What the Rise of Amazon's Cloud Computing Means to the Industry
Submitted by Bernard Golden on Wed, 2010-08-04 18:17
This week brought a fascinating report from UBS assessing AWS from a financial perspective. I became aware of the report from the very perceptive blog by Om Malik. Essentially, UBS did some financial reverse engineering to calculate AWS revenues (which are lumped into Amazon financials in the "Other revenue sources" bucket), margins, and growth prospects. The report is excellent and offers some (relatively) hard data to back up what everyone intuitively knows: AWS is knocking it out of the park. Certainly, compared to the general growth of the technology industry, AWS is showing a much higher growth rate.
Is Amazon pricing ruining the enterprise market?
This is the plaintive lamentation from an executive at one of the world's largest technology firms. Boiled down, his whimper can be summarized as "Amazon only charges $.085 for a processor hour; I don't know how to make money at those price points."
The UBS report offers some extremely illuminating information on this front: it calculates that AWS gross margins on its services are in the 50% range -- which, by the way, are significantly higher than the other portions of Amazon's business. In other words, AWS provides greater margin and is thereby a more attractive part of its business. It is likely to put more resources into this area of the company.
One misapprehension about AWS is who is using it. Many people think only startups or the SMB market puts applications on AWS. I can speak from our client interactions as well as discussions with many other players in the cloud market that the belief that "real" enterprises aren't using AWS is wrong -- and the big tech company's executive's jeremiad evinces that fact.
Consequently, the pain that technology executive is suffering is only going to get worse. And the need to figure out how to respond more critical.
Amazon's offering is straight out of Clayton Christensen's The Innovator's Dilemma. The enterprise offerings from these large companies "overserve" their customers -- even the simplest needs are accompanied by multiple interactions, visits, contract negotiations, conference calls, etc., etc.
However, there is a lot of pushback from many enterprise customers that Amazon needs to modify its approach to serve real enterprise needs.
How can one reconcile these facts: widespread success and customer (or prospect) belief that Amazon falls short of what it needs to deliver?
My sense is that the term "serving the enterprise" conflates a number of elements.
One element is the very real needs of enterprises in areas of compliance, verification, etc., etc. Often these issues manifest themselves in questions regarding storage location and recoverability, data center practices and security, and so on. All very valid. And no question that Amazon's rather opaque response to these kinds of questions does fall short of what is appropriate. I believe that there should be a way for customers to get more insight about these elements.
Another element is what could be termed "personal attention and custom terms." Enterprises are used to negotiating contract terms to suit their goals. Entire procurement departments exist to (and justify their existence by) haggling over prices, payment schedules, invoice formats, and the like. Beyond this, enterprises like to have roadmap alignment and architectural discussions, with all the trappings of on-site visits, company-specific checklists, executive briefings, and so on.
In some enterprises, this interaction extends to a desire to obtain commitments from suppliers to specific contract obligations like liability, etc. In these cases, enterprises wish to strike a custom agreement with Amazon (and any other cloud provider they might select) to address individual needs and concerns.
With respect to these latter two enterprise objectives, Amazon is quite stringent in its position: it does business a certain way, at certain price points, and it's not customizable.
The challenges of this approach for enterprises was highlighted this week by the kerfuffle in the trade press about Eli Lilly supposedly pulling the plug on its AWS relationship due to an inability to obtain a commitment from AWS with respect to storage security and financial penalties.
The fact is that the nature of computing is changing. Portions of it is being commoditized. Expecting to garner the benefits of commoditization while still wanting the hand-holding and custom arrangements of the old style of computing is a recipe for disappointment.
Amazon is not going to change its approach. Its margins on what it offers show that it's succeeding with its style of doing business. For vendors, this means that a new (low) price hurdle is being set for commodity compute services. They need to figure out how to create a competitive offering at a similar price point. For users, this means that the traditional expectations of custom agreements and services need to be adjusted in light of the reality of what commoditization really delivers.
I believe that the custom agreements and extensive personal attention and interaction will be offered by some cloud providers -- but they will come with a cost. The cost will either be implicit, in the form of higher compute costs, or explicit, as an invoiced service offering. But either way, the old model of overserving customers, with accompanying high prices, with no customer choice about whether or not to select the personal interaction and custom agreements, is heading toward obsolescence.
This new model will impose challenges and confusion, for both vendors and users, no question about it. But just as the rise of Wal-Mart bifurcated department stores into value and personal attention, decimating the middle-of-the-road approach in the process, so too will streamlined cloud computing do the same to the enterprise market.