IP Traffic and Pricing Trends

June 14th, 2008 by · 14 Comments

Ike Elliott and Dan Caruso have been going back and forth on whether there has been a resurgence in IP traffic and revenue from it over the past couple years.  First Ike challenged the idea that growth is happening by looking at Level 3’s IP/Data revenues and their total lack of movement in 2007, but Level3’s numbers include other stuff that makes it hard to read.  Dan then offered Cogent’s growth as an example, but Cogent’s numbers are skewed by their pricing strategy.  They grew revenue for the last few years at a fixed price, having forward-priced the market.  Those chickens came home to roost recently with slower traffic growth at Cogent, leading to pricing changes there – but as a proxy for industry trends, Cogent is hard to use because their pricing only changes every few years but then all at once.  Ok, so why am I rehashing this?

Because there is a company out there whose traffic growth and pricing has followed the industry.  One which reports its IP revenue in an undiluted form, one without acquisitions that affect it, and which can be traced back quarterly over the past 17 quarters.  That company is Internap.  Internap is Cogent’s opposite, they are a Tier-3 operator, they buy IP Transit from folks like Level 3 and Cogent (and a dozen others), blend them together for increased reliability, then resell it for a premium, and they do it in the datacenter only.  What this means is that their pricing and traffic trends mirror those of the raw bandwidth industry.  Add in the fact that they report their IP revenue number without other crap thrown in, and you have a proxy for the relationship between IP Transit traffic, pricing, and revenue over the past 3+ years.  Here it is:

Internap's IP/Data Revenues

As you can see, Internap’s IP revenue was dead flat for two years at $25M quarterly, nothing could budge it.  Then, in 2006, something happened at last – yes, traffic grew faster than pricing declined – and since then Internap has been doing better and better.  At least until last quarter when they screwed up their integration of the Vitalstream CDN business and there was collateral damage.

In my book, this is evidence for Dan’s case, that in mid 2006 there was a resurgence in this sector.  One that continues today, Cogent’s and Level 3’s pecadillos not withstanding.  But that growth, while real, remains a bit weak yet – about 10% a year at Internap for the last two years.  Their real growth product these days is colocation.

If you haven't already, please take our Reader Survey! Just 3 questions to help us better understand who is reading Telecom Ramblings so we can serve you better!

Categories: Datacenter · Financials · Internet Backbones · Internet Traffic

Join the Discussion!

14 Comments So Far

  • Dan Caruso says:

    Thank you Rob. Ike, shall I make reservations?

    Rob, colo might be good (not my space) but you are overlooking pure unadultured bandwidth, which is growing just dandy. I cannot share all my evidence as much of it was shared with me privately or under NDA. Ike’s original ascertions were broader than wholesale Internet access, though he narrowed it to just this service when I pushed him. Transit is like wholesale minutes–constantly beat back by competition due to the low switching costs. My statement always centered on bandwidth broadly defined. Despite Ike mimicing an Ostrich with his head buried deep in the sand, the bandwidth market is just fine. Don’t tell anyone though as this is the secret to Zayo’s business plan.

  • Rob Powell says:

    Colo has been great for the past few years, but I do realize the limitations of the sector there. I’m not actually overlooking unadulterated bandwidth, I also know it is starting to really kick in – in the right places of course. As Willie Keeler said, “hit’em where they ain’t”.

  • Frank A. Coluccio says:

    An interesting set of analyses, guys. Thanks.

    Rob, I’m a bit unclear of what it is the graph is telling us. A part of my confusion may be due to the fact that I can’t read the vertical axis clearly.

    The rising level is what, exactly? Am I looking at gross revenues of the company, irrespective of the annual increases in deployed infrastructure? Is it a reflection of pricing at the unit level?

    Surely infrastructure growth through capital construction must come into play and would drive the opportunity for more revenues to be generated. At some point you’ve got to normalize those numbers somehow to show the deltas in unit pricing/revs, especially if gauging year-year profitability is the issue.

