate push behind AI in film may be entering a new phase. It now feels less about visibility and more about caution.

Even without Microsoft, the Marché leaned heavily into technology. The newly launched Village Innovation covered one thousand square metres and featured tools ranging from AI script analysis to virtual production stages.

But alongside the excitement came concern. Many panels struck a cautionary tone, highlighting unresolved questions around copyright, labour rights and creative control.

Some speakers warned against using AI to mimic the styles of known directors. Others called for clear ethical guidelines and legislative safeguards. AI in film has grown from an ignorable novelty to a real, scary legal, artistic and industrial challenge.

5. Cannes is turning into a conference as much as a market

This year’s Marché featured a noticeable shift in programming. Alongside the usual meetings and screenings, there was a marked increase in the number of live panels, talks and curated public events.

A cynic might conclude that this was a deliberate effort to give attendees another reason to buy a €439-€600 Marché pass rather than conduct their business on the other side of the security barrier (read: Caffe Roma).

But also it reinforces the notion that the Marché is a space for thought leadership and strategic conversation, not just transactions. In practice, it means more people sitting in rooms listening to peers, and fewer hours on the floor chasing sales.

And it is working. Many of the talks were packed, with strong line-ups and genuinely engaged audiences. Panels covered everything from international co-production to AI ethics and sustainability. They featured high-profile names and plenty of substance.

However, while the number and quality of sessions have grown, I researched who the speakers were and found that the platform still belongs to a narrow group. Across all Cannes conference events in 2025, around 85% of speakers were white, and male voices still dominated the majority of events.

Gender representation is improving, with women making up nearly half of panel participants, but racial diversity lags far behind.

In addition, men are more likely to showcase work than participate in “talent development”, with the inverse being the case for women.

6. Poster design is finally catching up

One of the more surprising shifts at Cannes this year was the noticeable improvement in poster design. For years, the Marché has been known for its comically poor sales artwork.

Read more

Cinemas are packed with re-releases in 2025, offering everything from cult favourites to landmark anniversaries. Titles returning to the big screen include 28 Days Later, Demon Slayer: Mugen Train, Dogma, Brokeback Mountain, and a full Studio Ghibli slate celebrating 40 years of the iconic animation studio. Whether it's the 75th anniversary of Sunset Boulevard, the 60th of The Sound of Music, or the 50th of Jaws, this year’s calendar is stacked with chances to revisit classics in theatres.

If the past is any indication, some of these will bring in millions while others will collapse entirely.

In the same year that Coraline earned over $29 million in a limited run, Snow White struggled to attract more than 17 people per screen.

Revenge of the Sith packed cinemas two decades after its first release, while Morbius failed a second time despite widespread internet attention.

Understanding why a re-release works means looking closely at timing, audience connection, presentation quality and cultural context. Success depends on alignment. When the right factors converge, even a film from fifteen years ago can feel like a significant new event.

From looking at the data, I see four conditions for success:

1. A strong emotional connection. Films with a lasting fanbase or cultural presence attract audiences who feel personally invested.

2. A clear temporal hook. Re-releases work best when aligned with anniversaries, seasonal slots or sequel build-up.

3. A presentation that adds value. Upgrades like 4K, 3D or added footage give returning viewers a reason to see the film again.

4. A cultural connection (that has grown over time). Movies that have grown in reputation or visibility since their original release benefit from renewed interest.

Let’s look at each in detail…

1. A strong emotional foundation

The starting point is affection, meaning that movies with a devoted fan base or enduring cultural presence have a clear advantage. This is not to be confused with critical acclaim or box office history - we’re talking about lasting emotional value.

For the re-release of Coraline, Laika built its re-release campaign over five years. The studio posted regular behind-the-scenes content and built a following of more than ten million across platforms. By the time the 15th anniversary arrived, fans were ready. The film returned with 4K and 3D upgrades, new footage and a limited run through event cinema distributors. It earned more than $29 million globally and became the most successful classic title for Fathom Events to date.

Revenge of the Sith followed a different path but relied on the same principle. The perception of the Star Wars prequels has changed since their first release. What once felt polarising now feels personal to a generation who grew up with them. Disney released the film in May 2025 to mark its twentieth anniversary. It aligned with Star Wars Day and included surprise cast appearances. The run earned over $42 million and helped the film cross $900 million lifetime.

