Cisco’s broadband bikini revisited

27 Jun

A few weeks ago, I wrote about the latest internet usage report from Cisco and how it would inevitably be misconstrued by some folks.

Some of the issues I identified with the network equipment maker’s Virtual Networking Index included the unlikelihood that Canadians use the internet more than Americans; Japan’s similarly head-scratching low usage; the unexplained projections of how traffic is expected to grow in different countries; and why internet leaders such as the Nordic countries were grouped into a generic “Western Europe” category that may have skewed the entire study.

Of course, the data was indeed misconstrued by some pundits, including The Economist and Duncan Stewart in The Globe and Mail, who claimed that “Canadians are the third biggest consumers of gigabytes on the planet.” A few days ago, I spoke with Thomas Barnett, senior manager of service provider marketing for Cisco and Arielle Sumits, lead analyst and prime architect of the VNI about some of the issues I raised. Here’s an abridged version of that conversation, which was marked by some illuminating information, followed by some concluding thoughts.

The Economist’s chart showed that Canadians use more data than Americans. Is that correct, and if so, how so?

Sumits: That was actually a per capita chart. They didn’t exactly make that clear in the captions, they said per person. Per capita does change the results. I think the way you did it in your article, looking by households, is more accurate. If you look at per user and not per capita, then Canada ends up below the U.S. It’s still higher than you might expect but there is some other evidence that Canadian usage is pretty high. Comscore has been measuring Canada versus the U.S. for some time and they claim that Canadian usage is twice that of elsewhere in the world. Higher video minutes, particularly.

Americans have considerably more and better online video options with Hulu and deeper selections on Netflix and iTunes. How is it they end up using only slightly more than Canadians?

Sumits: It is a little bit surprising. A lot of the sites that are driving traffic in Canada don’t appear to originate in Canada. There are a lot of video sites particularly in Asia that provide a lot of video content that do appear to be driving quite a bit of traffic in Canada. Even though the legit content availability is lower, users [in Canada] do appear to be finding other ways to get content.

Barnett: Canadians, from an internet perspective, have been some of the most prolific users of Facebook and YouTube. One of the things we found this year is that the Netflix phenomenon is having an effect on traffic. That sort of long-form video is driving three times more traffic than the short form, which are the YouTube clips. That said, Netflix has recently been offered in Canada but YouTube is still a huge resource, as well as people posting videos on Facebook, which could be one of the big drivers for the Canadian traffic growth. We have 16 countries that we cover in detail and that’s based on the data we’re able to get. While we do look globally by region, there are 16 countries that we’re featuring in our report in some detail and that may be one of the reasons that of the countries that we cover, Canada comes up where it does in the rankings.

How do Canadians fare in peer-to-peer usage? Are they disproportionate users of things like BitTorrent?

Sumits: I wouldn’t say so. By far the highest number of peer-to-peer users are in Asia, followed by Europe. The U.S. is the lowest percentage, below 25% [of traffic]. Canada is higher than the U.S. but not as high as Asia. I looked at that and thought it might be contributing to the higher amount for Canada but other things [apply]. There does actually seem to be a correlation between cold weather and internet use.

Japan ranks pretty lowly in internet usage in your report, but that’s because wireless isn’t counted. Is wireless accounted for separately in the Virtual Networking Index?

Barnett: Yes, we actually do. Japan, for some of the reasons you mentioned in your article, there definitely is a different culture and usage of mobile technology there. One comparison for example that shows the stark difference between Canada and Japan specifically is that in 2010, mobile data traffic represented just 1% of Canada’s total IP traffic. In Japan, mobile data was 2.8%. Even though we’re only talking about 1.8 percentage points, the difference was huge. If we extrapolate that out over the forecast period, by 2015 mobile data traffic will be just 3% of Canada’s total IP traffic but it’ll be 12% of Japan’s. The use of video is obviously a huge driver of mobile data traffic, but particularly in Japan based on the networks and speeds they have available. That’s happening much more in Japan than not only in Canada, but in the United States and Europe.

