JP Morgan Pounds Final Nail in Bandwidth-Centric Model
ragingbull.lycos.com
We believe that new challenges have arisen to face Level 3 as the character of the U.S., and by assumed extension, the global data market changes. As prices for bandwidth and related services have declined, industry revenue growth has slowed, and we project both trends to continue. It is unlikely that there is a rising tide hiding around the corner that will float all the industry's boats. Bandwidth pricing strategy will not drive industry-wide growth so much as it will be a meaningful factor in how the fruits of industry growth are partitioned among competitors.
A differentiated, well-articulated and well-executed strategy for accessing the enterprise market is critical for achieving the scale required to create not just a surviving but also a thriving business. Large enterprises have a vastly different hierarchy of needs than the web-centric market that was once at the heart of the Level 3 strategy, and Level 3 must address these needs itself or align itself with a new breed of customers/partners that have the skills but not the assets to create value-add, end-to-end enterprise solutions. Level 3's global footprint and integrated colocation capability must now be matched with a newly articulated, enterprise-focused strategy on which the company can rapidly execute.
Level 3 has been identified, and has identified itself, with a substantial number of industry hot-buttons such as "bandwidth elasticity" and "silicon economics," but we believe these concepts simply have been justifications for the company's core belief in horizontal market disaggregation. The company optimized its network and systems on the premise that once scaled, there is very real economic opportunity for such an operator. Level 3 identified and executed early on a plan targeting the rapidly growing web-centric industry that required bandwidth and space and saw these as its most important factors of production. For this model, we believe Level 3 was ideally suited, but this model is over.
In order to scale, and in order to make a wholesale strategy viable, some form of access to the large enterprise customer will be the critical centerpiece to success, with no shortcuts in sight. This is a fairly large shift, in our view. Our previous belief was that large-scale intermediation would separate backbone players from their end-users, and that a new, disaggregated middle-class of service providers would prove a voracious consumer of bandwidth-related products. The hoped-for explosion in interconnected intranets, or extranets, has collided with the hard-nosed practical problems of managing and securing one to several connections, let alone dozens or hundreds.
The enterprise as a complex and demanding customer base will quickly winnow these players down to a handful. In contrast to the web-centric buyer, the large enterprise buyer does not see bandwidth as the critical material consideration in its day-to-day operations and is generally more patient and conservative in choice of providers. Like the web-centric customer, the enterprise customer sees the totality of the communications infrastructure as a critical input to their activities.
A new class of competitors rising to meet the emerging and entrenched, enterprise-focused networking companies. As corporations embrace the benefits of greater connectivity over time, they will do so through applications. Increasingly complex applications layers, we have learned, entail levels of management complexity dwarfing the networking challenges that enable them. Traditional IT outsourcers, such as IBM, Electronic Data Systems, and Computer Sciences Corp. are well positioned to assist enterprises take their systems and data onto the network - perhaps better positioned than many network providers will prove to be at taking the network into the IT and data rooms.
The backdrop for this industry revenue decline is bandwidth pricing, which is 30% to as much as 80% lower on a price per megabit mile basis year over year, depending on capacity. We have been unable to identify the link between price and volumes that might allow the industry to price its way onto a higher growth plane. Simply put, the web-centric industry that was the most sensitive to pricing is disappearing, and we cannot identify a potential replacement. Enterprises, while sensitive to price, have a massive number of factors that enter into their bandwidth growth equation beyond this element, which slow their adoption of new services. In the enterprise market, the tail cannot wag the dog.
Corporations, either at their locations or their own data centers, are expected to control roughly the same percentage of total traffic in 2005 (72%) as they do currently.
The implications of this shift are meaningful for service providers because a corporation is a completely different animal from web-centric companies in terms of telecom service buying behavior. While large corporations generally decide which vendors to use for different services or products (such as hosting, transport, voice), they tend to buy these services as part of a managed service offering from a single company. In addition, although the larger corporations tend to outsource the management and storage of their web traffic, these companies usually are unwilling to trust outsourced service providers with more mission-critical data, which in the case of large enterprises tends to be much larger than web traffic in terms of volume. The lead time to secure new business in the large enterprise market is long.
1. There is an overcapacity situation in the colocation market.
2. Colocation is not as attractive to enterprise customers as anticipated.
Most large enterprise traffic is internal and resides on servers in internally housed data centers that the companies are rarely willing to outsource owing to security and reliability issues. While most IT managers/CIOs at large companies are willing to outsource their web page traffic to colocation or hosting providers, they are unwilling to outsource more mission-critical data, which actually comprise a majority of a company's traffic.
We adjusted down our long-term outlook for revenue from Level 3's IP/colocation business because we believe our previous forecasts were inconsistent with our research conclusions that there is a substantial oversupply of space globally impacting rack revenue, that there is a deceleration in the trend toward outsourcing server capabilities, and that there is a declining trend in the bandwidth being generated per server.
Gary's View - Welcome to the party JP. Readers of this newsletter have been discussing this for over two years. Where have you been?
Anyway, now that I am done gloating, let me net this out. JP Morgan is saying that the bandwidth-centric marketplace is dead, and that if the companies modeled on the bandwidth-centric model are to survive, they must adapt. JP still thinks they can survive selling just bandwidth (I am doubtful) but the message is clear - the 1st generation of Internet applications is over. To survive - well, let me quote - "As corporations embrace the benefits of greater connectivity over time, they will do so through applications. Increasingly complex applications layers, we have learned, entail levels of management complexity dwarfing the networking challenges that enable them. Traditional IT outsourcers, such as IBM, Electronic Data Systems, and Computer Sciences Corp. are well positioned to assist enterprises take their systems and data onto the network - perhaps better positioned than many network providers will prove to be at taking the network into the IT and data rooms." |
