The Need for New Delivery Models
Recently ordained the topic du jour, it is now apparent that software as a service (SaaS)—also known as on-demand, hosted delivery—models of enterprise applications are not the passing fads that the Y2K or dot-com phenomena of the late 1990s were. The rationale thereof, however, depends on whom one asks for an opinion (that is, whether the opinion comes from someone of the burgeoning SaaS camp or from the traditional and still dominant perpetual license-and-install on-premise camp).
According to the SaaS proponents, the reasons for on-demand hosted delivery models of enterprise applications becoming a full-fledged paradigm shift are the hyped (and often confirmed) benefits for user businesses. While much more information can be found in Software as a Service Is Gaining Ground, in short, the benefits are largely related to the model's keyword “simplicity.” It might even be ironic that this “phenomenon,” “inflection point,” “disruptive force,” or “revolution” in business IT is not about the “next big thing” or “killer application.” If anything, many features and much of the technology had to be stripped down for simplicity's sake, as well as to avoid the usual pitfalls that user companies have experienced with traditionally complex and unwieldy on-premise business software.
For the purposes of this article, the terms “SaaS” and “on-demand” are used interchangeably, but those who wish to learn the differences between the two may refer to What Is Software as a Service?
SaaS applications are conveniently leased by users; they are hosted and remotely maintained in a multi-tenant utility manner by the software vendor as opposed to a user enterprise's internal IT staff. Here is a good analogy to explain what is meant by a multi-tenant utility manner: tenants of a residential apartment building share the same utility services, such as 110 volt electricity, with each tenant using different appliances. The applications are then delivered over the Web as a service.
In other words, under the SaaS model, the software is installed, operated, and maintained on the vendor-owned servers, which are monitored and maintained by the vendor's personnel. While the software is still configured and integrated to the customers' needs, implementation and integration challenges are substantially reduced.
The fact that a Web browser is the only required software at the user's side has propelled Salesforce.com's now well-known mantras of “no software” and “end of software.” While each of these mantras is, of course, a marketing stretch (given that the vendor is still a software company), it is understood that there is no need to have a box-load of installation compact discs (CDs) in house. Bundled with this is the intuitiveness of the user interface (UI). For example, if an intelligent person knows how to navigate through Amazon.com or eBay web sites, then he or she can quickly grasp the use of a SaaS application. At least in the realm of sales force automation (SFA), users have often shown a preference for simpler and inexpensive, “good enough” SaaS offerings over the underused (but much riskier) offerings from traditional counterparts (see Application Erosion: Eating Away at Your Hard Earned Value).
Further, although it is arguable that on-demand software and services are not necessarily a cheaper total cost of ownership (TCO) option in the long run, they seem much less expensive and less capital-intensive in the short term, and they are easier to get up and running. In other words, a SaaS deployment may not be an eventual lower TCO value proposition, but it does allow for more predictable expenditures for users, who in turn can avoid overpaying for unneeded capacity, since there is no need to buy any hardware or underlying IT infrastructure.
In addition, SaaS applications have quicker time-to-value (or, at least, quicker time-to-discontinuation, in the case of a mismatch). This is made possible by the “Internet-native” and multi-tenant product architecture that departs from those of the failed application service providers (ASPs) and similar hosting companies of the dot-com era. These outfits did not fundamentally change the architecture of their software applications; they simply resold client-server applications of the day to user enterprises that did not necessarily want to house them on their own premises. The up-front and ongoing costs of hosting single-tenant applications have apparently proved to be too much for many ASPs to viably hold out. For more information on architectural nuance differences, see What Is Software as a Service?
To that end, SaaS customers today typically buy monthly subscriptions rather than pay huge up-front license fees. For larger corporations, these license fees can easily amount to millions of dollars (especially when one factors in the cost of implementation, which is often in multiples of the software cost) and require lengthy selection processes, contract negotiations, and board approvals. If the software does not perform as promised, or if it has served its purpose, the user companies can simply stop paying, or at least serve the required contractual notice to the vendor.
