The App Revolution (in Filmmaking)

The following article is published in Filmmaker Magazine. Fall 2012, Vol. 21 #1.

There is a saying I once heard: “Once you change the method of distribution, the product has to change.”

This itself is a take on the idea that distribution defines the product. You see this around you every day in the products you buy. Cars are influenced by the dealership networks that sell them. Phones by the mobile network operators and the choice of computer you use at work by whatever the IT department or value added reseller prefers to work with. Mass market restaurants offer what can be sold by wholesalers–typically frozen, long shelf-life staples. Almost every product category is shaped more by what can be distributed than what can be produced. That’s simply because in mature economies distribution is harder than production. In consumer products it requires access to wholesalers who then require access to shops who themselves have access to prime real estate which attracts foot traffic. Production only requires capital. Distribution requires relationships, often exclusive ones.

This pattern is even more pronounced when looking at media products. Production is arguably easier since it’s constrained by talent, which is fungible. But distribution is even harder as it is addressing bigger audiences in shorter time frames. You see this lopsided balance of power in the abundance of books being written relative to those being published. There are thousands of films produced and hundreds get distributed.

But the saying suggests that if distribution were to change then the product itself would change. Indeed, if you can sell ebooks direct, then they tend to evolve into new genres (e.g. Fifty Shades of Grey). If you can sell cheap adult video online it tends to evolve into new genres as well (I’ll leave examples to the imagination.) YouTube videos quickly cluster around “Fail” or “Win” compilations which evolved from America’s Funniest Home Videos. They get millions of views. Even before the Internet, the availability of cable created the genre of music video, which created the first music broadcast alternative to radio. And of course, cinema itself redefined theater once it could get shown to millions rather than thousands. The new methods of distribution of media affected what gets produced rather than the other way around. Consider the converse: innovative filmmakers who try new storytelling methods are stymied by a lack of acceptance by existing distributors and find their material languishing in festivals or perpetual cult status.

So we can re-state the saying to a new “Law of new media”: Once you change the method of distribution, the medium itself has to change.

I gave examples of how Internet distribution changed some media formats. However, the Internet itself has not had any effect on long-form cinema so far. In fact, you would be able to tell if Internet distribution did work by watching if the product would change. So far, it seems, at least to me, that the commercially successful product has not changed. One reason has been that film is still a very difficult product to distribute digitally as most of the world does not have the capacity to buy, consume and enjoy films digitally. Another is that the current distributors have had an allergic reaction to the option. As existing products have been crammed into the new distribution chain (e.g. Netflix) the established distributors are experiencing channel conflict and are raising the costs of digital distribution rights. Netflix has seen its costs triple as it ate into Cable market share. Fundamentally, the incumbent distributors are reluctant to open a channel of distribution which conflicts with or reduces margins of the existing channels.

This brings up another corollary: fundamentally, you can’t move an existing media to a new network. You have to think of it as a deeply rooted-in system, and it’s just not going to like moving to another environment. You have to uproot this huge tree, and it just won’t come out and if it did it will not take root in a new place.

The only way to create this new value network around new distribution is to plant a new tree.

This is the problem for the technology industry today. The only way technology companies can solve the “content is king” dilemma (which, by the way, is an euphemism for “Hollywood is king and Silicon Valley isn’t”), is by allowing new content to flourish on top of new platforms. It’s not about moving the old content to the new platform. It’s about allowing new content to be created exclusively on the new platforms. This new content will therefore, by our law, be a new medium.

For me, apps are that new medium. Apps offered a new way to package entertainment for “niche time”. The early fart apps evolved into a whole new genre of games which are now challenging the established game studios. The result is that a generation of developers has moved from building PC or web software to apps with cloud back-ends. The first to switch to this new medium were those with the least to lose and with the most modest of goals. Apps have now a more valuable industry than digital music with over 55 billion installs between the iPhone and Android ecosystems. Consumers are increasingly spending their spare time in front of apps. In the US “app consumption” increased from 43 min/day in 2010 to 94 min/day at the end of 2011. This number already exceeds the time spent in web consumption. I find it difficult to doubt that app time will exceed TV time, especially as TV time is increasingly a dual screen experience.

(Feels like it needs one sentence explaining why. Is it because of their accessibility to small developers, or their ability to reach mobile handsets natively, or because of consumer trend?) If it’s absolutely impossible to move long-form video or long-form cinema to this new medium in such a way that everyone in the world can consume it on every device, then the app ecosystems will take up that challenge and say, “There’s a job to be done, and that’s to enertain people for an hour-and-a-half with characters, plots and wonderful stories, and we will make that happen, and we’ll call this thing an app.”

