The Long Tail of Web Services

Google is a “global technology leader”, the tech company to watch in the 21st century; and Amazon is an online retailer—so 1999. Right?

A casual perusal of the blogosphere and IT news sites would seem to suggest such a consensus. Amazon gets occassional kudos for its web services initiatives like the Simple Queue Service (SQS), the Elastic Compute Cloud (EC2), Alexa Web Services, and it’s Simple Storage Service (S3)—though many wonder why they’re spending so much money on something apparently not core to their business (something like $2 billion so far, according to a CNET News article).

But when people think of strategic technology leadership and innovation, they generally think Google, not Amazon. There’s no shortage of hype about Google becoming “the web platform that powers much of what people do online”, the “Microsoft-killer” with its very own OS (this one is particularly absurd), and “the locked-in platform on which web applications and services are all built.” These are only a few selected examples of which there are countless others on any given day.

Now there’s no denying that Google has a lot of smart employees and a lot of cool technology. However, little of that truly innovative technology does anyone but Google any good. For example: the Google File System is by all accounts way cool, but noone outside Google is ever likely to benefit from it. And their expertise in such hardcore tech neither directly generates any revenue, nor builds any profitable relationships with users, advertisers, or partners. An MIT Technology Review article, What’s Next For Google quotes a couple of employees claiming: “Google’s leaders believe that the company’s expertise in infrastructure—knowing how to build and operate those 250,000 servers—constitutes a competitive advantage more important than APIs or standards.” Uh oh…

Amazon, on the other hand, has apparently made a commitment to turn their technical know-how and serious real-world experience into a profit centre. Since it costs so much money to develop the kind of high-performance, highly-distributed, fine-grained Service Oriented Architecture for their own operations, it makes sense to monetize it. As this ACM Queue interview with Amazon’s CTO Werner Vogels (who also maintains the personal blog All Things Distributed) makes clear, these people know as much as anyone about how to make composite web services actually work on a massive scale. And they realize that this is a strategic asset with real growth potential. As Vogels says: “I think our biggest success has been that Amazon has become a platform that other businesses can benefit from.” And: “We see Amazon.com as part of the larger Internet ecosystem, and we want to stimulate innovation wherever possible.”

Contrast this view with the recent kerfluffle over Google’s abandonment of its SOAP-based search API. The issue here has nothing to do with the technical supremacy of AJAX over SOAP, and everything to do with the fact that—despite its self-description as a “global technology leader”—Google makes 99% of its revenues from advertising (did you know they own a radio advertising subsidiary?) With such a business model, only eyeballs matter. Machines and software processes cannot view ads, cannot be influenced by branding messages, and do not make purchasing decisions. For Google to become the sort of web services platform so many people seem to expect, they’d have to radically change their business model. As it is, there’s absolutely no advantage to Google doing anything other than forcing more and more human eyeballs to view more and more ads. Having worked in advertising for several years (I’m in recovery now, thanks!), I know as well as most that doing so requires making decisions that are directly counter to the interests, needs and wishes of your customers and users (no matter how much account directors may protest that consumers actually want what’s being shoved down their throats.)

So, to end this long-ish rant, don’t look to Google to provide any real technology leadership anytime soon (cutesy AJAX email interfaces really aren’t going to change the world—sorry.) Instead, consider what you might call “The Long Tail of Web Services”: “Ultimately, Amazon wants to build up a massive partner network of tens of thousands of third-party outfits, ranging from corporations to developers and even hobbyists. ‘I’ve always thought that high-volume, low-margin businesses are more defensible,’ Bezos said.”

Coming soon: a tutorial on running a network of JXTA super-peers on Amazon’s Elastic Compute Cloud.

Update: there’s a new, related post, here.

5 thoughts on “The Long Tail of Web Services”

  1. Vanessa, great post. I couldn’t agree more. Google, Yahoo, Microsoft and Ask are all in the same business as far as search goes. They want eyeballs and advertising revenue. You’ll see them all do what Google did recently and pull back all of the programmatic controls/access to search. APIs into their search are contrary to their business model and they don’t want to enable a future competitor to build off their backs.

    Meanwhile, Amazon and Alexa are taking the polar opposite route. Give programmers the keys to the kingdom and see what they do.

  2. Congratulations Vanessa!

    It’s an outstanding post! I’ve never thought about this point of view, but it really makes sense!
    When I first read about the innovative services of Amazon I remember that I kept asking myself: why Amazon? And after a few minutes, I just figured out that Google (hopefully) doesn’t have all the smart employees in US yet.

    I’ve been pondering about using Amazon’s EC for a while, so I am eagerly waiting for the JXTA Elastic Cloud post. Please, don’t forget it.:-)

  3. Thanks, Edward. Don’t worry, I won’t forget the JXTA/EC2 post :) It’s just a little time-consuming to put together a proper article.

    I just noticed yesterday that you can share EC2 images you make, so I may be able to create a sample RDV/Relay image and make it available that way. I have a little more research to do, however. So much work, so little time…

Leave a Reply to Edward Cancel reply

Your email address will not be published. Required fields are marked *