I’m Neil Hunt. I lead the Product Development team at Netflix - that’s the team that creates the technology that delivered 6.5B hours of streaming entertainment worldwide in Q1. Today, I’d like to talk about what I think TV will look like in the future - perhaps up to a decade out (I don’t think my crystal ball is very good beyond that)
Internet TV means watching what you want,when you want, andwhere you want it.
Current TV is limited by the tyranny of the grid --where there are only 21 prime-time hours a week... Any show that occupies one of those slotsneeds a huge audience to pay it’s 1/21 share of the infrastructure of the channel. But any show NOT in prime timehas a hard time connecting with an audience at non prime-time hours.
But with Internet TV, no appointment is necessary - fans can enjoy a show at any time. The fixed cost to carry an additional show is small, since it doesn’t have to displace something else in the grid. That allows for shows to be created to a more diverse set of audience tastes. Who would have thought that Lilyhammer, a show shot with half the dialog in Norwegian with subtitles, could attract a significant American audience -- it certainly couldn’t have on a linear channel In the UK, Breaking Bad was unsuccessful attracting an audience on linear TV (broadcasters tried twice), but the show was a blowout success on Internet TV. (I’m happy to add that last night, we learned that Netflix has won a BAFTA award in the UK for “Breaking Bad”)Internet = No Appointment Necessary
You might argue that escaping the tyranny of the grid is a property of On Demand TV, not Internet TV But Cable On Demand has struggled to be successful because the discovery process -- finding what to watch in a large catalog -- is hard without a rich interactive UI and a direct relationship with the user.
Internet TV at scale generates a rich stream of “big data”: I’m making it up: People who watched West Wing later watched House of Cards A subset of the people who watched Breaking Bad bailed out disproportionately at minute 14 Those who quit before minute 5 statistically were likely to enjoy Walking Dead weeks later Some of the audience who loved How I Met Your Mother likely share the household with someone who loves Turbo… We can use this data to learn what will appeal to each viewer.
Our vision is that you won’t see a grid; nor will you see a sea of titles;you’ll just see a few choices that are perfect fits depending what mood you are in and who is with you. I think this vision is possible; But I think we are only a small part of the way to getting there… At Netflix, we put as much effort into building UIs and personalization technology to get to this goal, as we put into delivering the bits to a streaming player. We are making incremental progress each month:more than 80% of what users watch has some element of algorithmic choice behind it.
A common criticism I hear is that we carry lots of old shows. Some call them junk.
I see it differently. There are no bad shows, there are only shows with small, but devoted, audiences. Internet TV can afford to carry many small shows, with a small budget, that reach a small audience, because on-demand can aggregate the audience over time, and distance, without the grid. That’s why titles you’ll never see on linear TV -- because the audience is too small -- are available on Internet TV Personalization helps find the audience passionate about niche show, and puts it in front of them when and where it makes sense. And the smaller productions -- like The Square -- an important documentary about Tahrir Square and the Arab Spring -- can find their way to an audience that might have been impossible to reach in a linear world.
TV 1.0 was broadcast using radio waves to antennas on the roof, so it was kind of hard to get people to pay for it - which means that it was all supported by advertising. Even now that TV is mostly delivered via cable or satellite systems -- that can collect payments, most content still carries the ads, because the same channels are still delivered free to air. Internet TV is finally divorced from the broadcast technology, so that we have the choice to offer consumers - ad-free with subscription (like Netflix), free with ads (like Hulu), or a hybrid model. I’d have to say, many consumers seem to love the ad-free model, so perhaps the Geico, Zales, Chevy, and Wendy’s of 2025 will have found a new place to market their products!
Or perhaps not, because free is powerful. Technology can work here too -- the same personalization that can find the perfect show to put in front of you may be able to show the Chevy ad ONLY to the people who might be enticed by a new car, and charge higher CPMs to make up the difference.
The pre-modern solution to personalization is channels. CNN for newshounds. MTV. History channel. Golf channel. Military Channel, Etc. Aggregating content that appeals to a specific demographic helps the audience discover other relevant content, but it’s very one dimensional.