2. A clear temporal hook

The next thing that a successful re-release needs is context. The best examples are tied to a specific moment, such as an anniversary, a new sequel, or a seasonal window. This timing creates urgency and gives the audience a reason to return now rather than wait for streaming.

The 2012 3D version of Titanic grossed $350 million, with the 2023 reissue adding another $70 million. In both cases, the presentation quality helped, but the timing was essential. The campaigns positioned each release as a celebration, not just a reprint.

The 2022 re-release of Avatar was timed to prepare audiences for the sequel. IMAX and 3D formats drove most of the sales. In some markets, they accounted for more than 75% of tickets. The film earned over $30 million in its opening weekend alone.

3. A presentation that adds value

Modern audiences expect more than nostalgia. The theatrical experience must offer something that streaming cannot match. That might include visual upgrades, exclusive footage or a communal setting.

Read more

It shows the amount of time between when movies are released in the US and when they come out in Japanese cinemas. The line is mostly flat, dipping only slightly in the past few years as the effect of the Covid lockdown ripples out.

But add some context, and everything changes.

There are two things which jump out when you compare it to other countries:

1. Firstly, movies come out in Japan a lot later than they do in other countries.

2. Secondly, the delay has largely stayed consistent, bucking the global trend.

France, the UK, South Korea, and dozens more have almost eliminated the delay as global releases have become more common. Studios want audiences everywhere to see blockbusters at the same time as it helps marketing, reduces piracy, and builds momentum.

Yet Japan hasn’t followed that trend. While the rest of the world has moved toward simultaneity, Japan has stuck with a schedule of its own.

Why Japan still releases foreign films later

I’m going to go through things in a little more detail but let’s start with an overview as to why this happening.

Japan’s release pattern is a deliberate strategy shaped by the way the industry works, what audiences expect, and how success is measured. The key reasons are:

1. Distributors want proof before committing. A film's success overseas is often used to shape the marketing in Japan. That means waiting to see how it performs. If it wins awards or does well in the US, that becomes part of the campaign.

2. Delays don’t hurt box office performance. Data shows no significant drop in revenue for films released late. That removes the main incentive to change.

3. Localisation is slow and detailed. Subtitles and dubbing for Japanese audiences involve more than translation. Scenes and references are adapted culturally. This takes time and adds to the delay.

4. Domestic films dominate the box office. In 2024, nine of the top ten films in Japan were local. Distributors and exhibitors give them priority. Foreign films that aren't major franchises often come second.

5. Releases are timed around holidays. Golden Week, Obon, and New Year are peak cinema seasons. If a film doesn’t fit one of those windows, it gets delayed.

6. Screen availability is limited. Japan has fewer cinema screens per person than most developed countries. Theatres have to be selective. That creates a queue, and foreign films often wait their turn.

7. Long theatrical windows remove time pressure. Unlike in the US, where films are pushed to streaming quickly, Japan keeps them in cinemas longer. There is no rush to release everything at once.

Let’s look at each of these in turn to understand Japan’s film market a little better.

1. Distributors want proof before committing

Japanese distributors often delay the release of foreign films until they see how they perform overseas. This “wait and see” approach is common, especially for films that are not part of established franchises. Success abroad provides validation, which is then used in Japanese marketing.

A film that becomes a hit in the US or Europe gives the distributor material to work with. Japanese movie audiences are accustomed to seeing claims like like:

• “全米が泣いた” meaning “All of America cried”

• “全世界No.1大ヒット“ meaning “The No.1 Hit Worldwide“

• “アカデミー賞長編アニメーション賞受賞“ meaning “Winner of the Academy Award for Best Animated Feature“

These claims can’t be made until the film has run elsewhere.

By the time Everything Everywhere All At Once opened in Japan (nearly a year after its US debut) it had won multiple awards and built strong global buzz. That success was part of the local campaign. The same strategy was used for Oppenheimer, which received a Japanese release date only after its Oscar run.

Distributors also use overseas performance to scale their investment. If a film underperforms in its home market, it might still open in Japan, but with fewer screens and limited promotion. If it flops, it may not open at all. This cautious approach helps avoid expensive failures and tailors each release to what the Japanese market is likely to tolerate.

It’s worth noting that franchise films are often treated differently. Marvel, Star Wars, and similar blockbusters are often released in Japan simultaneously with the US, or even slightly earlier. Their success is already assumed. For everything else, proof comes first.

