The On-Demand Cult

During Dreamforce there was a panel discussion for media and the financial and industry analysts. Some of Salesforce.com’s senior executives were there for a panel discussion, which included Chief Financial Officer Steve Cakebread, cofounder and Executive Vice President Parker Harris and Vice President of Marketing George Hu, as well as company CEO and founder Marc Benioff.

The panel was one of those things that most companies do in one form or another at user conferences. They are an important part of a company’s outreach to some constituencies that are very important in helping shape the company’s image. Salesforce both taped and broadcast the discussion and you can hear the whole thing here.

Despite the fact that there were several senior executives present, the invited guests really wanted to hear from Benioff, and he was happy to answer their questions. Dreamforce was a big moment for Benioff and his company as they were carefully trying to reposition themselves from being solely a CRM company to being something more akin to what Benioff called a multi-application and platform company.

Single vs. Multi-Tenant Architectures

For the purposes of this piece, the most interesting part of the discussion begins around the 45-minute mark of the taping. At that point an analyst asked a question about other delivery models that used single tenant architectures as opposed to the multi-tenant architecture pioneered by Salesforce.com and others in the early years of this decade.

While the difference between the two might seem small, there is a great chasm between them. Single tenant effectively makes a big server operate like a bunch of little servers — each company subscribing to an application gets its own database and virtual machine. In contrast, multi-tenant gives all users the same access but segregates data. Multi-tenant is harder to do, but the benefits include scalability and cost effectiveness, and each of those is a major benefit.

Benioff’s answer to the “single vs. multi” question took two tracks. First, he pointed out that single tenant had been tried before by companies like USI and Corio and “they didn’t do so well” in part because they couldn’t scale sufficiently, which put pressure on their pricing models.

The other argument is practical use. If you look around at the major Web entities today — companies like Google, Yahoo, eBay and may others including Salesforce.com — their architectures are remarkably similar, they are multi-tenant.

There’s no doubt that multi-tenancy is the heart of on-demand computing, and, in my humble opinion, it has proven to be the solution that was needed by those envisioning a very different and more democratic form of computing many years ago.

Retreat Up Market

I bring all this up to contrast what one vendor was saying to analysts like me. I took a briefing with Callidus, a company with products in the compensation management market including sales performance management.

Callidus is a traditional software company that gained significant market share selling installable software. Not wanting to miss the on-demand wave, Callidus recently came out with an on-demand solution that implements a single tenant architecture.

They’ve also begun talking about a hybrid approach that incorporates on-demand and traditional solutions to give customers more flexibility and choice.

It’s a classic strategy by an incumbent to downplay the importance of a disruptive technology. Clay Christensen documented it in the “Innovator’s Dilemma,” and it is remarkable how consistent it is. Inevitably, it results in a retreat up market — a situation where the company loses market share at the low end while clinging to large late adopter customers. In this case, one can only assume Callidus is feeling some heat from Centive and Xactly, two multi-tenant vendors gaining traction at the lower end of the market from Callidus and moving upstream.

Disparaging the Disruptor

Another thing incumbents do in this situation is to disparage the disruptor, usually while trying to imitate the disruptor’s success. So, for example, in a short essay penned by Robert Youngjohns, president and CEO, they even refer to on-demand as “a cult.”

“So what exactly is it?” he asks. “At best, it’s a marketing catch phrase. At worst it’s a cynical attempt to separate eager VCs (venture capitalists) from their cash.”

Hmmm… I’ve heard VCs described in many ways in my career, but never as innocent babes in the woods. Try this: Repeat after me, “Those poor, vulnerable VCs!” Just doesn’t work, does it?

Even in the face of this disparagement and attitude toward on-demand, Callidus still made the decision to attend Dreamforce and to have a booth. Go figure. Callidus has a good traditional product and large market share, two things that financial analysts love, but that’s not the point.

At the panel I mentioned at the start, there were industry analysts and financial analysts. After all these years, the industry analysts are still ahead of their financial brethren when it comes to understanding on-demand, though the gap is narrowing. The finance guys still need education about the revenue model, and it is still a challenge for some of them to put a value on a Software as a Service company.

The point is the future — and I think Dreamforce gave us a pretty convincing view of what the future looks like. It’s all about multi-tenant, on-demand, utility computing.