Shawn   [01.25.07 12:47 PM]

FYI that your "Click HERE" link points to a miniscule, unreadable thumbnail of what appears to be an Excel table.

dwlt   [01.25.07 01:08 PM]

The other thing to bear in mind is that it probably took more time for the SaaS companies to persuade customers that they were trustworthy and that using their service was a sensible thing to do. Given the growing acceptance of the SaaS premise, you might expect that it might take less time and money today.

Dennis Howlett   [01.25.07 02:37 PM]

Apart from not being able to read the graphic Will is comparing apples and bananas. Where does this 'conventional wisdom' come from? It might cost little to do a Digg but it costs a heck of a lot to build business processes. What do you think SAPs $400 million investment into mid-range/on-demand is about?

Phil Wainewright   [01.25.07 02:39 PM]

Shouldn't those IPO comparisons be normalized against the prevailing level of the Dow or Nasdaq index at the time?

As to the conventional Web 2.0 wisdom, sure it takes less to launch a software business if you a) pay the founders zilch b) use open source platforms c) use open source tools d) host your servers at $200/month. But building revenues, that never gets any easier.

What's changed more than anything else is that software prices have fallen (and in fact the SaaS model deflates them enormously) and that's why it takes so much longer to build a profitable business.

Will   [01.25.07 03:33 PM]

i fixed the link - glad it sparked some discussion if not universal agreement:)

http://i38.photobucket.com/albums/e116/pricewill/saas.png

Simon   [01.26.07 02:32 AM]

There is also an additional factor is that reliability and features of the average SaaS company is higher than straight consumer web2.0 plays.




Reliability is a huge factor. For most consumer services you don't need uptime of 99.999%, business plays (particularly ones aimed at medium to large business) must have that sort of up time. Each extra 9 costs significantly more than the one before.




SaaS isn't pioneering a new industry, it is moving into a mature industry. Which brings its own barriers to entry as customers have direct comparison when looking at trade offs and customers also have known requirements. SaaS for business needs to allow businesses to do what they currently do. Few businesses with establish software platforms (Oracle, SAP etc.) will switch to a service that does one thing well but they still need to use their existing platforms for everything else. It doesn't make sense economically nor from a internal process point of view. I'm not saying that every SaaS needs to replicate every feature on a platform only the significant features.




Bearing these points in mind, services such as EC2 and S3 offer the opportunity to reduce the cost pressure of reliability. As businesses par down what features are "must have" and support open data formats the minimum feature set for SaaS will slowly reduce making it easier for these companies to reach the profitability.

Jesse Kliza   [01.26.07 01:56 PM]

Here's a great post that I just replied to, that I think really dovetails nicely with Will's:

http://www.saasblogs.com/2007/01/25/it-costs-more-to-be-a-saas-company-how-platforms-may-fix-that/

Dennis Linnell   [01.27.07 08:57 AM]

Are you sure Will Price's analysis is economics? It's risky to draw any conclusions from the data presented there. Price listed six public companies used for the SaaS analysis and let's assume the "universe of 1990s client/server companies" is the eleven companies mentioned in http://willprice.blogspot.com/2006/12/when-it-goes-right-what-does-it-cost.html

Looking at the client/server list, I see a strong tilt towards security and infrastructure software in contrast with the application orientation of the SaaS list. What's more, nearly all the listed client/server companies became recognized brands, while only one of the SaaS companies enjoys any name recognition at all. (How many of those stock symbols did you recognize?) I think the small sample size and selection bias in both analyses makes them, as physicist Wolfgang Pauli once famously said, "not even wrong."

Will   [01.27.07 11:52 AM]

Dennis is certainly right to suggest that statistical significance requires a sample size of at least thirty. The challenge with the SaaS category is that the pool of venture-backed SaaS companies remains small, with Netsuite and a few others poised to file their S-1s.

For me, and others are certainly free to differ, the data and the public record suggest that SaaS companies require significant capital investment and that the pioneers in the category were forced to not only sell a new delivery model to customers (ie hosted) but also to invest in bespoke data center infrastructure to deliver the application. As Adam Smith(who was an economist:)) famously said, specialization is a function of market size, and as the SaaS market both grows larger and matures specialists will need to emerge to handle certain key functions of the stack that previously were vertically integrated within a given vendor. This specialization will reduce costs and drive break even thresholds down. As investors and software developers, the question is where to specialize and how to attract developers to a given platform. Apex's focus on the developer community is an example of this specialization in action - both with respect to function and to business model.