2. Delays don’t hurt box office performance

In most countries, releasing a film late risks piracy, spoilers, and lost momentum. In Japan, that risk does not appear to apply.

A 2018 study by the University of Amsterdam surveyed nearly 35,000 people across 13 countries and found Japan had some of the lowest online piracy rates in the world. While piracy was widespread in places like Spain and Canada, it was rare in Japan. The report linked this to strong enforcement and cultural respect for copyright, both of which help explain why Japanese audiences still show up for delayed theatrical releases.

Spoilers are a concern, but not at the level seen in the US or UK. Many viewers wait, even if they already know the story. As long as interest remains, the delay does not cut into revenue.

3. Localisation is slow and detailed

Preparing a foreign film for release in Japan involves much more than translation. Subtitles and dubbing require careful adaptation, and this process takes time. Cultural references, humour, and even on-screen visuals are often changed to match Japanese audience expectations.

In Inside Out, Pixar replaced a piece of broccoli with a green pepper in the Japanese version. The change was made because green peppers are more commonly disliked by Japanese children. In Zootopia, a moose newscaster was changed to a tanuki, a raccoon dog familiar from Japanese folklore. These edits are minor in appearance but require planning, animation updates, and voice re-recording.

Language structure adds to the delay. Japanese uses three writing systems and follows different sentence structures from English. Translators must restructure dialogue, observe strict character limits for subtitles, and handle nuances such as politeness levels or gendered speech.

Japanese audiences expect this level of care. Poor localisation can damage a film’s reputation. Distributors know that rushing this process risks losing audience trust, so they take their time. The result is a polished product, but one that arrives later than in other countries.

Read more

A few weeks ago I examined what Netflix’s patents could tell us about the future of film and streaming.

Amazon’s video-related patents, however, make Netflix’s collection look like a tiny start-up. Amazon has filed almost fifteen times as many patents in similar areas, covering everything from development algorithms to distribution systems.

After going through 7,378 of these patents, I found a series of significant shifts that may soon change the daily reality for filmmakers.

Here are some of the highlights at each stage:

• Development. Automated idea generation, data-driven story testing, predictive casting, and built-in tools for testing commercial appeal before you even write a script.

• Pre-Production. Fast creation of 3D models for sets and costumes, automated shot lists and schedules, real-time digital breakdowns, and asset tracking driven by AI.

• Production. On-set recording that feeds straight to the cloud, real-time shot tracking, dynamic metadata capture, panoramic and multi-angle capture, emotion and performance monitoring, and live accessibility tools.

• Post-Production. Scene-level microsummaries, AI-powered dubbing and subtitling, automated video quality control, edit tracking, and embedded rights management.

• Distribution. Adaptive streaming with multiple versions, dynamic in-stream monetisation, instant push to video-on-demand, heavy encryption and access controls, and platform-driven promotion tied to viewer engagement data.

Let’s look at each stage in more detail:

Development

The development stage is moving towards a system where audience data and predictive tools are at the centre of project decisions.

For filmmakers, this means new projects start with market analysis, automated idea generation, and early indicators of commercial potential. The decisions filmmakers make about stories, cast, and even branding will increasingly reflect the results of these tools.

Here are a few choice examples:

Predicting success before the script is written

Patent US10600060B1 introduces a system for predicting viewer engagement and content performance before a film is even made. It models viewer responses based on demographics, prior behaviour, and historical analogues to similar content. For development teams, this means a concept can be scored for its likely impact before commissioning a treatment, let alone a full script.

Algorithmic generation of market-ready ideas

Patent US11551096B1 takes this a step further with an automated design engine. The system scans market trends, social signals, and user behaviour to generate ideas for new characters, plot combinations or even entire stories. Each combination is scored for potential market fit. For producers, it’s a tool that can suggest viable development directions grounded in audience demand.

Testing ideas with synthetic audiences

With WO2024258464A1, Amazon has patented the use of synthetic data to simulate audience responses. Instead of waiting for real feedback from test screenings or pilot releases, creators can preview how different audiences might respond to key plot points, visuals, or messaging. It lets studios stress-test risky ideas at concept stage, without real-world exposure.

Casting that starts with the algorithm

Amazon’s development tools even reach into casting and prep. Patent US10475185B1 describes automated scene-specific tagging of participants and actors. In practice, this allows systems to recommend talent or casting tape selections based on early story drafts, visual material, or comparable roles, long before the formal casting process begins.