Or perhaps it’s more appropriate to say “a different channel for everyone” Personalization lets us create 48M “channels” for 48M users -- mixing and matching some kids content (with the right educational tilt), some docs (about specific topics), drama leaning towards particular actors and directors and so on. Personalization is a big deal, and something we don’t really expect, so it’s hard to appreciate the impact it can have - and is already having!
As channels have less relevance, so too do bundles of channels. Bundles make sense because they spread the cost and friction of managing the infrastructure and the billing relationship across groups of channels that appeal to clusters of users. But the Internet reduces the costs to the point that there is no saving to pass on to the consumer any more. Unbundling will be painful for the incumbents, so it won’t happen quickly. But the competing Internet TV businesses start from an unbundled structure, and there won’t be anything that bundles them up, so the power of the bundle is likely to reduce gradually over time.
TV used to be tied to the antenna on the roof. Then to the cable box. Both wired to a big glass tube in the living room or den.
Nowadays, TV is untethered within the home by Wifi, and untethered outside the home by SmartPhones and mobile devices. Wireless bandwidth will improve dramatically with new technologies, much smaller cells, so we will be able to watch TV while waiting in the security line at the airport. (Sadly, I also expect that line will probably be much longer than it is today). Perhaps some of you are old enough to remember when Bluetooth headsets were brand new, and folks walking around talking to themselves in a loud voice would seem kind of wierd Well, I think you can look forward -- sometime soon -- to seeing someone staring intently into nowhere, wearing their Google Glass, and enjoying Orange is the New Black. (Hopefully, they aren’t driving at the same time…) In 2025, it won’t seem so strange.
Well over 100M Smart TVs will be sold this year, bringing the installed base up to quarter of a billion, one per three Internet connected households, and by 2025, finding a non-smart TV might be as hard as finding a Blackberry phone today.
The interesting question that begs an answer is: “who will own the main Smart TV interface?, and who’s apps will be featured?” Your cable company would love to own it - keeping its bundle of channels front and center in your home. Any number of Internet TV services would love to own it - Netflix sure would… Your TV manufacturer would enjoy owning it - and finally gaining a share of the recurring revenue that they have for years enabled, but not participated in. And of course, Google wants it, because it’s a part of the world’s information that needs to be organized (and monetized). And Apple wants it, because it could be so much cooler, and there are 97 buttons to be cut from your remote control. I’ll stop short of predicting a winner here, but expect some interesting competition in this space.
In the US, Netflix accounts for about a third of the last mile downstream traffic But that is sort of missing the point, because Netflix averages 2Mbps, while the typical consumer pays for 8, 10, 25 Mbps -- so Netflix is more like 1/10th of the bandwidth that the consumer has purchased.
Now the typical Netflix traffic stays off the Internet highways, travelling only the last few miles on ISP connections, because we already bring the data to the ISP wherever ISP wants to take it, from a topological point of view. The Net Neutrality debate is about whether your ISP has the right to limit or prioritize some kinds of traffic, or whether it must provide a best-effort service for everything within the physical constraints of the pipe. I think that you the consumer have paid for 10 or 20 Mbps of last-mile capacity to your home, for your ISP to charge content owners as well is double dipping. If there were competitive services to your home, the market could take care of disciplining IPSs who overcharge or under provide, but for most users, there is no choice, no competition, and so for Netflix, there is no alternative route to reach you.
So the right answer is for the ISP to accept the data – that it’s customers have requested and paid for -- at zero cost, and deliver all traffic with “best effort” at equal priority – no fast lane, but no penalization.
I don’t know which way it will work out, but I would observe that the future of TV involves a lot of trafficover pipes that are owned by the same businesses that drive the present of TV, who have a so-called “terminating access monopoly”. If we want a vibrant market with innovative new businesses helping shape the future of TV, we should be thoughtful about managing the gatekeepers. Well that was heavy, let me move on to a less controversial topic...
