Streaming Video Can’t Scale At Cable TV Quality, Will Never Replace Traditional TV Distribution

Almost five years to the day, the NYT published an article proclaiming that, “TV is becoming obsolete” and was joined by plenty of other media outlets claiming that within a few short years, streaming video could displace the traditional means of video distribution. While some in the industry still want to set false expectations that streaming media technology is somehow going to replace the primary means of delivering video to the living room, the fact remains that five years later, cable TV is here to stay and is still the primary way to get video into the home.

While reading a post on another site on a similar topic, a reader left a comment saying, “streaming is still in its relative infancy”, a false statement I hear often. Some seem to want to use the argument that streaming media technology is new and will improve as a reason why it will one day replace cable TV distribution. But in reality, 2013 marks the 19th year since streaming media was first used on the Internet. And while the industry has made advances and will continue to do so, the improvements in the technology over the last few years have been small. Compression technology gets better, protocols start to get standardized but year-over-year, we aren’t seeing huge leaps and bounds. That because today, the primary building blocks of all streaming media services, like storage, encoding and delivery can’t get that much better. Infrastructure vendors are spending most of their time to make these services more reliable and scalable and others are using these fundamental technologies to build out ecosystems, protect content and monetize it. But it can only go so far.

Streaming media has limitations and that’s not a bad thing, you simply have to apply it to the best set of applications as you would any other technology. But many are hell-bent on the concept that one technology has to replace another, when in fact, most times, one complements the other. Streaming media is never going to be as reliable, scalable or as high-quality as cable TV, even in the future. Those who suggest that do so as they don’t understand how streaming works and the limitations to the technology. Think of all the companies over the past 10 years who said their technology, be it better compression or P2P delivery etc. could solve the issues with delivering high-quality video to “TV sized” audience. They couldn’t. And even with guys like Netflix taking the smart approach of placing caches for free inside ISP networks, there is a limitation to what can be achieved.

In 2009, Akamai, the largest CDN on the Internet peaked at 7.7M simultaneous video streams during the Obama inauguration, for all of their customers combined. It was so much traffic on their network that they had to cap customers and they made it clear that there is no such thing as “unlimited” capacity on the Internet. Last year, YouTube claimed they did the largest live event on the web with 8M simultaneous streams. So three years in between events and the largest live event numbers didn’t grow that much. Of course years later, both Akamai and YouTube’s network have grown and their capacity improved, but the 8M number puts things into perspective.

CBS Sports said the NFL Playoff game (Broncos vs. Jaguars Ravens) from this past weekend had 35.3M viewers and peaked at over 40M. Even if half of those were simultaneous viewers, that would be nearly 18M streams, for one show. Multiply that times the number of shows going on at the same time and the numbers are huge. The Internet can’t and won’t be able to handle that kind of volume, now, or in the future, at cable TV quality. People love to argue the point, but the numbers, facts and data proves otherwise. When you turn on the TV, it works and you know what type of experience you will get from an HD channel. On the web, you don’t even know what HD really means as some call video HD when it isn’t, when compared to broadcast standards.

With streaming, there is no guarantee what the experience will be. Content services like Netflix and Amazon and CDNs like Akamai are working hard to give consumers the best experience they can, but they are working with technology that has limitations when it comes to scaling. Sometimes you can get good quality video with no problems. Many times you can’t, or you don’t have a consistent experience or there is some other computer/browser/player/app issue. There are a lot of moving pieces in the entire video ecosystem as opposed to cable TV which has very few.

It seems many outsiders want to judge the “success” of the streaming media industry and the technology on how quickly it can “displace” cable TV as a broadcast distribution medium. What a false idea. The fact is, the technology is already proven and for the right applications, we’ve already seen plenty of success stories. All streaming services combined, the size of the industry is in the billions of dollars, content owners are starting to get paid and business models evolve each year. That’s the definition of success. For those in the industry who understand this, don’t let those who try to set false expectations of streaming replacing cable TV distract you from staying focused and applying the technology where it is best suited.