Conversely, in the “old-school” world (that is, on-premise perpetual license), software was not only initially expensive, but it would take months—sometimes years—to install, with many user companies deciding to subsequently keep it and suffer in silence with a suboptimal system than try to rip-and-replace it with yet another potentially substandard system. The pervading idea that replacing an enterprise resource planning (ERP) system should be approached as if it were brain surgery (that is, to be performed only when the patient is in a coma or is dying) has come in handy for many laggard vendors for a long time.
Benefits of SaaS
The times, however, seem to have changed. Most recent studies by analysts from Gartner, AMR Research, Aberdeen, Saugatuck Research, or THINKstrategies concur that in five years or so, a quarter of all enterprise system deployments may be SaaS-like, even though such deployments currently take a mere single figure percent share of the entire enterprise applications market.
One reason for this predicted increase is that the SaaS segment's current rate of annual growth is much higher and faster compared to that of its on-premise counterparts. Thus, while the prevailing on-premise camp, lead by SAP, Oracle, and Microsoft, has a valid argument in ascribing the ongoing success of SaaS to effective dissemination of information by the likes of Salesforce.com, Google, NetSuite, or RightNow rather than to the pervasiveness of this paradigm shift, on demand is a powerful enough trend that these incumbents have been scrambling to come up with a strategy to deliver some on-demand products.
As a case in point, in addition to SAP's hybrid multi-tenant CRM On-Demand product and Oracle's multi-tenant Siebel CRM On Demand product, other good SaaS examples are Microsoft Windows Live, Microsoft Dynamics CRM Live, and Microsoft Office Live—all online counterparts to Microsoft applications that reside on personal computers (PCs).
During its December 2006 Analyst Summit, SAP previewed a hosted service designed for businesses with less than 1,000 employees. With the product available on a hosted and managed basis only, each user enterprise can have its own separate copy and database (instance) running on a dedicated blade in a large server, and a separate database for storage of its data. SAP maintains the software directly, fixing bugs and providing updates without taking the system down. The systems are configured to match the users' needs, but modifications are allowed. The solution offers a choice of pre-configured industry business models that the customer can select as a starting point for further configuration. If configuration is not enough, the customers can write extensions and integrate through Web services.
SAP believes hosted delivery offers a number of benefits, such as allowing the vendor to closely monitor the performance and characteristics of the software, and to fix functional or architectural problems more quickly. It also provides a much faster feedback loop for the development team as opposed to a typical in-house deployment that traditionally takes months, or even years, to get the software implemented, gather user feedback, make the software changes, and have new releases deployed. In many cases, SAP introduces product improvements without customer involvement or disruption.
In any case, SaaS users can benefit from a completely improved user experience—one that will allow companies to focus on their competencies rather than worry about upgrades, patch management, backups, etc., since these are conducted by the vendor for all subscribers, and often unobtrusively and unbeknownst to users. Vendors too can benefit from the SaaS model in terms of more effective research and development (R&D) efforts. Vendors that have developed their solutions using a multi-tenant architectural model exclusively have the prospect of releasing new functional enhancements more frequently than their on-premise peers. There are examples of SaaS pioneers similar to Salesforce.com and Employease (now part of Advanced Data Processing [ADP]) that have delivered over 20 major releases and several dozen smaller monthly enhancements over the past several years.
Employease cites a baseball analogy, depicting traditional software vendors as trying to hit a home run every 12 to 18 months, and vendors of SaaS as hitting several singles per year during the same time period, to show the difference between the two. In other words, vendors can score many more runs and strike out much less often with the SaaS model, as the reality is that even the winners of the Major League Baseball (MLB) World Series have been the teams that knew how to run the bases rather than those with overpaid sluggers that did not perform under extreme pressure postseason. It is also worth noting that SaaS helps resolve the problem of software piracy, which has become all but accepted in some emerging markets.
These advantages, bundled with more predictable recurring revenue (versus the cyclical and unpredictable income of on-premise purchases), are hard to ignore by any serious enterprise application player. Under this approach, a customer's monthly subscription and hosting fee is substantially less than the typical onetime license fee, but over the life of the subscription, this monthly fee generally produces equal or greater revenue to the vendor. The lower up-front costs and integration hurdles tend to reduce customer approval requirements and to shorten sales cycles. All revenues associated with subscriptions are recognized ratably over the life of the arrangement, including software license and set-up and implementation services, which should result in more stable and predictable revenue (albeit also with more deferred revenue) for SaaS vendors.
