Amazon and AOL Provide Scaled Infrastructures for Startups

A lot of people have great ideas for startups. But even if you have a great idea, the path from brilliant conception to realized product to product recognition by the customer base to successfully scaled product delivery is a many faceted obstacle course. It's no wonder that so few startup companies realize the ultimate objectives of the founders.

One of the more successful paths for startups today is develop a unique feature or implementation or add-on that is related to an already established application or service. With this strategy, the startup attempts to ride the coattails of success with enough added innovation to eventually inspire a larger company to buy the startup and integrate its innovation into the core project.

If you're working on something more novel, you may face big company problems that have to be solved by a small team.

Creative startup problem #1: undifferentiated heavy lifting

At the Web 2.0 Summit in November, Amazon.com's Jeff Bezos talked about providing startup companies with a quicker path from idea to successful product. He said that 70% of a young technology company's time, energy, and money goes into what he collectively terms "undifferentiated heavy lifting."

Undifferentiated heavy lifting is all the work that must be done in order to bring your product to the marketplace, but which has nothing to do with your product itself. For example, if your product is a web product, then you have to think about servers, data storage, an operations infrastructure, etc. Everyone who has worked on these aspects of web service products will likely agree with Jeff's assertion that selecting, configuring, and operating the platform itself can occupy the bulk of a startup's energy.

Purchasing and maintaining the platform also involves significant risk for the young company. A company I worked for prior to and during the dot com bust spent enormous amounts of money purchasing and housing a state-of-the-art data center that was going to be the platform on which their new highly-scalable business-to-business product was going to run. The product that was going to run in the data center was never finished, the data center sat idle for some years, and eventually the computers themselves disappeared!

Creative startup problem #2: reinventing the wheel

Businesses that have a long history typically also have a large existing code base that has met the needs of the past and has been adapted and supplemented as needed to meet today's requirements. A startup does not have this advantage. Everything has to be built from scratch.

This is one reason startups embrace open source operating systems and software. Open source components provide reasonably tested functionality at the right price: free. Still, a lot of open source software has not been tested in the high-stress operational scenarios that will confront any startup whose application achieves mainstream success.

Building the application or service that supports the startup's core value added functionality can easily occupy the bulk of software development effort, if suitable open source components are not available. In a very real sense, this effort to reinvent a scalable software infrastructure is effort lost and time lost in the race with larger companies who are the startup's competition.

It's a troubling dilemma. To release a product that won't properly scale is surely a path to failure. But if you delay the release of your product while you do all the tedious durability testing and performance tweaking that's necessary to squash out all the little slivers and bugs that can crack your infrastructure foundation to pieces under the heavy operational loads that come with success, someone else may well own the marketplace by the time you bring your product to market.

Veteran heavy hitters step up to the plate: Amazon and AOL

From out of the blue, two "dark horses" have appeared, offering startups proven infrastructures that can be purchased on a "pay by the drink" basis. Amazon.com and AOL have both been written off by some people as relics of the Internet's infancy. Both companies succeeded by providing a path that reduced the complexity of the Internet and hence eased the discomfort in using the Internet for non-techie people. AOL provided a way to get online and use the Internet without ever leaving the familiar territory of AOL's site. In one place, you could find just about everything you wanted online. Meanwhile, Amazon made online purchasing feel safe and secure for people, allaying the fear of what might happen once you entered your credit card number into a browser window on your computer.

Unlike many technology companies, the customer base for both AOL and Amazon was the mass populace, which at the time when the companies started was not highly technical. People were interested in the online world, but computers were very frustrating. AOL made it easy by shipping CDs people could just plunk into their computers. The CD installed everything necessary to get you online and connected to the very diverse online world AOL was creating on its ever-growing server infrastructure. Stick in a CD, answer a question or two, plug in your modem wire, and the next thing you knew you were online! Fantastic!

Amazon did something similar with retailing, starting off with books, but then quickly expanding its catalog. Again, a big part of what made it successful was ease of use, but an even more critical element was reliability. Amazon didn't just throw up a web site, take in orders and credit cards, and hope that not too many flubs occured. Rather, from the start, Amazon made the necessary investment in infrastructure to ensure that transactions proceeded without error, making Amazon customers feel secure and comfortable about making purchases using their computer and the Internet.