Can you explain how you forecast traffic increases? Traffic in Canada is expected to triple while in the UK it is seen quadrupling by 2015.

Sumits: Historically, growth rates tend to hover around 30 or 40%, so tripling or quadrupling is consistent with that. Some of the drivers are broadband speeds in particular, and that’s something that may be contributing to some of the country trends that you see. Deployment of super high speeds with fiber significantly contributes, the number of connected devices are expected to double worldwide. All of these are drivers of the growth rates. Another reason is that lot of the video traffic in particular is also migrating from one medium to another. If it’s going from TV, which is very efficiently delivered and just sent across the network once and then split out at the last minute, if they’re moving to an on-demand or internet platform, that all becomes unicast traffic. It may all be the same from the user’s perspective, but it’s contributing to the traffic growth.

Why do you group Western Europe together in your reports? The Nordic countries are generally thought of as internet leaders – wouldn’t including them change your stats considerably?

Barnett: We start by having a model and we feed into that a variety of syndicated research forecasts from a variety of vendors. To some degree, while we can have reasonable information, the countries that we’ve selected represent the countries where we can get detailed information on those countries through some of our syndicated third-party research sources. Part of it is just a coverage perspective of what we’re able to get.

Sumits: It’s no judgement on the importance of those countries. It’s sort of like Asia, where South Korea is generating so much traffic. The same thing could happen for Europe. When we can get solid enough data we can include them. We just decided to start with the G8 plus five and we’ve added a few as we’ve been able to get more data.

Barnett: Just this year, because of the demand we had from Australia and New Zealand, we were able to go out and find… it’s not just finding data, it’s finding credible data. As we’re able to add new countries, we certainly will. Given the feedback you’ve given us on this particular conversation, we may look at Scandinavian countries for a number of reasons to see if we can’t include them in future forecasts.

How should people read the VNI? What should they take from it and what shouldn’t they?

Barnett: One of the major shifts we see is that Asia-Pac with its larger population and its rabid willingness to adopt a variety of different consumer devices will by 2015 be the largest traffic contributing region. We’re seeing a changing of the guard from that perspective. I wouldn’t say it’s a kind of sky-is-falling projection. We’ve seen over time service providers continue to innovate, transition and transform their networks. There will be challenges to support all of this traffic, but it is just an evolution.

Bottom line: I think there was a lot of good information in that interview, with a couple of key points to take away. Firstly, while U.S. entertainment companies claim that Canada is a piracy haven, the numbers seem to indicate that things aren’t as bad as they’re made out to be. On the other hand, with peer-to-peer file-sharing being higher in Canada than in the U.S., that would seem to reinforce the notion that having legal content easily and inexpensively available seems to be the best counter to piracy.

More importantly, the conversation reinforced what I concluded in my original blog post – that while Canadians are indeed prolific users of the internet, they certainly are not the “third biggest consumers of gigabytes on the planet,” as claimed elsewhere. They are, in fact, projected to land middle of the pack of a select group of countries that is, at this point, fairly limited.


Posted by on June 27, 2011 in cisco, internet


4 responses to “Cisco’s broadband bikini revisited

  1. Duncan Stewart

    July 1, 2011 at 3:00 pm

    Hi Peter,

    I’m a long-time fan of your work. But I believe that your comment above (that I misconstrued data and “claimed” something to be true that is untrue) is unfair and inaccurate.

    Let me explain why.

