Has IBM already blown the Software as a Service hosting opportunity?

I saw a post from Microsoft’s Don Dodge today that made me think.

Another dramatic change is coming. It doesn’t have a name yet, but indications are that Microsoft will build out the biggest hosted computing infrastructure in the world that will enable new hosted SaaS applications for consumers, small business, and even enterprises. This is a multi year, multi billion dollar infrastructure project to support the Windows Live and Office Live innovations that are yet to come.

So Microsoft is coming: the company seems to understand the urgency in some of trends sweeping the industry. IBM though has seemingly failed to create a competitive advantage through its own hosting capabilities, because it has focused too much on its Fortune 500 sweetspot and not enough on the markets where the rest of us live.

Firing salesforce.com as a partner in favour of Siebel, because one package allowed consultant customisation over user customisation, short term consulting money over long term small slice subscription revenues, was a bad move.

Meanwhile, not working out a model for hosted IBM software makes even less sense, especially given constant competitor FUD that IBM software is hard to install. In 2006 its still hard to get access to hosted IBM Software, whether from DB2, Lotus, Rational, Tivoli or WebSphere. What better way to get IBM Express products in customer hands, for example, than with online trials?

I regularly defend IBM Software Group from accusers who says its just a bunch of components ready to be installed by expensive consultants. I wish I could defend IBM Global Services’ hosting business..

I asked a while ago Can IBM Thrive In A World Where The Average Deal Size is 2c? But people will pay for hosted applications, with decent service level guarantees. Its a volume opportunity though, rather than a fat heat, fat deal, business model. Cote’ points to some revenue models here.

One would have thought that software as a service was a natural outgrowth of IBM’s On Demand messaging. I guess not. Now IBM is acquiescing OD in favor of the less differentiated "Innovation" message. Its a shame to see IBM now following HP’s campaigns of two years ago: it may be time to reassess its advertising agency relationships..

So what about IBM’s mania towards service-oriented architecture (SOA). What if customers need a host for the service IBM helps them build, or they build using IBM tools, or even tools from a different vendor, for that matter? This won’t be your grandad’s outsourcing model. SaaS, SOA, service management are all being mashed up and remixed.

RedMonk has long argued that IBM needed more touch points with "consumer business" in order to understand the world’s requirements better. Google is not a consumer app. Flickr is not a consumer app. eBay is not a consumer app. They are architectures of participation. They span customers and suppliers. The vast majority of people in the world work for small and medium sized businesses. If you’re planning to build every service to suit the requirements of Charles Schwab or eBay you’re going to build products that only truly suit the fat head.

Its disappointing that a company which has so many native advantages as a potential host of user and third party generated services is failing to nail the space, or even clearly understand it from a market inflexion point of view. IBM hosting has too many barriers to entry, it has not removed enough barriers to participation.

So what can IBM do?

Well for one thing- its surely time to make a major investment to transform developerworks from an education and download website (Web 1.0) into a platform (Web 2.0).

IBM could acquire salesforce.com – and spin off the "salesforce" [if its still worried about being seen to sell applications]… keeping appexchange, with its model of "Browse, Share, and Install On-Demand Applications". A youtube for service developers.

IBM should not really expect to make money by educating devs – developerworks classic should be an investment by IBM, not a line of business with financial targets to meet. Education drives sales, but its shouldn’t be considered a revenue generator. If IBM can turn its ISV and developer programs into something more hosting-oriented it can then charge for that.

As argued above IBM should also, in my opinion, make a much more agressive commitment to have all of its software products available as hosted offerings. It can learn from doing so and reach out to other service providers.

Maybe its not too late, but IBM has certainly squandered some real advantages in terms of a hosting and hosted model.

I wrote recently that IBM’s mainframe group had been invited to visit the Googleplex. Why do you think Google might wanted to see some big iron? My bet is that Google is planning to host more interactive, scale up transaction-oriented services, something its architecture is not currently designed for. Serving search results across a vast index is not the same as serving transactions. Perhaps Google needs help to become a little more enterprisey..

IBM however needs to start thinking less enterprisey and more servicey. It needs to consider some new business technical models for customer engagement and participation. To sum up, software hosting as a service should be a natural fit for IBM, but for some reason it is not. IBM really needs to push on, or it may find itself on the outside looking in as the opportunity explodes.

SAP’s Jeff Nolan is right to say its about software service networks, not product stacks.

Disclaimer: IBM SWG and Microsoft are clients.

Tags: IBM, Software as a Service, SaaS, Google, Microsoft, Don Dodge, eBay, Web1.0, Web2.0, DB2, Lotus, Rational, Tivoli, WebSphere, salesforce.com, appexchange, Googleplex