SaaS: it could fizzle out on the user desktop

What is old is new again and hiring software – which has been around almost since computing’s “Day One” – is reinvented under names such as cloud computing, software-as-a-service, and others.
##CONTINUE##
All kinds of statistics are tossed around: we read, for example, that “35 per cent of companies by 2011 will be using SaaS”. Today, line-of-business SaaS applications probably account for less than 3 per cent of total enterprise desktop applications. If we assume that the ratio will grow in favour of SaaS, it still has profound implications for integration and productivity.

For starters, there is little or no integration capability for existing end-user, desktop applications that are outside of SaaS – which means the whole pile of an enterprise’s legacy applications – mainframe, Windows-native, web, PowerBuilder, Java, even DOS. The list goes on and on….

In evaluating integration needs, I have seen enterprises with 500-plus different desktop applications. I have seen as many as 40 applications on one user’s desktop. Sure, the average is around eight or 10.

But “integration” often sits squarely on the shoulders of the user, in the form of cumbersome workarounds. Manual copying between windows, flipping and toggling between applications, getting lost, missing critical steps in workflows.

This is disastrous for any service-oriented business, such as a contact centre, where time and customer satisfaction are paramount. Why, in 2009, do people think that copy-and-paste is an integration strategy?

Regarding integration, I would not discount service oriented architecture, because exploiting services – opening any applications, including a desktop, as a service, or consuming services on the desktop, is a very real and desirable “last kilometre” of SOA. But pure SOA solutions tend to be based around the server, expensive, and often difficult and expensive to modify.

Add SaaS, and integration is even more elusive. SaaS in a single browser window or virtualised bubble might solve some problems – upgrades, licence control, maintenance, for example – but creates others. How will it work with your other legacy applications? It could put you back to square one, productivity-wise.

There are some bright spots. Salesforce.com, for example, has an application programming interface suite and I think it’s one of the best there is. Even then, integrating SaaS APIs with legacy applications may require significant IT skills that enterprises do not have – and users do not have the patience to wait two years for some kind of solution.

But to use Salesforce.com as an example again, look at how it works with Microsoft Outlook. You can seamlessly integrate the two with a couple of clicks. If only you could do that between any applications, with universal APIs available at the desktop.

This is why OpenSpan has spent seven years or more building a technology that can quickly mate and harness any application set, at the user level, where productivity, compliance and regulatory demands are greatest. It lets companies use what they have, so that SaaS or other cloud applications are a full asset, not a partial hindrance, to enterprise operations.

With SaaS, businesses need to be very careful what they ask for. If they use other applications in the enterprise, then before they put a single SaaS screen on users’ desks, they should make sure it’s not a stand-alone application-in-a-box. The users might not thank you for it.

-----------------------------
BY Francis Carden, founder of OpenSpan, an integration specialist
Source：Financial Times

© Copyright The Financial Times Ltd 2009. "FT" and "Financial Times" are trademarks of The Financial Times Ltd.