Collaborative writing with built-in commercial analysis

Amazon’s story collaboration tool, US9553902B1, enables real-time story development with predictive scoring. It lets multiple contributors work on story elements while algorithms monitor for “completion likelihood” and “market resonance.” Authors are shown live metrics on how likely a project is to be finished and how much reader interest it’s likely to attract. Feedback loops track whether a story concept has enough commercial pull to move forward.

E-commerce hooks embedded from the beginning

Amazon’s design of “shoppable” story experiences begins in development. Patent US10021458B1 and its sibling US9973819B1 describe interactive commerce layers built directly into video content. The patents cover overlays, pop-ups, and in-scene product tagging linked to an ecommerce platform. For filmmakers, this means that brand integration, merch tie-ins, and product placements are no longer late-stage add-ons but they can (and increasingly must) be planned from the pitch deck.

Pre-production

Amazon is working towards a pre-production period that is ever more digital, rather than physical.

Planning now includes building virtual sets, running automated breakdowns, and using AI to organise schedules and assets. This means preparing every element of a shoot with digital models and linked metadata rather than traditional paperwork and manual management.

Here are a few of the pre-production patents which caught my eye:

From photos to fully navigable sets

Amazon has filed patents for generating 3D models of spaces and people using only 2D images. These models can be built from photographs or standard camera footage then used for digital scouts, set design, costume planning or previsualisation.

Systems described in US11367263B1 and US11783560B1 enable fast conversion of flat images into scaled 3D environments and character models. This allows production teams to walk through sets, test wardrobe fits and stage scenes virtually before anyone has stepped onto a location.

Instant storyboards from scripts

Previsualisation also benefits from automated shot planning. Patent US9992556B1 describes a tool that generates storyboards directly from screenplays. It identifies scenes, characters, and likely camera placements, producing basic shot lists and visual outlines. The director can refine these, but the system handles the initial legwork. This turns script breakdowns and blocking into a faster, more iterative process.

Real-time scheduling and casting logic

Patent US10257266B2 focuses on tracking actor resources during preproduction. It connects availability, call sheets and location data to ensure that casting and logistics are synchronised. Changes in one part of the schedule automatically cascade through the rest of the plan, reducing conflicts and manual rework.

Cameras that prep themselves

Amazon’s imaging and sensor patents include systems for camera calibration without manual intervention. US9986233B1 and US10050429B2 describe methods for auto aligning and correcting camera systems including lenses used in scouting and previs.

Read more

If the United States were to define what makes a film “American,” it would have to choose which aspects of production and identity matter most.

Other countries have taken various approaches to this question, each balancing clarity, flexibility, and enforceability.

Here are the main factors in play:

• Financial Spend. Many countries tie national status to where the money is spent. Germany and Ireland require minimum local expenditure to unlock tax benefits.

• Ownership and Control. Some systems focus on who owns the rights and who makes the key decisions. Canada, for example, demands that the producer be Canadian and hold copyright for 25 years. France and Spain also scrutinise company structure and control.

• Key Personnel. Nationality or residency of the director, writer, lead cast, and heads of department is often central. Brazil uses strict ratios, while Canada and Australia assign points to creative roles.

• Language. Language can act as a proxy for cultural identity. France offers 20 points for films in French, while Ireland’s Section 481 scheme awards points for English, Irish, or European languages.

• Cultural Content. Some systems consider what the film is about, i.e. its themes, setting, and characters. The UK and Germany both use cultural tests that reward stories rooted in local history or society.

• Treaty Co-Productions. Many countries allow films to be deemed national if they meet the terms of bilateral or multilateral treaties. These usually require proportional financial and creative input from each side. Australia and France use these to expand their film footprint, especially for international projects with little domestic subject matter.

• Legal Registration. A few countries, like South Korea and Japan, use legal registration of the production company as a defining factor. This provides administrative clarity but does little to ensure economic or cultural benefits.

I have broken down the approach of 18 major film-producing nations below:

• Australia – Uses a flexible cultural test (SAC) or treaty co-productions, focusing on creative control, spend, and key personnel.

• Brazil – Requires Brazilian ownership and majority local cast/crew, with strict residency rules enforced through a federal certificate (CPB).