And that app will solve that job for that person. It’s what’s happening to the game console business. The new game apps are distributed differently, priced differently and produced differently. As a result they are used differently and make a different set of developers thrive.

The new app media will mean that we’ll just have to live in app universes, and go eat popcorn somewhere else. Microsoft and Google and Apple and Amazon and anybody else whose belly is at the bar has been scratching their heads wondering “How do we move that content over to us?” They’ve been barking up the wrong tree. The fundamental question isn’t how to cram old media into a new shiny screen. It is how do you let developers and artists and creative people create stuff that is new and compatible with this new way of living.

Think also about the history of media — movies haven’t been around forever; music hasn’t existed as a recorded medium forever. Not even books. What did people do before these forms of expression? We think of the use cases of going to the movies, or the use cases of sitting in front of a TV on Sunday night, or playing a console game. These things didn’t get done at some point in time. Nor will they be done at some future point in time.

In a few years we’ll move on and we won’t do those things. Each person will probably have his own screen in front of them–many already do–and whatever we do together we’ll do through a shared screen. We won’t have communal experiences that we associate with the old media. Certainly not in a theater. Maybe not even in a living room. And that is how we will move into a new value network.

It is inevitable. As a technology provider, you shouldn’t try to squeeze a 19th-century model into a 21st-century model. Just let it go. Create a platform and give people the freedom to create on it. As a media creator, you should try to learn what the new model demands. New forms of interaction, new story telling opportunities, new ways to communicate with audiences. Create experiments and learn. It’s the only way.

  • I started conversation with my students this semester by asking them, “Where are you online?” Reading this post, I realize that not once did we discuss apps. I look forward to asking them about this in the coming weeks. I was also shocked that about 80% of them had seen The Dark Knight Rises in a movie theater.

    I think you’re just getting started with this, but I’d like to read you digging in more specifically about the making of a movie and the ways apps would bring me the film. Are you suggesting I’d experience movies the way I listen to your podcast? That’s the primary way you see movies being experienced? And, I think you’re saying the 90-ish minute film won’t be something future consumers will watch as opposed to 3 minute content?

    Thanks for your good work!

    • Here’s an approach that I use right now:

      I buy training tutorials on various subjects… lasting 1-3 hours.

      Most of these are delivered as a series of podcasts which are downloaded to my computer (iPads).

      Now, you can treat these podcasts as a single “movie” with chapters for the various topics — typically 10-20 chapters.

      So, you can watch it as as an OD “movie”, but have the ability to pause it at any time grab a snack, whatever.

      The important thing is that You can be watching one chapter (or stream it to your iPads) while other chapters are downloading in the background.

      I think you could take the same approach with real movies delivered from the Internet:
      1) Download, save to disk (buffer) and begin playing the first few chapters and/or stream them to other devices over WiFi
      2) Download and save later chapters in the background

      This is, somewhat, similar to the way iTunes works for streaming movie purchases. But, I think a more flexibility is needed — and political issues with the content owners need to be resolved.

  • Great article! Thanks.

    Maybe (and I’m being quite simplistic) applications like iBooks Author is what we need to let everyone produce new contents in new forms.
    Of course, iBA is just a step from textbook to mediabooks and is very structured to simulate paper textbooks. But the result is not suitable for the old distribution type.

    Also, the “app usage” should be separated in different categories, because not everything is media consumption.
    It is like actualy saying “computer usage”: an iPod is a computer, a mainframe is a computer…

  • Minor: in sixth paragraph from bottom, is the following sentence supposed to be there 🙂 “(Feels like it needs one sentence explaining why. Is it because of their
    accessibility to small developers, or their ability to reach mobile
    handsets natively, or because of consumer trend?)”

    You say “We won’t have communal experiences that we associate with the old media.” Movies will survive, even if the distribution and format changes. Sports and drama shows will still be on something that will look like TV. So it’s not one-or-the-other, but the old will survive alongside and mixed in with the new.

  • To drag out one of my favorite hobby horses here, I think there’s a less-obvious implication to the new media distribution model which is akin to Horace’s comment about restaurants. The Internet distribution model has technical and economic constraints around bandwidth, especially in mobile use.

    I believe this will exert a pressure on entertainment media towards categories that favor lower bandwidth requirements and, for mobile, intermittent and unpredictable connectivity. I expect to see some selective pressure exerted against traditional streaming media, especially video streaming, because of this — it’s a case where the media was developed for a broadcast, not an individual, distribution channel. And it’s not a particularly good fit for the one session, one user model of Internet use.

    What does fit well into the new model? Text, because it’s very low bandwidth. Web content that’s not too heavy on single-use images. Any media where a bandwidth-expensive resource can be sent once and used many times: this includes music files, games or many different types, and possibly magazines (since print ads are often both graphically intensive and reused).