SaaS is the Real Deal Then?
Although the SaaS movement has not yet reached the tipping point of obliterating the traditional on-premise model, it has long passed the early enthusiasts' and pioneer stages; it is clear SaaS has taken off well beyond the point of no return.
In fact, Salesforce.com has been in business since 1999, currently has more than half a million subscribers in nearly 25,000 companies worldwide, and is soon to become a $500 million-(USD)-a-year company. Also, having realized that offering only a lightly customizable SFA functionality (which can be commoditized fairly quickly, even by some of the above-mentioned traditional predators) is not viable in the long run, the vendor has reached out to its independent software vendor (ISV) partner. To that end, Salesforce.com has created an ecosystem called AppXchange (recently renamed Apex), which currently boasts over 200 ISVs and 400 SaaS solutions that notably expand the original SFA scope of contact and opportunity management (see Software as a Service for Customer Relationship Management and Sales).
The idea behind the AppXchange platform is to allow partners and customers to build, share, and sell entire applications running within Salesforce.com's hosted environment. This has already prompted similar SaaS platform ecosystems from the likes of IBM, NetSuite, Progress Software, Jamcracker, Oracle, Microsoft, and WebEx (in alliance with Cordys, a provider of service-oriented architecture [SOA]–based development, deployment, and integration tools and infrastructure, founded in 2001 by Jan Baan).
NetSuite has been in business for at least as long as Salesforce.com has, and it reportedly has thousands of customers globally that are using its online applications. These applications include a broad functional scope for running a business well beyond the realm of SFA or customer relationship management (CRM) to tackle the important aspects of order management. The company has espoused SuiteFlex, an application development platform that specifically targets its channel partners. It offers partners the ability to build their own vertical applications on top of the NetSuite business management system, as well as an integrated, customized UI for their customers, and then to have it all available as an online, on-demand service delivered via NetSuite. The vendor has also launched Suite Bundler, which it defines as "service as software," and which should allow the channel partners to code and embed their knowledge of industry best practices and applications directly into the NetSuite system.
An indicator that some SaaS-enabled supply chain management (SCM) solutions are becoming "industrial-strength" is the impressive customer list of Click Commerce's demand chain management (DCM) SaaS solutions (see Will a Tool Manufacturer and a Supply Chain Software Vendor “Click” in Matrimony?). Then, in addition to WebEx's renowned web-conferencing software (with over 2 million subscribers so far), McAfee or Symantec antivirus software, which protect both corporate networks and home PCs, have for too long been used via subscription and "live update" features, representing another example of SaaS. Possibly the oldest and biggest SaaS outlet is ADP, the world's largest human resources (HR) and payroll ASP. ADP has been in business for nearly 60 years, and with about $9 billion (USD) in revenues, has thus far served nearly 600,000 clients worldwide.
Although most of the success of SaaS's market penetration can be attributed to the solution's initial appeal for individual users (free trials and low-cost single-user subscription fees, for example), there remains the hope of impressing the grassroots, and then winning business unit- and enterprise-level adoption in major corporations. Today, virtually any size company can be a SaaS user. For instance, Salesforce.com points out that it has a growing number of Global 2000 and other brand-name companies as its customers, including ADP, AOL, Avery Dennison, Nokia, Perkin-Elmer, and SunTrust, to name only a few. The same holds true for Click Commerce.
The scalability of the new generation of SaaS solutions enables users to test the reliability and performance of on-demand applications in limited (pilot) deployments, and to expand their adoption incrementally. The nature of on-demand applications has consequently allowed for the fragmented and decentralized software procurement model, often conducted by business unit operational business managers rather than the traditional IT ones. In fact, some believe that the future evolution of IT (which has so far been driven mainly by underlying platform shifts, such as mainframe, client-server, and Internet to SOA [see Architecture Evolution: From Mainframes to Service-oriented Architecture]), will be much more about a new economics, such as the paradigm shifts of open source or SaaS deployment models.
For more information and to start your own custom solution comparison, please visit