Wednesday, January 24, 2007

The Economics of SaaS: We Need a Platform

Recently, I wrote a post titled “When it Goes Right, What Does It Cost to Build a Great Software Company?"

Based on a universe of 1990s client/server companies, my analysis found that:
  • Median Capital Raised: $10.1m
  • Median 1st-4th Year Revenue Ramp: $.1m $.8m $6.9m $21.6m
  • Median revenue at profitability: $48m
  • Median Years to Exit: 4
  • Software model supports the creation of great companies on <$15m of capital
The post begged the question regarding the metrics associated with SaaS. This post, based on a review of the current public SaaS comps, is the SaaS analog. The analysis includes: WSSI, CRM, OMTR, RNOW, TLEO, and VOCS.

While the SaaS companies grew up in a different IPO market, the results suggest that SaaS companies take:

  • 1.6x longer to get liquid
  • 3.65x more capital
  • 1.75x more revenue to hit profitability
  • Salesforce, for example, raised $64.52m in equity, to Peoplesoft's $10. Websidestory raised $43m to BOBJ's $5m.

Click HERE to see data behind analysis

SaaS is here to stay – the advantages to customers and vendors are well established.

The remaining challenge, however, is how to build viable SaaS companies more cost effectively. Will the rise of AMZN S3 and EC2, Apex and AppExchange, etc eliminate the need for bespoke infrastructure investment. We will stop stop asking about MSDN subscriptions and ask instead about AMZN subscriptions?

One can only hope some form of platform infrastructure emerges to accelerate SaaS companies development. If not, the merits of SaaS will be challenged by the time, capital intensity, and delayed profitability of the model. Platform companies – Powersoft/Sybase, ORCL, MSFT – drove down the costs of building client/server application companies. The industry needs the SaaS analogs to unleash the power of the model at the cost optimal level.

A simple analysis holds that Fixed Costs/Gross Margin = breakeven revenue. While for SaaS this is a somewhat circular calculation (as in SaaS fixed costs are amortized into COGS), the rise of platforms will drive down fixed and allow SaaS companies to reduce capital required to get to scale. Fixed costs must be reduced in order to unleash the full power of the model and the rise of platforms will reduce the bespoke investments historically required to build SaaS companies.

Now, the data above is for public companies who came to prominence in very challenging times. We are active SaaS investors and are already seeing the fruits of leveraging commodity platforms and partners to build companies more cost effectively. However, the public record to date suggests that the SaaS industry remains relatively immature without the obvious parallels to the client server tools and server companies that drove the success of huge numbers of application companies in the 1990s.

3 comments:

  1. Of course you can't draw parallels between SaaS and upfront license models. Why bother when they're fundamentally different?

    It's not the economics of SaaS at all but the economics of startup.

    "...Platform companies – Powersoft/Sybase, ORCL, MSFT – drove down the costs of building client/server application companies..." Are you serious? Do you know what it costs to buy Oracle licenses? Do you know the total stack cost for MSFT? Today. A very different story when you wind the clock back to the C/S development days.

    Where's the inflation adjustment?

    "...record to date suggests that the SaaS industry remains relatively immature without the obvious parallels to the client server tools and server companies that drove the success of huge numbers of application companies in the 1990s...." And what do we have today? Very few standing and outside of MISO?? Not a lot.

    I'm sorry but this is a misleading
    analysis as it stands.

    ReplyDelete
  2. Anonymous3:39 PM

    I really liked your analysis on SaaS, but I think that it was a bit unfair to SaaS. Most of the differences has to do with comparing companies from different eras (the Walter Johnson / Roger Clemens, Bill Russell / Shaq issue). It seems as if the differences more adversely effect the SaaS world and the companies that were build during the biggest tech dry spell over the entire period.

    Time to liquidity - from 2001-2003 there were very limited liquidity events; these companies were growing and profitable and in a "normal" market would have been public; a way to show it would be the sales at liquidity for the two groups
    Capital Raised - If the IPO window was open, this number would be decreased; also although SaaS cos are "financing their customer" compared to client server - there are other benefits
    Ramp - although impossible to do it would be interesting to use a like accounting method (same rev rec) for ramp; my guess it that SaaS scaled faster due the shorter sales cycle; same for profitability.

    Just some thoughts.

    ReplyDelete
  3. Jesse Kliza1:55 PM

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

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

    I would love to hear your comments on it.

    ReplyDelete