  • Excellent article Dan, it is very accurate!!!! Thank you for writing it!!!!!!!!!

  • Anna

    Dear Dan, I have to respectfully disagree with your statement. “Streaming media is never going to be as reliable, scalable or as high-quality as cable TV, even in the future.”

    Today it has been proved that delivering video over the internet is more scalable, much less expensive and can offer better quality and richer experiences than cable.

    How scalable?: All current Cable users + Netflix users = 33% of total peak bandwidth at any time.

    How much cost effective?: at least 5 times less expensive. (Media Delivering cost)

    How much better quality?: Video: Better than 1080i and up to UltraHD. (Cable has not even a chance to flip 1080i to 4K) Audio: Up to 11.1 Channels if the decoder support it.

    How much better experience? There is new technology capable to change channels many times faster than cable over the Internet. On top of that, add all the interactivity added value by integrating this service over the Internet.

    Maybe you are referring to Live TV. But this niche may belongs not even to Cable TV but to over the Air signals. (Which are free). However I will not so be sure to say “never” after saw the tremendous progress by some company in Silicon Valley.

    • Hi Anna, your not comparing apples to apples.

      All of Netflix’s content is delivered over the public Internet. Cable TV is delivered over the MSOs private network. You can’t lump in Netflix + Cable TV subscribers together.

      As for the cost effective argument, you have it backwards. Every time something is delivered via streaming, there is an additional cost. With cable, it does not matter how many hours of video someone consumes as there is a fixed cost. Streaming is much more expensive.

      Quality wise, there is no comparison. Based on Netflix’s own data, their average video stream is delivered at just over 2 Mbps. Cable is delivered at multiple times that. I don’t know what you mean when you say “UltraHD”. While you say cable does not have a chance to go to 4K, neither does streaming. The demo Samsung showed with Netflix, of 4K streaming at CES, was not delivered over the public internet.

      There are enough facts and data in the market to show what is really taking place and streaming media will not replace cable TV as the main distribution medium.

      • Anna

        Dear Dan, You are right, we are not comparing Apples with Apples.

        I love your blogs and hate to be arguing with you, but this time, I have to disagreed with some of your conclusions.

        All I can say is that the word “Never” is never going to happen. What I saw a couple of months ago, changed the fundamental core and made me rethink of what is now possible.

        • It’s ok for us to disagree, different opinions is fine, it’s not really arguing.

          • Fernando

            Hi Dan, thanks for another excellent post.

            Sharing the concern Anna manifests (“never say never”), I’d like to ask you to comment this: “Multiply that times the number of shows going on at the same time and the numbers are huge. The Internet can’t and won’t be able to handle that kind of volume, now, or in the future, at cable TV quality. “. Wouldn’t the creation of a Multicast Internet along with the growth of Fiber to the Home change the game in the future and make it possible for for Live streaming events (not Netflix on-demand, more like presidential inaugurations and sport events)?

          • Taro

            That is assuming you need to receive and watch 900 channels simultaneously, right?

          • Fernando

            Why, Taro? Multicast allows you, the end user, to receive only the channel you’re watching, given enough last mile bandwidth. When you change a channel, you’d subscribe to a new multicast group.
            It’s just not a reality in Internet today as ISPs usually don’t exchange Multicast traffic, and when they do it seems it is still in experimental phase. The number of available multicast addresses on IPv4 (537 million) may not allow for proper allocation for all simultaneous live events in the world. But if one day it matures and become feasible, it could become the quantum leap that would allow a “near unlimited” capacity for live streaming events. So Dan’s conclusions are true for the current Internet, and only a “new Internet” could overcome such limitations.

          • Hi Fernando, for more than 10 years now people have talked about multicast technology, yet very few networks have actaully implemented it. It’s expensive for them to upgrade all of their gear and who’s going to pay for that? How do they make that money back? Multicast technology, while valid, has never been implemented on any scale. It’s been around since 2000 and in 13 years, has no real traction. Remember that the best technology is not what always gets adopted. When was the last time you heard any MSO, ISP or service provider talking about multicast technology? No one even mentions it anymore as no one is doing it.