• Canada – Relies on a points-based system plus ownership and control rules, with separate agencies overseeing tax credits and broadcast eligibility.

• China – Grants nationality through regulatory compliance, censorship approval, and Chinese company involvement, not cultural content.

• Denmark – Applies a discretionary, case-by-case approach based on cultural contribution and Danish creative involvement.

• France – Uses a detailed points system focused on language, personnel, and production services, tied to its powerful subsidy mechanisms.

• Germany – Combines a cultural test with financial thresholds and minimum spend requirements, closely managed by federal bodies.

• India – Has no single cultural test but recognises national status via full Indian production or treaty co-productions certified by the NFDC.

• Iran - Nationality tied to state approval at all stages; ideological compliance overrides all else.

• Ireland – Blends cultural and economic criteria in both funding and tax systems, emphasising Irish creative leadership or contribution.

• Japan – Uses a light points system for co-productions and requires Japanese company involvement, with a focus on financing and rights.

• Russia - Seven legal criteria covering personnel, spend, language, and producer nationality.

• South Korea – Bases nationality on corporate registration plus creative and financial contributions for public support schemes.

• Spain – Enforces strict thresholds for EU personnel and production work, with mandatory certification for all national benefits.

• Sweden – Applies a case-by-case assessment focused on a Swedish producer and meaningful creative input from Swedish nationals.

• United Kingdom – Uses a detailed points-based cultural test or treaty co-productions, linked directly to access to tax credits.

Let’s look at each in detail:

Australia

Australia maintains a formal and highly consequential definition of film nationality, primarily for access to government financial support. A film must be officially recognised as ‘Australian’ to qualify for the Producer Offset, a refundable tax rebate worth 40% of qualifying Australian expenditure for feature films, and 30% for other formats. This classification also affects eligibility for Screen Australia funding and determines whether a title counts towards broadcaster content quotas. Two pathways exist: passing the Significant Australian Content (SAC) test, or being certified as an Official Co-production under a bilateral agreement.

Films that are not official co-productions must pass the SAC test, which is laid out in legislation and assessed by Screen Australia. The test considers five factors:

• Subject matter

• Production location

• Personnel nationality and residency

• Local expenditure

• Any other relevant details.

There is no points system. Instead, Screen Australia makes a holistic judgment, balancing cultural and industrial criteria. For example, a film might succeed based on Australian creative control and local spend, even if the story is not overtly ‘Australian’. This flexibility benefits projects with international ambitions but creates uncertainty for producers trying to predict outcomes.

For co-productions, nationality is assessed differently. Screen Australia evaluates the financial and creative contributions of each partner country using a points-based system focused on the nationality of key personnel. Directors, writers, editors, and lead cast all attract points. This route offers greater predictability and is often used for projects without strong Australian subject matter but with substantial Australian involvement. Once certified, co-productions are automatically treated as ‘Australian’ for funding purposes and bypass the SAC test.

Screen Australia is the lead certifying body. It administers both the SAC assessment and the co-production program. The Department for the Arts oversees the Location and PDV Offsets and manages co-production treaties. The Australian Taxation Office processes the actual tax rebate. The Film Certification Advisory Board provides policy advice to the Minister for the Arts.

Several high-profile examples illustrate how these rules play out. Baz Luhrmann’s The Great Gatsby and Elvis both qualified for the Producer Offset, despite their American stories and settings, due to their significant Australian creative control, crew, and spend. Conversely, Taboo, a documentary series produced by Australians, was denied access, reportedly due to concerns over non-Australian subject matter and creative control, leading to legal and public disputes.

Brazil

Brazil defines the nationality of a film through a structured legal framework administered by ANCINE, the federal audiovisual agency. The designation determines access to public funding, tax benefits, screen quotas and international treaty advantages. Films must secure a Certificado de Produto Brasileiro (CPB) to be officially recognised as Brazilian. This certificate is mandatory for exhibition on Brazilian platforms and for participation in federal incentive schemes.

To qualify as Brazilian, a film must be:

• Produced by a registered Brazilian company, with its headquarters and controlling interest held by Brazilian nationals or long-term residents.

• The director must be Brazilian or a foreigner residing in the country for over three years.

• At least two-thirds of the cast and crew must meet similar nationality or residency criteria.

• While language is not strictly required to be Portuguese, some bilateral treaties stipulate specific language conditions.