    And then there’s my personal dark-horse favorite, computer-generated animation, where characters and sets can be downloaded once, and used many times. This is best seen currently in 3D video games, which generally boast higher-resolution imagery than HDTV (though the models are not currently as detailed). This is an area that Moore’s Law favors as well, GPU power has been increasing rapidly, leading to more and more realism in CGI generated on the user’s end system. And Pixar and Dreamworks have shown that it’s possible to do impressive (and profitable) storytelling using something close to this model, though they still render the imagery at the source end and distribute it via conventional channels. Essentially, CGI is the ultimate video compression technique — it compresses not only visual images, but the entire 3D world into a relatively small number of bits representing the models and their states. No, it’s not going to replace true cinematography… yet. But for entertainment purposes, it may be “good enough”, and have other substantial advantages versus streaming video.

    • KirkBurgess

      The cgi technique is interesting. It could actually become an absolute killer feature of live sports viewing. Imagine if live sport can be tracked in real time and a user is able to select any camera angle they wish as at their end its a photorealistic 3D recreation.

      • televised sport will all be cgi soon enough

  • Sam Penrose

    Among your very best. Thank you.

  • Good article!

    Couple of observations:

    The entry cost of producing a “movie” has decreased exponentially in the last few years.

    As an example, the cost of editing a movie 10 years ago was between $25,000-$125,000 per seat.

    Today, you can get a professional editing software for $300 per seat.

    Cameras were very expensive — today, something as inexpensive as an iPhone camera can be (and are being) used.

    End-To-End Digital has already been done!

    All digital: capture; edit; distribution and presentation with no physical media.

    The Girl With The Dragon Tattoo was shot in super hi-res 5K, then edited and delivered in super hi-res 4K. This was digitally distributed to theaters who had the capability to present 4K.

    At the same time, sub-scale versions were made for 1080p, Blu-Ray, DVD and Academy format.

    The film was shot in Sweden and the UK. Sub-scale dailies were sent to Hollywood via Podcasts. (5K/4K was sent through the mail, and arrived a few days later).

    Have a look at this — it is very informative and entertaining:

    Michael Cioni prepping for a 4 K World

    Michael has also developed some iPad apps to present Dailies and Video Assets on iPads via WiFi.

    Digital Delivery Bandwidth Issues Can be resolved!

    I am a firm believer in “necessity is the mother of invention”.

    When I was growing up (and into my early adulthood), airline travel was quite different than it is today. If you wanted to fly from, say, Minneapolis to New York City — you took a direct flight. Like a train, the plane went directly from Minneapolis to NYC (you didn’t need to change planes or trains). Also, like the train, the plane would stop at various cities between Minneapolis and NYC… say, Detroit and Pittsburgh — but you had the same seat on the same plane.

    As the industry evolved, and air traffic increased a “bandwidth” problem arose — the pipes (planes) were under utilized in some areas, and backed up in others.

    To address this, the present “spoke and hub” system was designed — where smaller planes from outlying areas fly into major hubs, and the passenger changes planes to go to another major hub or outlying area.

    Mostly, this works pretty well — the pipes (planes) are full and the bandwidth [usually] efficiently handles the traffic load.

    The reason I described this, is that the digital distribution pipe/bandwidth issues could be addressed by a approach similar to the spoke-hub solution. High-speed backbones would be used to distribute content to central hubs — where they could be streamed to multiple lower-speed spokes.

    In addition, at a slight loss of convenience — a single sharable stream could be started at predetermined times (say, every 30 minutes) and multiple viewers would watch the same stream — analogous to broadcast TV or a cable schedule.

    My points in all this:

    1) The cost of entry to digital video is minimal
    2) Digital has arrived
    3) Technology exists to resolve the one-to-one * digital distribution and consumption

    * Or, as illustrated in the classic “Naked woman – Engineer” problem — we can get close enough!

    • Indeed, your mentioned approaches to digital distribution (cacheing/distribution hubs and multicast) have been thought of and tried, but for some reason, haven’t caught on, except in some niche areas.

      Multicast has both technical issues (what happens when a packet is dropped on one path but not another?) and economic ones (the carrier doesn’t know how far out the multicast will spread, and it’s a bandwidth multiplication effect that can easily clog the network.

      The hub and spoke, or more generally, hierarchical re-transmission, design has been more successful, but it doesn’t scale terribly well to millions of users in arbitrary locations. And it still overloads the last mile to the user, which is where the most expensive bandwidth bottlenecks usually are. As a rule in the Internet, core bandwidth is *much* cheaper than edge bandwidth; there are efficiency of scale effects in play there.