In both cases, the companies earned the trust of their customer bases. Why leave the simplicity of AOL where everything's right there for you, available using a simple keyword, for the complexity of web browsers and URLs and all that other stuff technology people talked about? And Amazon: well, of course you could trust Amazon.com with your purchases. But, would you really want to submit your credit card number to some little company's web site where you couldn't even find an address or a phone number? There were so many stories of fraud...

Foundation of customer trust: reliable infrastructure

The foundation for the trust their customers gave to both AOL and Amazon was based in great measure on the focus both companies place on establishing and maintaining highly reliable hardware and software infrastructures. When people dialed into America Online or went to Amazon.com, they did not have to cope with problems they saw elsewhere on the Internet: things like nothing happening after a link was clicked or a button pushed, strange boxes with obscure messages popping up, or pages that never finished loading. These things simply didn't happen when you were using America Online, or AIM, or visiting Amazon.com.

The front cover article about Amazon in the November 13 issue of Business Week includes paying attention to customer needs rather than watching competitors as one of the "best practices" that have been key to Amazon's success. But how can you deliver what your customers want if your service is not reliable? If your service isn't reliable, many potential customers will give up and never even see the innovation that is your company's fundamental value-added entity. This is especially true if you're an unknown startup trying to gain traction.

Amazon's muck

At the Web 2.0 Summit, Jeff Bezos represented Amazon's Amazon Web Services with the statement "We make muck, so you don't have to." The opening statement on the Amazon Web Services front page is:

Amazon Web Services provides developers with direct access to Amazon's robust technology platform. Build on Amazon's suite of web services to enable and enhance your applications. We innovate for you, so that you can innovate for your customers.

Two new Amazon products are key to taking a significant portion of the burden of undifferentiated heavy lifting off the shoulders of startup innovators:

• Amazon S3: the Amazon Simple Storage Service "is storage for the Internet. It is designed to make web-scale computing easier for developers."
• Amazon EC2: the Amazon Elastic Compute Cloud "is a web service that provides resizable compute capacity in the cloud. It is designed to make web-scale computing easier for developers."

Can these services really help a startup? While not free, the pricing is low enough that Amazon is already attracting a rapidly growing customer base among startups. Here's a quote from the recent Business Week article:

"Everything we can get Amazon to do, we will get Amazon to do. You're going to see all kinds of startups get a much better and faster start" [by using Amazon's services].

Chris MacAskill
SmugMug.com

AOL's already-scaled services

AOL has virtually turned itself inside out in terms of its software platform. AOL now offers developers access to the already-scaled services that were previously utilized by AOL internally to provide content to AOL's subscribers. To do this, AOL has applied industry standards to build APIs into their services, documented the APIs, packaged them, and made them available to developers. Go to the dev.aol.com Available APIs page to see a list of the currently available APIs.

All of the APIs come with the right pricing standard for experimentation by developers and entrepreneurs: free. And the usage limits for free access are high enough that you can get your prototype up and running, servicing a viable user community while you experiment with and tweak your core innovation, before you'll have to think about paying AOL for extended API calls as you bring your system live.

Just innovate (and leave the plumbing to us)

The already-scaled hardware and software infrastructures offered by Amazon and AOL provide startups with a way to focus on their own particular innovation, their own value added proposition to customers. It may surprise many to see this coming from these two companies, both of which have received their share of frowns on Wall Street and from the technical crowd.

Few would deny that both AOL and Amazon have excelled in producing high quality, highly reliable, fully-scaled infrastructures. Those who have worked on infrastructure understand how difficult this is to accomplish. These highly reliable infrastructures are now available to individual developers, entrepreneurs, and startups.

Tim O'Reilly calls Amazon "a pretty serious dark horse" in the race to build the future platform for the Web. I'd say the same statement applies for AOL, which comes into this "race" with an already enormous Internet presence: consider AOL.com, Mapquest.com, sites served by Userplane (for example, MySpace), and AIM instant messenger, for starters.

Both Amazon and AOL have configured their proven, working, reliable infrastructure platforms to enable innovation by others, with minimal or no up-front cost. Sounds like a very good deal for startups to me.

-- Kevin Farnham
O'Reilly Media