• The location of filming is not sufficient to establish nationality.

• For co-productions without treaties, stricter rules apply, such as a minimum 40% ownership by the Brazilian company and adherence to crew ratios.

• Separate rules also exist for works seeking "independent" status, which excludes links to broadcasters or pay-TV providers.

ANCINE is the primary decision-maker and enforcer. It verifies applications, issues the CPB, and oversees treaty and non-treaty co-productions under Normative Instructions 104 and 106. It also administers the national film fund (FSA), enforces screen quotas, and registers all foreign content exhibited in Brazil.

One practical example is the case of a Brazilian-Israeli co-production where the treaty allowed the original soundtrack to be in either Portuguese or Hebrew. Such flexibility does not exist in non-treaty co-productions, which face stricter requirements. The film I'm Still Here, nominated for an Oscar, drew attention not for failing to qualify but for raising questions about representation and eligibility in the wider industry. Conversely, no recent high-profile disputes over CPB issuance were found in the current research.

While the classification rules are clearly outlined in law, the application process relies on evidence submitted by producers and interpreted by ANCINE. Most disputes arise over co-productions or proof of residency. The agency has responded by digitising and streamlining CPB processes, especially for low-risk content such as music videos and pornography, which form a significant volume of applications.

Canada

A film’s nationality in Canada is critical as it unlocks a range of subsidies, tax credits, and domestic broadcast quotas. Two main bodies handle classification: the Canadian Audio-Visual Certification Office (CAVCO) handles eligibility for federal tax relief, while the Canadian Radio-television and Telecommunications Commission (CRTC) governs what counts as Canadian for broadcast regulations. A third body, Telefilm Canada, manages co-productions and provides funding. These systems mostly align but have key differences in ownership and control requirements.

To qualify as a Canadian production, a film must earn at least six out of ten points in a prescribed system. Points are awarded for:

• Canadian citizens or permanent residents occupying key creative roles such as director, screenwriter, and lead performers.

• To pass, one of the two lead performers and either the director or screenwriter must be Canadian.

• However, meeting the point threshold is not enough. The producer must be Canadian, have full creative and financial control, and own the copyright for 25 years if the production is claiming tax credits under the CAVCO regime.

• The production must also spend at least 75% of its labour and post-production budgets in Canada.

• Language is relevant too. Only films in English, French, or Indigenous languages typically qualify.

• Some genres are explicitly excluded, such as reality TV, advertising, and live sports.

CAVCO is the key certifying body for tax purposes. It also administers a separate regime for service productions that are foreign-owned but use Canadian crews. The CRTC handles certification for television broadcasts and recognises most CAVCO decisions but does not require the 25-year copyright rule. Telefilm Canada plays a pivotal role in approving treaty co-productions and offers funding to projects that score higher than the minimum or meet alternative criteria.

Schitt’s Creek is often cited as a fully qualified Canadian series, achieving 10 out of 10 points. Yet it deliberately avoids visible Canadian markers. Conversely, The Handmaid’s Tale and Turning Red failed to qualify because they were American-owned. This gap between technical qualification and cultural resonance is at the heart of the system’s controversies.

Critics argue the current model allows productions that feel American to qualify, while disqualifying authentically Canadian narratives created with foreign financing. The point system is also under pressure to evolve, with the CRTC exploring reforms that would expand the number of roles counted or introduce cultural criteria.

Streamers like Netflix have clashed with Canada over the requirement that a Canadian entity hold the copyright for 25 years. This demand conflicts with their global licensing strategies. The recent Online Streaming Act has brought new players into the regulatory net but faces ongoing legal challenges from industry giants.

China

China does not use a cultural or points-based system to define a film’s nationality. Instead, a film is treated as Chinese if it is produced by a registered mainland entity with official permits and passes state censorship. Co-productions with foreign partners can achieve equivalent status if they meet strict legal, financial, and ideological requirements. National classification in China is not about language, themes, or personnel but about regulatory compliance and political alignment. This affects everything from market access and censorship rules to eligibility for funding and exemption from import quotas.

• For co-productions to be treated as Chinese, the film must be jointly financed, typically with at least 25% of the investment from a Chinese company.

• It must include Chinese elements, such as story content, cast or crew, and shoot or post-production within China.

• There are no fixed percentages for every criterion, but guidelines suggest that one-third of major actors should be Chinese and that principal crew be roughly balanced between Chinese and foreign nationals.

