Software as a Service
by Judith Hurwitz, CEO
I imagine a conversation between two CIOs 15 years from now. They are comparing notes and reminiscing about the days when they had to manage all their hardware, business applications, network infrastructure and the like in their own data centers. It?s painful to remember the complexity of those days. Life is certainly a lot easier now that we can leave computing infrastructure to the professionals?
I predict that Software as a Service?the delivery of application functionality through a subscription ? will be the norm. The notion that companies will own and run their own applications will be a relic of a bygone era.
Ironically, the computer industry started out with this model. Until the mid-1970s, most companies that computerized their businesses subscribed to timesharing services. In those days, computers were simply too expensive for one company to own. Although hardware costs have plummeted, running software applications and maintaining performance standards cost companies plenty.
When the pioneers in software as a service such as Salesforce.com first entered the market, many CIOs were skeptical. In fact, the requirements for software as a service were largely driven by business unit management who demanded efficient implementation without the overhead of running systems. It is little wonder, therefore, that the earliest success with software as a service has come from the Customer Relationship Management (CRM) market.
Many IT organizations reluctantly went along with the experiment, many harboring the expectation that once these applications became ?mission critical? they would move to a software licensing model. It didn?t happen. Salesforce.com, for example, was able to persuade management that the cost to license their software would be prohibitively expensive (if it were even an option) ? especially when adding in the costs for hardware, systems management, backup, and the like.
Two key issues fundamental to the growth SaaS are licensing models and the partner ecosystem.
From a vendor perspective, one of the most important aspects of SaaS is the end of the perpetual license model for software — a radical departure from the way customers have typically acquired software. Customers will slowly become accustomed to this model. Initially, we have seen acceptance in markets such as CRM. Many customers like the idea of being able to try an application before they invest in implementation and training. They also like the ability to use an application they need twice a year without having to buy a license.
One of the major differences between SaaS models and conventional licensing is pricing. Typically in SaaS, customers pay based on number of users. There may be variations in this flat rate plan if an application requires huge volumes of storage. For example, one user might want to store a terabyte of data. Invariably, there will be an extra charge when storage is part of the equation. We predict that it will take as little as five years and as long as ten before customers are comfortable giving up the one time payment of the perpetual license. This movement from the perpetual license goes hand in hand with issues of service and support. The stakes are high. If a customer comes to rely on a service platform, availability is paramount to success. For example, Salesforce.com was severely criticized by customers when it suffered a number of service outages.
We predict that the partner ecosystem in which software companies partner with infrastructure vendors to bring products to market will be the lynchpin of the SaaS market. As with conventional software, it will be difficult for emerging players to go it alone. We expect to see many large infrastructure players make a big play for this market. SAP, for example, is beginning to offer some of its applications as a service (starting with CRM); IBM is expanding its PartnerWorld Industry Network program to a wide array of SaaS players. IBM is offering both hosting services directly for its SaaS software partners and offering its ISV partners? software directly to customers.
Hurwitz & Associates believes that SaaS is here to stay and will continue to grow and envelop more areas of software. Here are scenarios that make good candidates for SaaS:
- When applications are used by a highly distributed workforce that needs to be able to access the applications from many different computing environments and where backup and service levels need to be guaranteed
- When applications are used infrequently but their use is critical
- When extra capacity (hardware) is needed such as grid-based systems to support customers? peak load
- When applications offer shared services between partners (various collaboration applications, and ecommerce, etc.)
Does it always make sense to adopt Software as a Service? This is a tricky question. Long term, software as a service will benefit enormously from Service Oriented Architectures (SOA). One of the benefits of SOA is that it includes a set of defined interfaces between business services. This will help insure that services provided by various vendors will be able to work together without requiring customized coding. In fact, SOA will create a new channel for SaaS as Value Added Resellers, global systems integrators, and Regional Systems Integrators provide services to companies that want to leverage these offerings while adding connective tissue and additional functionality.
What are the areas where Software as a Service is gaining traction outside of CRM? We?ve seen: Human Resources applications, applications that help companies comply with governmental regulations such as Sarbanes Oxley and SAS, data analytics applications that are used periodically to analyze complex customer data and service management applications designed to monitor issues such as application and system performance, just for starters.
The bottom line. The momentum is building for software as a service. We believe that this model, as subset of SOA, will dramatically change the computing landscape. The only issue that threatens this model is if vendors are unable to create and maintain a reliable, predictable level of service.