It's a whole 'nother ball game

When Steve Ballmer was gearing up for Microsoft's launch of on-demand services last October, I don't think he really believed some of the claims he was hearing from the Microsoft Live team. Bi-monthly product release cycles? His choice of metaphor to describe these faster product updates was eloquent: he called them 'twitches'. It sounded a bit derogatory, as if these frequent, mini-updates weren't really on a par with Microsoft's usually megalithic, four-to-five-year release cycles.

He had so little confidence in the short timescales that when he talked about them publicly'This is software, Jim, but not as we know it' he mentioned six- to nine-month intervals, which in turn led me to describe Microsoft's Live strategy as "barely twitching," because I knew that any serious on-demand vendor is constantly adjusting and improving their product, with the industry average at close to a six-weekly update cycle. Add them up cumulatively over the course of a few years, and those so-called 'twitches' will easily add up to just as much — often more — than any one of those oft-delayed, single mega-upgrades for which Microsoft has become infamous.

Compare Ballmer's imagery with MSN executive David Cole's this week in an internal email to colleagues: "rolling thunder" is his metaphor for the rapid product release program that MSN has been achieving. And Ballmer has had to put his skepticism to one side; Live really has been achieving a lot, with more than a dozen new products introduced since its launch last November (even though, as I've frequently mentioned, it has to do a whole lot more yet before its on-demand offerings come up to the mark).

A lot of software bosses over the next year or so are going to be making the same journey of discovery as Ballmer has begun. The received wisdom about on-demand applications are that they're just a subset of conventional software, differing only in the mechanics of delivery. The people who run SAP, for example, still believe that, and I'm sure Steve Ballmer was still struggling with that mindset six months ago during the preparations for the launch of Windows Live.

In fact, on-demand applications are a whole 'nother ballgame — which is why personally I try to avoid the popular phrase software as a service (SaaS) since I feel it's a phrase that's born of the 'nothing changes' mindset. As I once wrote of the move to services architectures, this is Software, Jim, but not as we know it.

A gathering called the SaaS Summit, which took place a couple of weeks ago in Napa Valley, California, gave attendees a chance to explore the differences and discover what it really means to be an on-demand (or SaaS, if you must) vendor. Several of them have been emailing me their impressions. What comes over most is how much there is to learn about being an on-demand vendor compared to the conventional perpetual licence software model. The way that you price, how you partner, your go-to-market strategies and, of course, how you develop and operate the software, are all quite distinct from how traditional software vendors have worked.

Mark Miles, CEO of Baridata, which offers a 100% hosted medical enterprise solution, told me:

"The summit really opened my eyes to how unique the SaaS model is. We are fully embracing the SaaS model and have adjusted our marketing and pricing strategies as a result [of attending the Summit] ... This has resulted in a complete revamp of our pricing model (moving to per unit pricing) that we will use at launch and beyond."

Jeff Kaplan from THINKstrategies, a consulting firm that specializes in SaaS and IT service, and who spoke at the conference, highlighted the importance of working on ecosystems: "SaaS players that fail to build broad-based partners will fade away," he said, citing Salesforce.com with its AppExchange platform as the pre-eminent example but also noting "IBM and Microsoft understand the SaaS market dynamics and importance of ecosystems, and they are focusing on enabling SaaS as much as offering proprietary SaaS solutions." OpSource, which hosted the event, was also doing a good ecosystem job, he said. He has jotted down some more impressions of the event on his own blog.

At least there are beginning to be some established models out there for new players to follow. Mike MacDonald, CEO of Visual Mining, highly rated a presentation by OpSource's Dave Engelbrecht on how to make the SaaS model work:

"Dave devoted some time talking about how SMB players can gain advantage by 'drafting' (the autoracing term) behind the market leaders," he told me. "In other words, leveraging the marketing engines, pricing models, and customer experiences of Salesforce.com, NetSuite, OpSource and others that are making the SaaS model work."

It's still early days for this industry, though. "It was an early stage crowd," wrote Mike. "By that I mean that the attendees were largely from software companies either already doing SaaS, considering it, or scoping competition. Not too much in the way of large end-users."

Kaplan's impression was that "The SaaS market is almost at the ‘gold-rush’ stage ... players from nearly every vantage point in the IT industry are seeking to capitalize on the opportunity."

In some ways, this is a throwback to 1999, when there was a whole lot of excitement about ASPs, and similar conferences attracted lots of vendors but a dearth of real users. Long-term industry watcher Amy Wohl, who I ran into at many of those early conferences, was a speaker at the OpSource event so I asked her whether it was a case of 1999 all over again. "I think the OpSource conference had the energy of the 1999 conferences without the naivety," she told me. "Much more emphasis on business models using partners, focus on what the advantage is and who the customer is going to be. It looks good."