For a little over a decade now, software companies have been attempting to transition away from the feast or famine quarterly grind of perpetual licensing sales designed in the 1980s and perfected during the client/server decade of the 1990s. The first Application Service Provider (ASP) models at the turn of this century were ahead of the provisioning capabilities at the time, but were the start of a revolution that just may help shareholders not only survive but even thrive during this increasingly nasty worldwide recession. Or so we hope. Like all things evolving, there are variances and nuances to the recurring services software model that, for technology investors, may mean the difference between solid performance and "what just happened?" cratering.
There is a reason why, in recent years, analysts have fallen in love with the recurring services software model. At scale, a well designed model offers predictable revenue streams and cashflow, scaling leverage, and resiliency against the impact of economic slowdowns. More importantly, on-demand software is increasingly the way many customers can afford to pay for reliable technology.
The challenge of this model is that it is expensive to scale to a positive earnings inflection point where free cashflow gets generated. In other words, a lot of capital gets burned on the way to success. One of the pioneers of the model, and arguably one of the most successful perveyors of the Software as a Service (SaaS) is CRM on-demand software maven Salesforce.com (CRM-NYSE). The Company is legendary for taking on and pummeling CRM perpetual licensing heavyweights Peoplesoft and Seibel, which are both now divisions of Oracle (ORCL-Q). Considering that Salesforce.com is trading at 104x TTM P/E, investors still love the stock (even with a recent 40 minute service outage). Regardless of the economic conditions investors believe that, for the most part, Saleforce.com clients will pay a modest monthly fee and follow it into the cloud as the Company extends services through its app server infrastructure. In a final irony, Oracle is now stating that it is ready to take on Saleforce.com head-to-head with its own hosted services. See what Larry Ellison thinks. To get to its profitability tipping point, Salesforce.com burned through a lot of capital. NetSuite (N-NYSE), another on-demand pioneer still is - over $6 million last quarter. From 2003 to 2007, it was ok to burn through vast amounts of capital to scale. Is the capital still available to assist on-demand growth, and if not, can vendors leverage the model and adjust towards less lofty scale but greater profits?
Most software and media companies (there is not much difference anymore) claim to have launched, or are about to launch some type of recurring software service. As nearly the entire sector (Even Larry Ellison) transitions to the recurring revenue model, it begins to lose its purity. Once diluted, the model begins to morph and also lose some of its famous revenue attributes such as predictability, leverageability, and scalability.
After reviewing hundreds of these types of Companies, here are some points to ponder when yet another on-demand (insert service here) investment thesis is presented.
1. Renewal Risk. Although the recurring services model is fairly immune to the quarterly life-or-death struggle of the perpetual license unit sales, there remains some pretty significant revenue risk associated with multi-year contract renewals common to the model. I am aware of several SaaS vendors already beginning to negotiate major client renewals that are due in 2010. Earlier stage vendors with customer concentration risk could suffer devastating consequences if key clients switch or abandon. Clients are aware of this and are eager to leverage this risk into better renewal terms.
2. "Devil is in the Detail" Risk. Not all recurring revenue is created equally. Variance in the granularity, scope, and input costs associated with the recurring revenue streams create variability in risk. As well, key terms such as length, minimum guarantees, automatic renewals, and capital equipment commitments impact the marginal value of of recurring service revenue to investors.
> Granularity: monthly revenue can be charged at the server level, on per seat or subscription basis, or at a transaction level. At the highest level (server level), a multi-year SaaS contract would be less sensitive to economic cycles, including short-term economic downturns. Typically, transaction models are consumer facing and more economically sensitive, and are likely to vary with demand elasticity. The more transaction-oriented the license, the greater the potential risk may be to earnings as activity declines. However, there is a contrarian perspective...
>Scope: During bad economic cycles, sometimes it is better to be under the radar. Data and application services that cover a broader scope of bundled functionality may be at greater risk of contract erosion through client unbundling. A primary example is financial data services. Stressed brokerage firms are cutting back budgets by actively unwinding premium bundled seat licenses from providers like Thomson Reuters (TRI-NYSE). In the meantime, low-cost individual data subscriptions are escaping the purge by being relatively unnoticed (at least for now). Ironically, the more entrenched a vendor, the more likely that it will be targeted for ad-hoc contract renegotiations while the unnoticed little app runs quietly in the background.
>Bundled capital risk: many recurring services require some type of specialized hardware or firmware for the service to work. In order to secure multi-year services contracts, vendors will often eat the equipment costs through financing, and pass on the expense through increased recurring licensing fees. In capital constrained times such as now, equipment bundling removes purchasing barriers, which is positive. On the other hand, this tactic reduces gross margin benefit, while increasing overall capital risk to the vendor. If the financing cannot be secured in reasonable timeframes, it increases contract and deployment risk.
>Term: longer-term contracts, such as 5 year and 3 year terms are typically lower risk that 12 or 24 month terms in a downmarket. The main risk related to long-term deals is clustering. Having multiple long-term contracts renewing during 2009 is probably not a good spot to be.
3. Stage of Development: This is an interesting variable for investors. A mature on-demand model can generate between 20% and 30% pre-tax margin. However, the greatest earnings lift comes from Companies who have hit the earnings inflection point within the previous 2 quarters, or are about to hit it within 1 quarter. Doubling, quadrupling EBITDA with similar increases in cashflow are common at this stage, even with modest growth. Earlier stage investments into rock solid solutions with the potential for market dominance could create the most long-term return if one has a 5 year investment time horizon and is patient. There is not much patient in these markets.
The recurring services revenue model, as it has evolved since 2000, has not been tested during a significant economic contraction such as the one that we are currently experiencing. One could expect that financially distressed clients could simply walk away from contracts, or force renegotiated terms. However, there is a lot of margin to play with. I am aware of some on-demand vendors already deciding to cut back on growth initiatives in order to maximize margins on forward contracts. Whereas topline sales are forecasted to grow by less than 10%, bottom line marginal contributions are forecasted to triple or even quadruple. If investors are looking for growing cashflows, then technology companies with strong recurring revenue streams may still be a good spot to look.
I do not own shares in any of the companies discussed in the post.