In trading some e-mails with Jim Crowe recently, he was nice enough to spend some time to answer questions about Level 3's (NASDAQ:LVLT) CDN business for publishing here on the blog. While Level 3 has been pretty quiet in regards to their CDN business and the customers they have been winning, make no mistake that they plan to be a serious competitor. And I'm not just repeating the company line. I see it from talking to customers, learning of new contract wins and by seeing all of the pieces of the CDN ecosystem they are finishing to put in place.
While Level 3 still has some work left to complete before they make the hard push into the market, even reading their Q2 Content Delivery Network Newsletter gives you a good indication of just how much they have gotten done. Their CDN is now running Flash Media Server 3 (FMS3), they added a CDN footprint in Asia, they finished the buildout and launch of their origin storage offering, they added the ability to limit the download speeds for progressive files, added byte-range request functionality (which is seeking into a file even if it has not fully downloaded or stored on the Level 3 network), added geo-intelligence rules to block access to media and enabled reverse proxy ingest amongst other things. And shortly, they plan to roll out detailed URL-level reporting, support for Silverlight HD, enable better customer self-servicing options, implement live WM push functionality, improve their live reporting features and add the ability for customers to move content from hot to cold storage.
And if you look around, you'll see that they are already pushing out their new CDN messaging. On their new recently launched website, the main message on their home page is "Connecting Content Creators to Content Consumers" and talking about their content ecosystem strategy. A new Flash based presentation on their website, entitled "From Creation to Consumption", really lays out the strategy Level 3 is taking, and it's a smart one. Don't just push bits, figure out how to help customers create, ingest, store, manage, track and deliver the content. It's a good presentation and come Q4 of this year, you'll start to hear a lot more details about Level 3's CDN offering. And come Q1 of 2009, I expect Level 3 will become a serious player in the CDN market.
The following were questions I asked Jim specific for posting on the blog:
Question: While Level 3's CDN business does not contribute much to your overall revenue today, what impact do you see it having on revenue 2-3 years out? How big of a business can CDN become for Level 3?
Jim: We anticipate that the CDN business will become a major revenue source for Level 3 over the medium term. We expect the delineation between our Internet Transit business and our CDN business to blur in the fairly near term. At the very least, we expect our CDN business standalone to compare in size to our Internet transit business within the next few years.
Question: At what point do the economics of scale really begin to kick in for Level 3 in regards to owning the network and having a cost advantage for the CDN services? Many telcos say it is cheaper to own the network, specific to CDN, but can you quantify yet just how much cheaper it really is?
Jim: We benefited from the economic advantage of owning a network from the time we entered the business – we will enjoy our cost advantage from the first bit carried to the last. Consider that our network (all layers) has been operating at scale for some years now and we don’t need to wait for CDN traffic to grow materially to enjoy that advantage.
Two things are useful in trying to quantify that; most CDN providers break out, as a percentage of their overall costs, what they spend on network services (usually comprised of collocation, power and bandwidth). That portion of their total cost is the differentiating element. The advantage to us comes from the difference between the retail rates they pay, on a recurring basis, and our internal costs. [See the answer to a following question for a more detailed discussion of CDN economics.]
Question: As you know, many CDNs are entering the market and focusing only on price, even if they say they aren't. Since your costs should be lower, will you come to market anytime soon with a really cheap price, undercutting everyone else, so that you can grab market share faster; while others are trying that model and will lose in the long run, you own the network, so can you win on price and still make money, or at least break even?
Jim: We will certainly be very competitive on price. However, we have two other major advantages over most of the new entrants; scale and quality. We expect that many of the newer entrants will struggle to achieve a sustainable business model and, at the same time, provide the large and rapidly growing capacity that larger object (i.e., video) CDNs require. The customers we talk to are increasingly concerned about a provider’s ability to scale – today and more importantly tomorrow. We have already answered the question of our ability to scale. With regards to quality, we are used to building and operating very large IP networks. We know how to operate them while keeping the performance and availability levels very high. All of the independent performance measures on our traditional services as well as our CDN places our products amongst the best performing in the industry.
Question: When does Level 3 begin to really attack the market with more marketing and sales specifically around the CDN product, without bundling it into all of the other products and services Level 3 offers?
Jim: Making a full range of optical and IP services remains core to our strategy. This approach allows us to meet the complete needs of media and entertainment companies who generally purchase a range of services from CDN to waves and sometimes dark fiber for data center interconnection and other facilities.
While we have an advantage in selling to a large, existing customer base, we also have sales people actively seeking and closing new customers, some of whom may only purchase CDN. Because we have a portfolio of services to sell, we can leverage our ability to sell CDN and Internet transit to those customers that need both. We can also provide our Vyvx Broadcast customers with the ability to encode, deliver and store content. Based on the customer’s solution needs, we can bundle a host of services that we believe no other CDN provider is able to do today. We will continue to increase both the sales people selling and the marketing effort to raise the profile.
Question: To date, from what I know, most of your CDN business has come from your current customers for other services. When will Level 3 focus on going after the pure CDN customers who have no other needs like transit or colocation?
Jim: Clearly you want to sell as many products as possible, but there is a lot of just pure CDN business out there today. We have paid particular attention to our existing customers since we have many long-standing relationships and because many of them had been asking us to enter the CDN space. We also target new customers who we believe will value the scale and scope of our offerings. Funcom is a recent example of such a new customer.
Question: A lot of people want to compare the decline in IP transit pricing to CDN pricing. Since it is not an apples to apples comparison for most and since most CDNs don't own the network and you do, what insight do you have on this?
Jim: We divide the cost of a CDN into four primary elements: the cost to develop and deploy technologies such as intelligent traffic management and server cluster load management; IP/optical transport (i.e., bandwidth); the cost of CPU/storage in server clusters; and the cost to develop, acquire and protect intellectual property. Of these, we expect bandwidth and CPU/storage costs to dominate with the former the largest element of long run incremental cost. Underlying bandwidth costs have been falling quite rapidly, which have enabled a decline in IP/optical transport pricing. CPU and storage costs have a long and generally well-understood price performance improvement rate. These price performance dynamics are a fundamental reason that it increasingly more cost efficient to transport larger objects such as video over CDN networks versus other channels such as optical disc/physical distribution. For a more detailed discussion of these fundamental trends, readers can refer to the “August Investor Presentation” which can be found in the Investor Relations, Presentations and Events section of the Level 3 website.
Question: Can you say how much money you will put in the network this year specific to the CDN product, not including any previous acquisitions?
Jim: I can’t comment on that specifically as we do not break out our financials to that level of detail; however, we continue to make significant investments in our platform. We have expanded into Asia and continue to make capacity augments on all three platforms (streaming, storage, and caching) in North America and Europe. In addition, we have been investing in the infrastructure within our Level 3 colocation space where we locate our CDN nodes so that we can do “just in time” capacity deployments to uniquely serve the growth needs of our customers.
The From Creation to Consumption presentation that we spoke of really is at the heart of what we are trying to achieve and why we are different from other CDN providers you comment on. We are trying to simplify the distribution chain associated with using an IP network for the delivery of rich media content. For this part of the market we believe that, over time, the distinction between IP Transit and CDN goes away. We will simply talk about Internet delivery of content. It is all about efficiently moving bits increasingly dominated by video and other rich media.