          • Fernando

            I agree. It is still a distant dream.

  • Steve

    As much as I hate to say it, and I’m from the streaming space, the technology will never support the kind of video quality and scale that cable TV infrastructure can. In 2000 the average video was delivered at 300 Kbps. Today it’s 2,000 Kbps. That’s not a huge advancement in quality 13 years later. The data supports Dan’s conclusions and while many of us in the industry may want to argue with his point, because we want our businesses to grow faster, the fact is that Dan is right and is being realistic.

    • Taro

      If the reality is that we are only getting 2Mbps, regardless of the much larger figure my “ISP” has been selling me for years, then the FCC and the congress should jump to this case now. (You should be quite upset too! unless this is not affecting you)

      In the last months, I have been enjoying Netflix a lot. The video quality they have now is better than Cable and have no complains whatsoever. I just which that Netflix can offer a premium package option to get the “newest content” I want.

      • It’s not that the average ISP can only give you 2 Mbps, it’s that Netflix makes a tradeoff about how good the quality of the video needs to be, versus how much it costs them to deliver it. The higher the bitrate, the more money they have to spend. With cable TV, it’s a fixed cost and my MSO does not spend more money to deliver me a 720p channel versus one in 1080p.

  • I disagree with whats being said in the article because it’s based on data from several years back and makes linear guesses into the future while, ironically, evolution follows an exponential pattern. I’d never dare to say that something is “never” going to happen in the tech area.

    Besides that, there is the fundamental concept being missed: it is not only a technological change, what’s actually much more interesting is the cultural shift of actively choosing what you see and creating your channel rather than just making a selection from a number of programa being played at one time. The article misses the point that, thanks to streaming, providers will only need to stream what users want to see, saving resources that otherwise would be streaming content all the time but nobody watches. It is a behavioral change in how we consume content, and that changes the paradigm so much that again any comparison with data from the past is not applicable.

    • Hi Chema, the YouTube data is from exactly 3 months ago, it’s some of the most recent data the market has. Behaviors have changed in how we consumer data, but one set of behaviors has not replaced the other. People watch a lot of video online, but they still also watch TV. One medium has not replaced the other and we have the data to prove it. There are more than 100M households in the U.S. paying for cable/satellite TV service, with not even 1% of them cancelling their subscrption, over any one year.

      • While respecting your expertise on the subject, I find your fact highly questionable “more than 100M households in the U.S. paying for cable/satellite TV service, with not even 1% of them cancelling their subscription, over any one year.” Take a walk around campus/dorm or talk to twenty-something people, you’ll find out how many of them actually have TV (cable/satellite) subscription. They might have multiple channels on at the same time because they already have it (parents, dorm), but they don’t actually actively subscribe for it. This generation is the future, and whichever technology does not captivate the future generation is destined to end/shrink. I think your points should be taken in the short term, not the long term. In other words, you’re saying streaming videos are not overtaking as fast as we originally thought, not in a span of 5 years. This is important for investors because joining a market too soon is just as bad as joining too early.

  • I posted some thoughts on this article in response –

  • adamsb6

    I don’t think there’s a physical bottleneck that would prevent a transition to all-streaming all the time for content that isn’t live.

    Cable TV is already a set of hundreds of digital streams. Users can queue up HD streams from their cable provider at will. The bandwidth to deliver those streams to the home already exists. Some homes would need to have their cable TV bandwidth reallocated for use as Internet, but many already are capable of pulling down multiple HD streams in addition to digital cable.

    The broader Internet probably would not be able to handle the load of all TV viewers hitting distant endpoints, but that’s what edge caching is for. The most popular content would be cached nearer to the user.

    Live content gets a little trickier. Multicast isn’t reliably supported across the Internet. However, the cable companies definitely have some kind of technology to deliver a series of packets to many millions of digital cable users that describe the same live video streams. It’s most likely that they’re already using an IP network for delivery, given the period in which digital cable was developed.

