Alright, first I promise that this is not another bash Lawson CEO Harry Debes blog, no matter how tempting that might be it would be just too easy and too much piling on…besides what could I say that hasn’t been said? If you haven’t read about it, try my friend Paul Greenberg’s blog.
This is a question that I’ve pondered for quite awhile as I’ve watched pure-play SaaS vendors grow and get traction in the software markets. First a few definitions are in order I think. When I say pure-play SaaS vendor I mean a software vendor that derives substantially all (I would have said all, but since Salesforce.com bought an on-premise software company recently I’ve added the “substantially”) of its revenue from subscription based licensing of its software products and the infrastructure necessary to run it. An on-premise software vendor then would be a vendor that derives substantially all of its revenue from SW products that are sold on a perpetual license + annual maintenance business model and of course hybrid vendors derive revenue from both models. Salesforce.com and NetSuite are examples of pure-play SaaS vendors, Oracle, SAP and Microsoft are currently on-premise but are moving to hybrid models.
I decided to do some basic analysis to see if I could start to understand what is happening to pure-play’s versus traditional on-premise vendors. This is somewhat complicated for a variety of reasons, pure-play vendors are at a much earlier maturity level and their revenue model is very different than the on-premise model that takes license revenue up front and then builds a recurring revenue stream around maintenance (support and software upgrades / fixes). The pure-play revenue model is built around a monthly fee from their customers with no up front license or on-going maintenance, and everything is included in the monthly fee, support, infrastructure and the software. The monthly fee is generally variable and is set by number of users and also may vary based on the functionality used. In addition, contracts for pure-plays can be 1-3 years (or possibly longer although I don’t know of any vendor that has managed to tie up customers longer than 3 years). With that in mind, what metrics would provide a meaningful comparison? I collected the following to use in this analysis. The data I took from Yahoo Financial rather than using IDC data for a couple of reasons but mostly so that it would be publicly available, tied to each companies reported fiscal year data (IDC tracks data on a calendar year basis) and so that I’d have matching employee count numbers.
Company———annual revenue–profit—employees—-rev / emp—profit / emp
Note: Revenue & Profit are in Billions USD, Rev/Emp and Profit/Emp are in thousands USD.
I used Salesforce.com since it is the most visible of the pure-plays, has the highest revenue and has been public the longest. It is the smallest of the five vendors examined, still it does seem to be a reasonable comparison, as far as snap shots go. From a productivity stand point there’s no issue (using rev / employee…although it does seem like all of the vendors could use a productivity lesson from Adobe). SAP’s the least productive using this metric, Adobe the most, but in the middle the other three are relatively the same. Salesforce.com was founded in 1999 and really started to gain traction in 2001 so let’s say they have worked 7 years to get to this point. It does seem that there is a difference in the profit model for pure-plays versus more traditional vendors. It’s not surprising though, if you look at the differences. SaaS vendors, at least the larger ones today, have a large investment in infrastructure and operations that traditional vendors do not have. That changes the break even point and the profit model. And as a SaaS vendor scales up for more customers they have to continue to invest in infrastructure. This is driven by number of customers, bandwidth and geographic location. Depending on the pricing model, the SaaS vendor would, at some point overcome the investment costs and at least in theory, grow revenue more rapidly than costs…when they reach a certain infrastructure scale level.
Now don’t misunderstand me, I do believe that the value of SaaS deployed software to customers makes a compelling business case. I do believe that SaaS deployed solutions will continue to grow and more and more vendors will start to offer SaaS solutions. But the numbers and the concept makes me wonder if there are some tweaks to the business model that are necessary to build a reasonably profitable business. One thought is that maybe the applications business alone isn’t enough, which would validate Salesforce.com’s platform as a service concept. If they can build an ISV ecosystem that is using their PaaS then they would have a very leveraged source of revenue / profit that could rapidly outpace application revenue…guess we’ll see how that strategy plays out.
The second thought for a new pure-play would be to use an alternate method of gaining the necessary infrastructure. Rather than investing in its own infrastructure up front, use a model where it pays for the use of infrastructure from a third party in a scaleable approach that increases only when the revenue / number of customers increases. This model is currently available from IBM, which already had a size-able infrastructure, and from Microsoft for its distribution partners. This approach could lead to break even / profit much more quickly and at the same time leave the management of the infrastructure to industry giants that have the expertise, the highest level of security, scale, etc.
The last thought I have around this for now is that perhaps the hybrid business model offers a more profitable and more easily scaled model. All of the large vendors are moving to this, mostly because of market pull for them to offer SaaS solutions alongside on-premise software. Certainly I believe this matches the buying / deployment model that we are seeing more and more. Customers are using on-premise for some applications and SaaS for others, this trend started a few years ago and is increasing as more of both types of solutions become SOA enabled (and thus much easier to integrate). It’s still early in the evolution of the business model to tell for sure which answer is the best, but it is an area to be watched.