On-Demand Integration: Salesforce, Hubspan, and the Future of Heterogeneity

Salesforce.com is nothing if not a font of great ideas whose time should have come a long time ago. Indeed, it almost seems as though Salesforce.com's top dog Marc Benioff is the only one in the kennel who's bark really matters: On demand CRM certainly was not invented by Salesforce.com, but it hardly takes a leap of faith to credit Benioff with truly creating a market that others had already been promoting for years.

So with this grain of salt firmly in mind, let's look at the latest press release from Salesforce.com. After much posturing and positioning, it apparently announces an on-demand integration service that complements the company's AppExchange eco-system strategy. This after a spirited, and confusing, defense of the company's multi-tenancy architecture, a point that might be relevant if only if were well explained. And some other posturing -- such as the "first on demand outbound messaging in the industry" -- which, forgive me, sounds like....a lot of posturing. 

But underneath it all, as with most of Benioff's ideas, there's a helluva good story, even it that story isn't just about Salesforce.com. Because application and process integration is one of the great opportunities for on-demand services. In fact, great just isn't hyperbolic enough. Putting application integration into the hands of a specialized vendor that can mediate the myriad APIs needed to hook up different parts of the enterprise into a semi-seamless whole promises to be one of the truly amazing opportunities for on-demand services. This is one of those no-brainer applications for on-demand (the other is analytics, in case you were wondering) that will become an even better reason to opt out of a major in-house development effort and leave the APIs to the professionals. 

Take a look at Hubspan.com if you want to see how this can be realized in a single corner of the application interface world. Hubspan has been out selling on-demand customer integration services for web commerce for a number of years, with some very seriously happy, brand name companies among its customers. The opportunity is simple: provide an on-demand service that lets a buyer use any procurement software his or her company wants for purchasing goods and services from any e-commerce website, regardless of the e-commerce software that vendor has chosen to use. The essentials of the transaction -- order, invoice, tracking number, approvals, etc. -- are translated and transformed by Hubspan, and the result is that the e-commerce buying and selling experience takes place without anyone on either side of the transaction even knowing how many APIs were traversed along the way. 

Hubspan, and Salesforce.com's more fledgling efforts, are harbingers not only of a brave new world of application integration, but are also the proving ground for the promise of web services and the prospects of a best-of-breed, one-size-does-not-fit-all future. This kind of on-demand service will be made easier -- and more necessary -- as web services and their usage proliferate. And, with the relative ease of integration that on-demand integration provides, more best-of-breed offerings will be able to live, and thrive, in a complex world made up of many heterogeneous, and interlocking, applications and processes. 

How well will Salesforce.com's efforts pan out? It's hard to tell, Salesforce.com's self-reported numbers are alway a little hard to parse, so we may have to just wing it and assume that this good an idea will have to succeed. Because putting application integration into the on-demand world is genuinely a great idea, one that might be even more useful than CRM on-demand. And, that, we can all agree, has been proven to be a pretty neat trick so far.