ANALYSIS
The DIY Argument

Earlier this month SugarCRM announced that it had secured US$20 million in a new financing round, bringing its total funding so far to $46 million -- a significant accomplishment for several reasons.

If SugarCRM is not familiar to you, it's not a new Caribbean restaurant in SoHo. I never bothered to find out exactly where they got the name, but Sugar is an up-and-coming CRM vendor with a very different take on the market. The company's signal differentiator is that it is an open source product. In other words, they freely give out the source code, and a community of like-minded developers contributes to the code base by writing modules or simply providing enhancements and fixes.

Give Me Code!

You might be familiar with the open source movement from experience with Linux, the open source operating system based on Unix. Increasingly, companies that want to run as lean as possible put their applications on Linux servers, and that includes on-demand solution providers, so there's a good chance that open source is already in your life.

What's interesting to me about open source, and Sugar in particular, is that in some ways it looks like a throwback. Unlike on-demand software that seeks to hide all of the complexity of the operating system, database and code from the typical developer, open source puts it all out there for anyone who wants it. From the looks of it, there is a thriving and growing community of users who still want access to the code.

Sugar appears to be well positioned to support all tastes and all opinions in the on-demand vs. on-premises debate. The company has neatly side-stepped the issue, effectively saying, "Here it is; do whatever you want." As a result, some customers operate in an on-demand fashion, while others have a traditional IT strategy that brings the technology in-house.

It might seem heretical to on-demand true believers, but there are many companies that either don't want to rely on "cloud computing" or who are constrained by regulations to keep close tabs on their data. Who's to say who is right? There are banking regulations and government edicts, especially in the European Union, about where data can be stored -- and for the sake of argument St. Moritz sounds a lot better to European ears than San Diego. There is absolutely no doubt that data center location will become a larger issue as American software as a service (SaaS) companies try to gain altitude in other markets.

All Are Not Created Equal

Executives at Sugar tell me they took the $20 million for expansion reasons, not to fund day-to-day operations, which to me says business is pretty good. Sugar is part of a growing list of companies that are trying to jump on the SaaS bandwagon while offering their own interpretations on how to get the job done.

You need to add to the list some heavyweight names like Microsoft (Nasdaq: MSFT), Oracle (Nasdaq: ORCL) and SAP (NYSE: SAP), but all are not created equal. As the riffs on SaaS proliferate, it will take a fine-tuned ear to listen to the value propositions and determine if said propositions are right for any particular circumstance.

Here are some things to consider: First, just because a vendor offers you an on-demand or SaaS solution, it doesn't mean that the offer is equivalent to another SaaS offer. SaaS is not table stakes. Consider who is doing the hosting and generally what's on the other side of the cloud. The gold standard is something called "mirrored data centers."

As the metaphor implies, mirroring means that the system is making an exact copy of everything that happens at another location. In the event of a catastrophe, the mirrored data center should be able to take over with minimal disruption. However, few SaaS providers have ponied up with the capital needed to build mirrored data centers. Often, if you read a vendor's IPO filings, you will see that one of the uses for the capital raised in an IPO is for just this purpose.

Dig Your Own Well?

So, one of the dangers of going with a reseller of a SaaS solution might be the reseller's ability to keep you running in the event of a disaster. Some vendors might counter by telling you there is a tape back-up, but then you need to ask how long it takes to read all the back-ups back into a working system.

The other issue offered by some vendors like Sugar and Oracle (with its grid computing architecture) is the ability to bring together multiple systems in a heterogeneous configuration that contains both multi-tenant systems and single tenant systems to provide a kind of de facto redundancy. If grid computing sounds appealing, I recommend you let the vendors tell you about it.

So, to net it out, it looks like the on-demand debate has some distance left to travel. The idea of a single, centralized utility model is certainly one possibility, but the market appears to still be experimenting, and Sugar's $46 million in funding is one example of how robust the experimentation is. To borrow from a utility model argument from a while ago, it appears that some people are still content to dig their own wells, eschewing town water. But maybe the metaphor is more like the argument for solar panels -- sometimes there's an advantage to doing it yourself.