• Recent draft rules for online content propose that only 25% of the main cast and crew need to be Chinese for web series to qualify.

• The most important hurdle is censorship. Scripts and completed films must be reviewed and approved before the film is granted permits for production and public screening.

Read more

"We want movies made in America, again!"

The statement lit a fire under an issue the industry has rarely confronted head-on. What does it actually mean for a movie to be made in America? Is it about where the cameras roll, where the money comes from, or who controls the creative vision? And how many so-called Hollywood films would actually pass the test if one were imposed?

Film production has become one of the most globalised industries on earth. Crews, financing, visual effects, post-production, and even actors routinely cross borders. Many American films are shot in Canada, the UK, or Australia. And while Trump’s proposal assumes a clear line between foreign and domestic, the modern reality of filmmaking is a blur.

Let’s see what the data can reveal about where they are really made, and what factors any American tariff system would need to contend with.

How many movies are made in America?

The phrase “made in America” sounds simple, but is hard to discern for most major movies. If we use IMDb’s country of origin, then around a third of global feature films could be classed as American since 2000.

If we focus just on movies from major American Hollywood studios, then that rises to more than four out of five (87.3% in 2024).

However, it’s not as simple as “American or not”. That’s because:

• Films typically have multiple countries of origin. Many films today are co-productions between two or more countries. This allows them to access multiple tax incentives, funding sources, and distribution markets.

• There’s no system for tracking it. There is no universal database or framework for defining or verifying a film’s national identity. IMDb allows producers to self-report, while awards bodies, festivals, and national funds each apply their own criteria. These systems often conflict, leading to different answers depending on who is asking and why.

• A global supply chain. Films are made through a global production pipeline. Just like manufacturing, cinema today relies on an international supply chain rather than a single point of origin.

• Streaming platforms blur boundaries further. Global streamers like Netflix and Amazon regularly commission films with no clear country of origin. A project might be developed in LA, produced by a UK company, shot in Spain, and marketed as “international content.”

• American studios regularly back non-American films. Hollywood studios often finance, produce, or distribute films that are not American in origin. This can include prestige dramas aimed at awards season, international co-productions, or local-language content for foreign markets.

How do you figure out where a film is “from”?

Any tariffs would require films to have a clear national identity. Currently, the United States has no formal definition of what makes a film American. There is no certification process, no threshold of domestic content, and no single agency responsible for determining national status.

The term “nationality” can mean different things depending on who is asking and what is at stake. Festivals, awards bodies, funding agencies, and audiences all apply different criteria, and the results can vary dramatically (I’ve added some controversial examples at the end of this article).

Many people first wonder in which country the actual filmmaking took place. Shooting locations are, in theory, easy to identify. A film either rolls cameras in a place or it does not. But this becomes complicated when productions are split across multiple countries or use remote teams.

When I looked at this a few years ago, I discovered that the average top-grossing movie is shot in 1.6 countries.

The UK and Canada are the top foreign locations for studio productions. The chart below shows the percentage of movies made by Hollywood studios, which have America as a country of origin on IMDb, that had at least one day of filming in the UK or Canada. As you can see, it’s significant. In 2019, 23.9% of such productions were shot in the UK, while for Canada, the figure was 19.4%.

These shoots are not always weighted equally. One country may host a few days of aerial photography, while another carries the bulk of the production. Some offer studios and crews. Others offer tax rebates or natural landscapes. Some supply specialist services like underwater filming or digital doubles. Measuring a film's production true footprint by location would require weighting the contribution of each territory.

Another method is to look at the primary production company. Some major festivals, such as the Berlinale, define a film’s origin based on the registered office address of its main production entity. This is easy to verify and administratively convenient. But it often fails to reflect the broader picture. Many films are made through international co-productions, with creative and financial input coming from multiple sources. Pinpointing the “main” producer is rarely clear-cut.

Official co-productions bring their own rules. Countries with bilateral or multilateral treaties often allow a film to qualify as “national” in more than one country. This provides access to local tax breaks and funding, but only if strict criteria are met. These include minimum financial contributions, nationality requirements for key crew, and points-based systems assessing cultural relevance. A film can pass as “French” under one set of rules and “Canadian” under another, depending on how the co-production is structured.

I looked at this topic a few years ago here Which countries most commonly team up to create film co-productions? As a good example, see below for the complicated relationship between movies made in France, Belgium and Luxembourg.