Novels are long and short, have chapters of different lengths, and have arcs that develop characters, plot lines, and stories in nested structures. Feature films vary in length and structure too, although the length tends to be between 90 and 180 minutes. But current TV storytelling is almost universally in 13 weeks of 21 or 42 minute episodes, where each episode closes out one story arc, and teases something new to bring you back next week.
Internet on-demand TV need not be bound by these conventions. When fans can binge several chapters in a sitting, the story can be richer and longer, and doesn’t need support for being viewed stand-alone -- it can assume that everyone saw the last chapter, and doesn’t have to fill in the gaps. Chapters can be longer or shorter, and groups of chapters can carry a single story arc with expectation that it’s all viewed in close proximity. Today, House of Cards and Orange are only just beginning to take advantage of these opportunities. But I’m optimistic that this will pull TV story telling forward and lead to something incredibly richer and more interesting. The stories we watch today aren’t your parent’s TV, and the stories your kids will watch in 2025 will blow your mind!
It’s an axiom of disruptive technology that new technologies often start cheaper, but less capable than the incumbent technology. So with streaming TV - it was poorer quality, slow to start, and rebuffered often. But technology improves inexorably, often following Moore’s law of twice as good every 18 months -- too slow to see it happen, but so fast that every day something seems new and different. So one day you wake up and realize that Internet streaming TV is actually ahead of the previous technology.
4k is double-double high definition - twice as many pixels wide, twice as many high. And 4k TVs are here now, affordable soon, and ubiquitous in a year. But the Internet is just about your only source for 4k content at home - While some cinemas have 4k projectors, I don’t think packaged media will go through another blu-ray transition, and the broadcast industry has a lot of infrastructure to upgrade to support 4k. 4k is still new. The cameras will evolve, and the sets will upgrade, and the DPs will develop new skills. So just like HD got better after it first appeared on the scene, 4k will improve over time.
But Ultra High Def isn’t just more pixels, it’s also better pixels. A development I’m excited about is extended dynamic range or high dynamic range. Content for your TV screen today is mastered for low maximum brightness, so the highlights are compressed to keep most of the picture visible. But the technology is developing to produce much brighter pictures -- thousands of nits instead of 100s -- and the result it a tremendous improvement in the realism of outdoor scenes - sun reflecting of water that makes you want to reach for your sunglasses, and bright colors in the highlights. Today the state of the art HDR TV is water cooled, and not really a practical consumer appliance. But we will see the first consumer TVs in 2015, and I think it will amaze you, because delivering high brightness is an area we haven’t addressed for decades, while we’ve moved from VHS 300 pixels to SD 700 pixels to HD 2000 pixels to 4k.
In the quest for realism, the next frontier is high frame rate, which gets pictures with less judder and smoother motion. Sports broadcasts are already all 60 frames/sec, and many TVs interpolate the 24 or 30 frames we use for drama up to 60 or 120 or 240 fps. We can do a much better job if we film at 60 frames from the start. This faces lots of resistance -- “it feels too real”. I think I can hear the same voices who declared that color is too real, and hearing the actors voices was too real...
The final frontier for display technology is to utilize the full field of human vision. Today’s TV occupies 30-40 degrees in the center of your field of view - that part where you can see detail clearly. But you have an additional 100 degrees of view, where your retina has different kind of sensors, that perceive motion, not detail. The next step is to shoot IMAX for home - where the story happens on the 40 degree high resolution screen in front, and ambient motion and immersion happens on a lower resolution screen wrapped around the whole 180 degrees in front of you.
I’ve talked about a lot of changes, some of which you can experience today, some of which seem far out. To paraphrase John Wanamaker, “half of all my predictions are going to be wrong, I just don’t know which half” But I think I can rely upon the prediction that TV of the future will be different, better, in ultra high definition, and delivered by Internet
The Future of TV - Neil Hunt - Chief Product Officer Netflix - Internet World 2014
Chief Product Officer, Netflix
(Presented at Internet World, NY, 19 May 2014)
The Future of TV
What you want
When you want
Where you want
Linear TV is Ripe For Replacement
The Tyranny of the Grid
Linear TV is Ripe For Replacement
The Tyranny of the Grid
Internet = No Appointment Necessary