      I’m not saying the problems are unsolvable, but most of the likely solutions have been tried and not gained much traction for one reason or another, often one that’s not obvious until the approach is scaled up. And we’re talking about very large scale here…

      Right now video streaming is working mostly because the number of active users is a relatively small fraction of the total user base — but it’s already using something like half of the core bandwidth. The core load issue can be fixed by the hierarchical redistribution model (indeed, Akamai and others already serve this sort of function), but this doesn’t work at the edge without massive investment (i.e. every DSLAM and cable head-end has to become a redistribution node, and even that won’t be enough for the cable case — you also need to deploy new bandwidth).

      The other elephant in the room here is mobile, where you need spectrum to move bits. Spectrum is not an abundant resource like fiber capacity, and you can only increase its bandwidth by moving to new, higher frequency bands (technologically tricky, but progressing slowly), or re-using existing frequencies in smaller and smaller cells. But that greatly multiplies the investment in cell hardware.

      So I don’t think the bandwidth problem is quite as easily solved as you assume. It’s not intractable, but it may take a lot more investment than people are willing to pay for, to get to that content. The costs *will* come down some more, but I’m not sure they will easily drop much more than another factor of two or so.

    • Walt French

      I hope that the media distribution effort isn’t as weak as the airline hub’n’spoke system is.

      Experienced flyers almost always avoid connecting flights when possible, since they add another hour (bare minimum) and > $50/hour incomes are the norm, so the apparent cost savings quickly disappear. And there’s always, as @WM mentions, the risk of packet loss, which in flying means a greater-than-doubled risk of cancellation/delay preventing you from where you want to go.

      But isn’t this what CDNs such as Akamai are built to do? Don’t they have arrangements with all the major ISPs to inject the content at one high-efficiency location into any one ISP’s regional net?

      Methinks the delivery mechanisms are not of the technical variety, but licensing, coverage and critical mass.

      • The CDNs really only offload the replicated traffic from the core backbones; they do nothing about the bandwidth issues on the last mile. So they do indeed help, but there are still significant engineering issues. Some of the access ISPs have talked about putting CDN caches in the last-mile headend nodes, but this is a massive undertaking, and I think the economics (and control issues) in the proposals have made this approach fail, so far. Technically, it would help, though.

        But even with head-end caches/redistribution nodes, there would still be scaling problems with last-mile cable infrastructure (on the user side of the cable headend), at the least.

        The whole problem is a multi-dimensional mess of economics, contracts, and technical issues, and optimizing across all those factors is indeed very hard.

      • KirkBurgess

        The bandwidth issue for some forms of video can be alleviated for some media if its content a viewer is likely to subscribe (episodic content, pre ordered movies, podcasts), as the media can be downloaded at a slower rate automatically and stored on the viewing device.

        This would leave more bandwidth for live streams.

      • That doesn’t decrease total bandwidth, if the content is only viewed once, but it will help the peak bandwidth required in the network, since the content can be downloaded at off-peak hours. I suspect there would need to be some modest economic incentive to encourage users to declare the desire to pre-download, and not just view stuff on immediate demand. Since networks are built to peak capacity requirements, and off-peak usage is much lower (you’d gain something like a factor of 3 or better, I think), this could be a major improvement.

  • raycote

    Horace McLuhan:
    “the distribution medium is the production message”

  • WFA67

    Of possibile interest:

    “Gross Misunderstanding: What journalists miss about the movie business”
    – Edward J. Epstein

    “…today, the six major studios get less than 20 percent of their total revenue from showing their films in American movie houses. Most of their money comes from another, nearly invisible source: licensing their intellectual properties.”

  • IBooks and Newstand has shown that the ‘app-content-distribution’ thesis is not complete. Just like in real-world a central location can increase activity. Like a mall (or a known retail epicenter) increases consumption.

    The traditional dynamic is definitely shifting, and it seems like its shifting to a more real-world dynamic.

  • james

    Is this attempt at “blending” the platforms the
    reason we see so many cable/satellite companies releasing apps that provide
    their customers with on-demand AND TV practically anywhere they go? Is
    this an attempt to keep all the branches on the same tree, as you put it?
    I work for a major satellite company, and when my DISH coworker friend suggested
    I get the free DISH Remote Access app I was hesitant. I downloaded it,
    and with my Sling Adapter I can access my entire TV system anywhere I can
    connect to 3G, 4G or WiFi. DISH also added the Blockbuster @ Home movie
    package, with over 100,000 titles. Is this conglomeration of media,
    through one provider, a positive response to the problems you’ve mentioned
    above? I know the customer benefits, but where do studios and distributors
    fall into this changing landscape?

  • oases

    Brilliant article.

  • oases

    Brilliant article.

  • Pingback: The App Revolution (in Filmmaking)()

  • Pingback: Aktuelles 14. Mai 2013 —