I'm here in San Francisco at SaaScon, which is the first major conference on Software as a Service since the ASP boom dot-busted back in 2001/2002 (by the way, I'm on the SaaScon advisory board — see disclosure page).
The opening keynote is Tim Chou, who led Oracle's application hosting division — originally called Business OnLine, now Oracle OnDemand. He left a couple of years back and published a book,
The End of Software, and now lectures at Stanford University on SaaS as well as being involved in several startups including on-demand business process vendor nSite, who I wrote about last year (as well as noting that the title of Dr Chou's book is of course a "complete and total misnomer").
The meat of Chou's presentation, which as I write he's just now got to, is the economics of software. He's talking about how IT budgets are largely already committed to maintaining existing software and hardware. Customers may be spending $x million on their IT, but virtually all that money is already committed to what they already have installed. There's nothing left for taking on anything new.
That's on the customer side. On the vendor side, out of every dollar a customer pays, about 15c typically goes towards research and development. But how much of that is writing new software?Probably only 3c goes on new development. The rest of it is testing, back-porting releases, bug fixes, etc.
Looking at the rest of that dollar, 25c goes on support. 40c is spent on sales and marketing. And most of that money is spent, says Chou, because vendors are "fundamentally disconnected from their customers." The challenge for vendors is to "collapse the supply and support chain."
One of the fundamental things that software-as-a-service does is, it allows vendors to standardize on a single platform and therefore allows "specialization and repetition:
"This is so fundamental to everything that happens well that humans have ever done. If you go in for surgery tomorrow, what's the first question you should ask the surgeon? 'How many times have you done this surgery? ... successfully.' Repetition and specialization are the key to excellence."
And the point of repetition and specialization is that it allows commoditization, just as it did in the microcomputer industry with the PC. Software costs can go from thousands per user per year as with conventional software down to a few dollars as in the case of on-demand volume players like eBay.
So what of the future? In the past, innovation has come from inside large corporations. Now, says Chou, innovation is coming from the bottom up, from the consumer end of the scale. He quotes the example of an Australian vendor of ostrich farming software. Delivering software as a service over the Internet provides the economies of reach and scale to make vertical apps like that financially viable.
Finally, there's the effect of delivering software in this model on the businesses that consume it. Chou relates an anecdote about a vendor of sales compensation software (as a service of course) saying that customers are now asking, "What's the best method of compensating my sales people?" Not "What is the best platform to run your software on?" or "Should I go .Net or Java?" What matters is allowing businesses to get at the information they need to do business better:
"We forgot it's not the software that's cool, it's the information that's cool. This is not just about operational efficiently, this is about doing business completely differently."
Chou finishes off talking about the multiplayer online game World of Warcraft. People are addicted to this because it brings people together to complete a task. And that's something that applies to software and business too -- in an on-demand environment, people's skills are visible, and they interact because different players have different skills. Look then at the potential for collaborative development, he goes on:
"Development communities and networks that allow people to build new software are just on the horizon. I like to say to people today, software is free."There are 15 million programmers in the world, and their numbers are growing, especially in Asia, he adds.
Closing point: "The user at the end of the day has been disconnected, has been disempowered. This model completely shifts that. We have the opportunity to shift the playing field here, to move from vendor-centric computing to user-centric computing."