    Also, your last sentence seems to contradict your main thesis – i.e., that the Internap numbers are purely bandwidth related — if colo services and rents are also included. Or am I reading that incorrectly? Come back …

  • Ike Elliott says:

    Thanks for the excellent research on Internap, it’s very helpful. But be careful with Dan’s comments…he likes to stretch the truth. My blog posts have been very clear that I was talking about wholesale Internet access, and I stand by what I said. Here’s the quote from the original post from my blog:

    “The problem for the wholesale Internet service provider industry, though, is that revenue is not growing at a corresponding rate, because wholesale Internet access service prices are compressing at a rate that very nearly offsets all of the bandwidth growth. So, for the provider of Internet backbone services, you get little or no additional money for providing 50% more capacity every year.”

    How many times do I need to write “wholesale Internet access” for it to be clear? Dan should just admit that he misread what I wrote and jumped to the wrong conclusion, assuming that I was talking about the overall bandwidth market, which I was not.

    To be clear, I never claimed the Internet was not growing quickly. Here’s the link to the post that started it all:

    It’s interesting to note that this post was written to refute a chart that seemed to claim a big jump of about 500% in Internet traffic in 2007 that Dan had posted on his blog. Dan then acknowledged an error in that chart and retracted it.

    All I claimed was that Internet traffic growth in 2007 remained in the range of 50%-60% per year, and that as a result, wholesale Internet revenues were not growing quickly. That’s a point your own Internap chart confirms.


  • Rob Powell says:


    The data shown are only the revenue Internap derives from IP/Data services. They break it out in their SEC filings and I collated it. The rising revenue that began in 2006 refers to a change in relationship between pricing and traffic – for 3-4 years prior to that pricing pressure had completely wiped out the effects of IP traffic growth despite any infrastructure deployed. I was attempting to show there was a turning point in this product at that time.


  • Rob Powell says:


    I don’t think anyone thinks revenue growth will ever approach parity with traffic growth, the point I am trying to make is simply that the relationship changed materially in 2006 for the better, in that wholesale internet traffic grew demonstrably faster than pricing declined. I felt that relationship was at the root of the matter here.

    I do understand the differences between wholesale IP transit and the ‘bandwidth’ that Dan is talking about. I may address that in a later post if I can condense it.


  • Frank A. Coluccio says:


    Thanks for the clarification.


  • skibare says:

    Rob, Ike, and Dan……..interesting conversation but the bottom line is “”WILL Level3 Make Any Money””?????? Seems Dan is doing well……..maybe a rising tide lifts ALL ships hopefully???????

  • Dan Caruso says:

    “Internet backbone providers are hoping that a surge in demand for bandwidth, driven by Internet video distribution and other large file distribution services, will reduce the rate of price declines while creating a big surge in bandwidth utilization.” This is a quote from Elliot Soft-Spitzer himself–if this is a narrow reference to Wholesale Internet Transit, than maybe Hillary really did dodge bullets in Bosnia.

    Ike, there was a mistake in the graph I posted. I immediately admitted this and corrected this. Perhaps you should follow this example, good friend.

    How about if I bring the wine to dinner, then you will only be burdened with a corking fee?

  • Where did you get this data, or more specifically, the denominator in the revenue/bit graph? It is an interesting chart.

  • Rob Powell says:

    The graph is not revenue per bit, it is simply IP/Data Revenues pulled from their filings and the recent corrections on their website. Revenue per bit would necessarily go down, since bits are increasing at 50%+ per year and revenue is not. The relationship that matters here is that revenue per bit is now declining more slowly than the rate at which the bits are increasing – hence the product of the two (revenue) rises.

  • Folks

    A confirming data point is coming from my reading of the financials. It looks like since SBC absorbed AT&T and Verizon took MCI, the backbone competition diminished and profits are going up.


  • coetsee says:

    Hai Rob..
    colo has been great for the past few years, but I do realize the limitations of the sector there. I’m not actually overlooking unadulterated bandwidth, I also know it is starting to really kick in – in the right places of course. As Willie Keeler said, “hit’em where they ain’t”.


Leave a Comment

You may Log In to post a comment, or fill in the form to post anonymously.

  • Ramblings’ Jobs

    Post a Job - Just $99/30days
  • Event Calendar