    1. Your quote from my June 23rd column is accurate: I did write “Canadians are the third biggest consumers of gigabytes on the planet.” But that quote lies within a sentence that includes a hyperlink to my June 9th column, which discusses the Cisco VNI numbers in more detail. In that column I explicitly acknowledged (in two different paragraphs) that the data was monthly per capita wireline usage.
    2. Next, in your interview with Cisco they suggest the Economist chart is deceptive because it is labeled “per person” rather than “per capita.” I have looked that up in a number of sources, and as far as I can tell the two terms are synonyms.
    3. You make an argument that the data would be more accurately presented on a per household basis, saying that including business usage is misleading and should be included. For some things, perhaps. But given that the UBB debate is most frequently framed around total Canadian competiveness, I would argue that including all internet usage is a more useful number. In fact, the whole focus of my second column was around whether or not Canadian data prices and caps were hindering Canadian companies. Making household data much less relevant than per capita, no?
    4. You make a fair point that the low Japanese number on the Economist chart is partially distorted by their high use of mobile internet data. But, once again, given that the focus of my column was on UBB and therefore wireline broadband, isn’t my use of the Cisco/Economist data appropriate?
    5. I also have to say that your conclusion – “Canadians are…projected to land middle of the pack” – is disingenuous at best. I never once in either column referred to the 2015 projections. I love Cisco, and I love the VNI, but making projections that far out is tough. I was writing only about where we are today. And today, based on wireline data used per person, Canada is number 3 out the 16 nations that Cisco has detailed and accurate data for. That ain’t middle of the pack, and the fact that we are ahead of G8 peers like the US, Britain, Germany, Japan and Italy is not trivial: of those 5 the only one that is even close to Canada is the US.
    6. As a final note, even though I don’t think “middle of the pack” is accurate, nor do I agree with your household or including wireless data choices, your conclusion is still remarkably similar to mine. Despite the OECD data suggesting we have high wireline data prices and too low caps, Canada is NOT a “digital backwater.” Whether we are in the top quartile or the second quartile is splitting hairs: we’re not lagging the rest of the world, as many of the UBB crowd allege.

    None of the above means that we should be complacent, or fail to invest in next generation networks, or that we don’t need to focus on innovation and the digital economy.

    Respectfully, and I look forward to reading more of your thoughts on this and other issues.

    Duncan Stewart

    • petenowak2000

      July 1, 2011 at 6:32 pm

      Hi Duncan, thanks for your comments. You make some good arguments but I think you may have missed the underlying point, which is that it’s hard to make claims about Canadians being the third-biggest data users in the world – whichever way it’s measured – when only 16 countries are considered. That’s particularly compounded when broadband leaders such as Sweden and the Netherlands aren’t included, yet the likes of South Africa and Russia are. I believe comScore stats generally show the same things, but unless I’m mistaken those numbers are even more limited, covering only 11 countries.

      That said, it shouldn’t be surprising that Canadians are big Internet users. We had broadband early and therefore took to using the Internet before many other countries. UBB is not likely to stop that, it’s just a question of how much it’s going to cost people to do what they’re accustomed to doing. Some data suggest we’re already paying comparatively more than peer countries.

      Is Canada a backwater? Of course we have better wired and wireless networks than most of the world, so definitely not in that sense. But we’re no longer the leader we used to be as numerous studies from OECD, World Economic Forum, Ookla, etc. have shown. What UBB could do on a competitive level is discourage high-bandwidth services from developing or coming here. Netflix’s recent quality degradation because of UBB wasn’t exactly a vote of confidence to other potential similar services.

      You may not think slipping from first to second quartile is a big deal but it is when your trajectory keeps aiming lower.

  2. Duncan Stewart

    July 1, 2011 at 6:49 pm

    Thanks Peter. I agree the data set is far from comprehensive…but (as I often say in my column) you work with what data you have.

    I also agree that we shouldn’t gloat about beating South Africa, India, etc. But being #2 of the G8 seems pretty good — top quartile again.

    My next column will actually address exactly your question: what is the likely impact of high prices/data caps on user behaviour going forward? It should appear next Thursday in the online Globe and Mail.

    Finally, let’s not get ahead of ourselves. Using my methodology, Canada is 3/16 today. Using yours, we are 4/16. Still top quartile. Where we will be in 2015 is really hard to predict — and this is from a guy who gets paid to make Predictions! 🙂

    Thanks for engaging, and Happy Canada Day!


%d bloggers like this: