What flavor is your ecosystem?

Marc Benioff, explaining Salesforce.com's programmable new Apex platform to Dan Farber last week, likened it to Windows: "We want to have the same open environment and to create competition." He genuinely wants the world to write its applications to Apex — even on-demand ISVs who are already well-established in their own right, such as Successfactors, one of the largest privately-held vendors in the field with over a thousand customers and more than 2 million users.

In fact, he even hinted darkly that it wouldn't be in the ISV's best interests not to write to the Apex platform:

"I think it would be wise to natively support our platform and have a version on theirs. There is no reason not to do both," he said. "I am sure their competitors will if they don't. It is cheaper to build it on our server, similar to Amazon with EC2."

It may be cheaper to write to Apex, but at what cost? An ecosystem like the one that grew up around Microsoft Windows created a virtual monopoly which, while it created opportunities for some ISVs, sounded a death knell for others when Microsoft decided to create competitive products of its own that ran on its platform.

Benioff, of course, has a fiduciary duty to Salesforce.com's stockholders to bring that kind of ecosystem into existence if he can get away with it — virtual monopolies are highly lucrative for those in charge. And he knows very well that the best way to win dominance is to get every other on-demand ISV of any significance to write to the platform. But ISVs know that too, and they understand very well that flocking to support the platform is akin to turkeys campaigning for a bigger role in Thanksgiving. Fortunately for them, Benioff's Windows analogy is false. There are two huge obstacles preventing any single vendor's platform from dominating on-demand computing in the way that Windows has dominated the desktop era:

• #1 The competition
  Right from the start, Microsoft had a built-in advantage because it already owned DOS, the original PC operating system. By the time Windows was ready for widespread adoption, its only significant rival was IBM's OS/2. Working with PC chipmaker Intel and top PC manufacturer Compaq, Microsoft deftly outmanoeuvred IBM and made sure that Windows triumphed. Today, there's no equivalent cabal of manufacturers controlling the Internet. Apex faces many more rivals than Microsoft ever did for Windows — most of them more nimble than IBM at the turn of the 1990s — and all of them playing on an even field.
• #2 The Web
  Not only is the Web is a much more open and collectively owned platform, it functions as a multi-tier platform that's capable of hosting many different interlocking ecosystems. Even though customers will tend to coalesce around just a few ecosystems rather than encouraging a confusing plethora to thrive, they'll avoid concentrating too much power in a single platform provider.

The Windows analogy is not only false, it's also inappropriate for the Web environment. It describes a walled-in ecosystem, in which every participant is dependent on the underlying proprietary platform. Why would anyone tie themselves to a single vendor's platform when the Web itself is already a vendor-neutral platform? Perhaps there's a place in the Web 2.0 (or I should say Web 3.0) era for that kind of ecosystem, but it's just one of a number of different flavors of ecosystem that vendors and users will choose from [Disclosure note: some of the companies mentioned below and other Salesforce.com competitors are or have been clients of mine — see disclosure page]:

• The all-embracing ecosystem: Several well-known vendors are attempting the same platform play as Salesforce.com's Apex and AppExchange initiative. They believe there's a market for a soup-to-nuts environment that provides everything from the database and the programming tools to the application infrastructure and the user interface. NetSuite SuiteFlex and WebEx Connect are both in this game, to name just two.
• The application ecosystem: Virtually every major on-demand enterprise vendor would like to see its own application become the gateway for accessing other applications and services. You can understand why. The owner of the customer relationship gets to collect the revenues and then distributes a share to ecosystem partners. It's an unequal relationship.
• The API marketplace ecosystem: Some providers believe there's a role for a neutral market-maker that provides the commercial infrastructure for customers to locate and sign up for API services that they can assemble into applications. A good example is the StrikeIron marketplace.
• The API developer ecosystem: Companies like eBay and Amazon are encouraging developers to write to their API as a means of driving more business through their platforms. It's the flipside of the application ecosystem, because it gives up the customer relationship to partners, but reaps rewards through maximizing transaction volumes.
• The utility computing ecosystem: Amazon Web Services is the best example of a provider that offers raw computing as a service. The infrastructure is shared, the rest is up to partners and customers. There's very little lock-in with this type of platform, as changing to another provider is a behind-the-scenes operation, much like changing from one web hosting provider to another.
• The service platform ecosystem: This is a more traditional ecosystem play, in which a software infrastructure vendor helps partners who are using its software but doesn't act as a service provider itself. This arrangement keeps platform lock-in separate from provider lock-in. Examples include Jamcracker and Progress Software.
• The service provider ecosystem: There's word that BT is building an online marketplace for on-demand applications. The service provider takes care of unified billing, hosted components and optional add-ons such as composite services. This approach locks you into the provider but allows choice between vendors.

I'm sure I've missed a few flavors still. The main point to make is that all of these different approaches can co-exist — and many vendors are already deploying several of them simultaneously. The nature of the Web is that customers will always be free to sign up to multiple ecosystems and, although convenience often dictates standardizing on a small number of vendors, flexibility demands having the choice to add others as required.

The most successful ecosystems will be those that offer the most popular balance between convenience and flexibility. Benioff of course is hoping the market will converge on convenience in the form of Apex/AppExchange, but if flexibility gets more votes then at least the Apex API will still allow Salesforce.com to play a part. What none of us yet know is where the market will draw the line, and until some favorite flavors emerge, all have a chance to thrive.