You could choose to follow the money. Countries contributing the majority of the budget are often assumed to be the film’s home. This logic drives many decisions around subsidies and eligibility for public funding. But modern films are rarely funded by a single source. Budgets may include pre-sales, equity investors, tax rebates, and soft money from several territories. Even identifying where most of the money comes from can be difficult, especially if financing is routed through holding companies or structured to meet treaty obligations.

The director’s nationality is sometimes used, particularly by critics and audiences, but is rarely decisive in formal classifications. The Oscar-winning Brokeback Mountain was directed by Ang Lee, who is Taiwanese, but no one questions its status as an American film. In contrast, Caché, directed by the Austrian Michael Haneke, was submitted to the Oscars by Austria but disqualified because it was in French. The Academy ruled that the language did not match the submitting country. This led to widespread criticism and a later rule change.

Language is another marker, particularly for awards. The Academy Awards’ International Feature Film category requires that more than 50% of the dialogue be in a non-English language. This rule disqualified Nigeria’s Lionheart, despite English being one of Nigeria’s official languages. Austria’s Joy, a film about Nigerian immigrants spoken in English and Pidgin, met the same fate. These decisions exposed how poorly the rules reflect the linguistic realities of post-colonial societies.

Sometimes the cultural content or aesthetic feel of a film shapes how it is perceived. A film may be classified as American, French, or Korean because of its themes, settings, or cinematic style. This kind of informal identification can matter a great deal to audiences and critics. Sergio Leone’s Westerns feel American to many viewers, despite being Italian productions. In other cases, films are labelled according to their narrative setting rather than their production background.

Then there is the question of post-production. Visual effects, editing, sound mixing, and colour grading are often done in countries that never appear in the credits as production partners. For example, a film shot in the UK may be finished in India or New Zealand, using remote teams and cloud-based pipelines.

Finally, the rise of AI in filmmaking is going to upend many of the foundational principles on which the current thinking is based on. If a shot is generated by a British person, living in France, on a Taiwanese computer, using a Chinese artificial intelligence model, trained on global training data, running on a Canadian server… then I’m lost already!

What are you trying to protect?

One aspect of the debate around country of origin is - what are you trying to protect? Three aspects to the film business could be prioritised in any scoring system of movie nationality:

• The economy. Movies spend vast sums of money and thereby contribute to a country’s economy, both directly (i.e., employing people, buying consumables, renting locations, etc.) and indirectly (i.e., the multiplier effect of the employees buying things, paying taxes, etc.).

• The industry. A product can employ thousands of people directly and via all the parts of the industry servicing the distribution, marketing, and release of those movies.

• The culture. Movies are a major cultural export, showcasing a country and telling the stories of its people.

Let’s quickly think about the challenges in building a system to protect each of these priorities.

Priority #1 - Following the money

A simple method for tracking money would be to look at the source of the funds. However, that would not achieve President Trump's stated goals, as Hollywood movies are mostly made with American money already - it’s just that they spend a lot of that outside the US.

Another method for an economy-focused system would be where the money is spent. This would have a heavy correlation with where the majority of the filmmaking was conducted, as between half and three-quarters of a film’s budget is spent during filmmaking.

While the UK does not pay close attention to the amount spent to determine whether a film is “British” or not, the figure becomes incredibly important when calculating how much of a tax rebate the production will receive. In response, US studios set up Special Purpose Vehicles (SPVs), which are British shell companies through which to spend the movie’s funds. This means that a British company is spending American money to employ people of many nationalities.

Priority #2 - Protecting jobs

Alternatively, if we were to value the industrial side of the film business (i.e., keeping jobs in America), we could focus on the nationality of the people who make the film and/or where they live.

Hollywood productions typically employ thousands of people in their creations, and they have a power law in their distribution.

When I examined this a few years ago, I discovered that 4.3% of all movie credits were for Marvel movies.

The challenge here will be determining the weight of each person’s contribution. Most nationality-based scoring systems for movies focus only on the leading cast, main creatives (i.e. writers, producers, and directors) and heads of departments.

Priority #3 - Protecting Culture

Hollywood has played a massive part in spreading the American ideal and values around the world. Therefore, you could set up a system which rates a film’s “Americanness” based on what’s in the film - the characters, stories, settings, etc.

Read more