Cisco Says Video Traffic Growing, But Where’s The Business Going To Come From?

This week, Cisco released their second annual visual networking index which predicts that by 2013, all forms of video will account for almost 90% of the total consumer traffic on the Internet. While that sounds like a big percentage, I'd like to know how Cisco came up with these numbers. Cisco's website shows a lot of charts and numbers, but then says the source for the data is themselves. Where is Cisco getting these numbers? How did they come up with them? What are they based off of? What is their methodology? Come on Cisco, show us the data behind this. More importantly, if Cisco's predictions are right, that's not necessarily a good thing for the industry.

We keep hearing about traffic growth due to video, but what we don't keep hearing from are content owners making money from all this traffic growth. If Cisco's numbers turn out to be accurate, that's a lot of additional video traffic that someone needs to monetize unless they want to be though of as someone like YouTube. A company that pushes a lot of traffic, but can't break even. Don't be fooled into thinking that just because video traffic is growing, so too is the revenue of the companies in the space. Video now easily accounts for more than 50% of all the traffic on the CDNs, but it accounts for far less than 50% of their overall revenue. That's a problem.

For all the traffic growth we keep hearing about, we also see companies like BT and others capping users or telling content owners they are going to have to help foot the bill for network upgrades. We've now got content owners doing 3MB streams and folks like Microsoft announcing they will do 8-10Mbps 1080p streams on the Xbox 360 later in the fall. All of this is great, but there needs to be some kind of business model behind this surge in traffic. If it's not figured by the ISPs and content owners, many are going to have a false sense of security thinking that because video traffic grows by a large percentage, the revenue of companies in this space will grow by the same volume. That won't happen. While the CDNs won't tell us how many total streams or GBs they deliver each quarter or each year, if they did, we'd see that they are delivering five or ten times more volume, yet their revenues are not going up by the same rate.

I'm all for video growth, but lets be realistic when companies like Cisco put out numbers like these and won't show us how they came up with them. Traffic growth without revenue growth really does not matter.

  • But how will Cisco’s investors see a future for Cisco without sales of gear to support all this growth in video traffic without all this growth in video traffic due to lack of a business model to support its cost?

  • Mark Dantche

    Their Crosscheck data seems pretty good, Youtube, comscore, etc. Also, if you take their data from 03 to 08 they were pretty accurate based on looking back at real CP traffic.
    But in the end isn’t all IP research done by Cisco the same as Cancer research done by Philip Morris?

  • Cisco says the date comes from, and I quote “The forecast is the result of (1) a model based on third-party analyst forecasts and (2) data we gather directly from service providers.”
    That’s great, but who are the analysts? What are their forecasts? How many of them are used? And how many service providers is the data collected from? And what is defined as a “service provider”? And what region of the world are the service providers in?
    Cisco might be right on all of their projections, but I’d like to see the underlying data and methodology, with details, on how all of this is collected and who it is collected from so I can make my own informed decision on the data.

  • You have to look at it from the back to the front. A company poses the question “how much do we need to be selling over the next 5 and 10 years to cause an increase in share price?” They calculate that number, and then work backwards to derive how much video traffic would be required to support these sales? Then analysts are hired to create a defensible forecast of traffic, which then justifies the sales. I’m not saying Cisco in particular did this, but it is very very common…

  • Steve, that’s a good one! I totally agree.
    That forecast is based on the past and today’s data. But no technology improvement is taken into account. If someone uses good old MS WMV9 at, say, 900kbps, and switches to h264 flash, he can achieve comparable quality at around 600kbps. Today h264 owns the online video. But, for example, still there is the Dirac codec ‘undiscovered’ on the Web. Not to mention the p2p tv’s, which utilizes the user’s bandwidth, not yours only.