  • The notion that video encoding technology and storage technology cannot get better is just plain wrong. New standards for video encoding efficiency continue to show big improvements:
    Ditto for storage, disk storage continues to double in capacity for equivalent cost every 14 months:
    The main limitation that video distribution runs into is the reluctance of residential ISPs to invest in infrastructure due to the monopoly/duopoly nature of list-mile wiring. There is no technical limitation that prevents improved content delivery.

  • Jakub Sroczynski

    ironically just read this right after reading the article above. If a given community can commit to investing in better hardware, or invite Google to do it (Check out Google high speed cable project in KC) then this will happen regardless of ISP willingness to play.

  • Is this just a technology issue, or is it at some point a business model issue as well? Seems to me even if, in theory, the delivery bandwidth problem were solved, you’d still have the challenge of measuring and monetizing the audience the same way the Pay TV industry (not just cable) is doing today. Online ad revenues are still a single-digit fraction of total ad revenues, online sub revenues are still a small fraction of total sub revenues, and none of the broadcasters seem to be in any hurry to ink exclusive distribution deals with online service providers. So there’s the financial motivation or ROI to build out Internet equivalents of bouquet services? Particularly now that MSOs are finally putting energy into multi-screen on-demand offerings of their own?

  • hootocol

    I remember saying that I would NEVER need to store more than my huge 640MB hard drive could store. I have to wonder if the author of this article is intentionally setting up some sort of “Remember when I said that? Boy, was I wrong.” scenario. I’m not suggesting streaming can replace cable now nor am I suggesting that streaming does not and will not continue to complement cable. I am, however, stating with absolute certainty that making statements suggesting that a newer technology cannot be improved to the point that it replaces an older technology is absolutely dumb.

    Something tells me that Dan has recently made a trip to the store to purchase film.

  • The author is 100% correct. The only thing that could fix it is if TCP/IP could run on a broadcast architecture. This is possible.

  • Cable TV and Broadcast distribution solutions needed in many markets, especially as the transition from analog to digital is still occurring and EAS are required

  • musingsofarationalmind

    My favorite part was “People love to argue the point, but the numbers, facts and data proves otherwise”… Sunshine you have little to no data, facts or numbers in here… It is one mass circle jerk with you saying something that has no source to back it up and then using that as a building block for a bigger un-sourced argument…

  • matt_the_katt
  • katie

    Lol I am reading this on my computer while in my living room watching a netflix original series have not had cable in a year and I never miss an episode of dexter

  • steve mankins

    To everyone praising streaming over provider based services keep in mind that your getting netflix, hulu plus etc through your provider…believe me, cable companies don’t care as much as you think about you cutting the cord, they will just up your internet and data fees. Not only that but ending in 2013 the big cable companies planned for 1-3% percent video sub losses and some saw as much as 5% gains. So the next generation as some put it is still plugging in.

  • The solution is massive capacity … Figure 300 million simultaneous viewers in the U.S. Figure 4x video using Google VP9 at 250kbs. Figure a single ‘head end.’ That is 75 Tbs. A single fiber today has 1 Tbs capacity today. A 100 fiber cable will meet the capacity requirement for the total U.S.

    And, of course, there will be thousands of ‘head ends’ so capacity is a problem no longer in search of an answer.

    A 1 Gbs to the wall system is widely estimated to cost $140 billion to go nationwide. That is $16.67 monthly over 60 months for 140 million wall outlets. Figure operating equal to capital, this is under well under $35 a month per outlet.

    The political battle to free the Internet from the clutches of cable and telecommunication companies is the real cost and the real barrier.

  • Real Logic

    Do you still stand by your extremely short-sighted view of this issue?

  • The quality of cable tv is much better as compare to streaming videos.

  • Leo Quiusky

    Do you think that with the constant technological advances made on the internet services, that it will be able to compete in the near future?