by: Christian Smagg
While recently discussing with Matthieu Hug, co-founder and Associate Director of RunMyProcess, how integration could potentially be a significant deterrent to effective adoption of SaaS solutions, I invited him to share his views on why SaaS is more than just another analysts’ fantasy:
50% of the mergers & acquisitions deals fail because of IT. That’s a pretty striking finding that should seriously question the way IT integration challenges are still being overlooked by many organisations. With such figures, it’s needless to look for responsibilities within IT departments or to point out changing user needs. This simply reveals organisational failures, and as such, calls for a re-thinking of some of our IT habits: drop complexity, huge projects and the dream for a single, exhaustive system; focus on short time to market and agility; don’t try to avoid errors and change but accept and manage them. Huge IT projects (calculated in thousands of man days) corresponds to what I would call a "pre-industrial" way of doing IT. Industrialisation is not about becoming complex and large, it’s about answering more needs, better and faster. Hopefully, IT’s "industrial revolution" is upcoming, and it’s called "Software-as-a-Service".
There are at least 3 reasons to back such an affirmation:
SaaS changes fundamentally the economics of IT. In a traditional IT department, the very large majority of the activity is usually focused on operations and maintenance, with business innovation representing a fairly small percentage of the overall IT budget. SaaS removes the need for large software and hardware investments and maintenance costs, and places the technical operations under the responsibility of the SaaS vendor. Overall, this allows IT to concentrate on business operations and innovation.
SaaS is the way to a Darwinian IT environment, where unused solutions disappear from your landscape immediately. Changing the service you use has a limited cost, changing a product you bought is plain money lost; when analysts reckon that more than 50% of bought softwares are never even put into production, this is more than a marginal gain. In simple terms, this simply means that SaaS vendors are doomed to excellence or death: that’s good news for the customers.
SaaS blurs the frontier between software and business service. How simple is it with traditional IT to add a new customer scoring functionality to a CRM system? Or to make an external partner participate to your core processes? In a SaaS oriented environment, there is simply no difference between the CRM service and a scoring service. Similarly, your customers’ or suppliers’ IT is a set of business services, which are not really different from the software or online services you subscribe to.
Overall this will drive corporate IT beyond the walls of the enterprise: future IT value will come from seamlessly linking and integrating legacy systems, B2B partners, online softwares and individual business services. There are obviously a couple of technical requirements allowing a seamless integration between applications, mainly around SOA (Service Oriented Architecture) for legacy systems. But that’s far from enough to get an operational "extended IT". Think of IT as an orchestra, where each "service" (being online software, legacy SOA, B2B services,etc) is a musician with its instrument, and where business users write down a changing partition. A single pianist can play a sonata but a symphony is out of its range; and as the number of musicians increases, they can’t play the music alone: you desperately need a conductor.
Similarly, you will naturally use several services, and thus you need to orchestrate them. Indeed, one of the most important requirements of the extended IT, services and web oriented architecture, is around the orchestration of all these services, or "integration" of them all in IT terms. Besides, this orchestration should be easy to set up and modify, controlled by business people rather than techies, and avoid software development as much as possible. Last but not least, the tools that allow to design and run this orchestration are only making sense if these are SaaS platform themselves (eg online and usage based). To use acronyms, such platforms are often called SaaS Integration Platform (SIP) or Internet Service Bus (ISB).
SaaS is an opportunity to take into consideration for one of IT’s most complex reality: quickly evolving business needs. A good SIP is the tool that leverages this opportunity and should be kept simple and ready to use.
Thanks to Matthieu Hug, co-founder and Associate Director of RunMyProcess for this contribution.
Original Post: http://www.saastream.com/my_weblog/2008/05/integration-str.html