Is it less expensive to use Software as a Service SaaS than to purchase software for use on your premises? Research firm Gartner Group has issued a warning to CIOs not to assume that SaaS will in fact turn out cheaper in the long run.
"In recent years there has been a great deal of hype around SaaS," stated Robert DeSisto, VP and distinguished analyst at Gartner, the information technology research and advisory company. "As a result, a great number of assumptions have been made by users, some positive, some negative, and some more accurate than others. The concern is that some companies are actually deploying SaaS solutions, based on these false assumptions."
SaaS is cheaper during its first two years of use, Gartner finds, but the total cost of ownership over five years would be lower for on-premises software. It also warned that while most users will assume that they will be paying on a 'pay as you go' basis, there are still likely to be contractual considerations. In "the vast majority of cases," Gartner says that companies are pushed to sign predetermined contracts with fixed fees.
In its report, Fact-Checking: The Five Most-Common SaaS Assumptions, Gartner also warned that SaaS is not necessarily faster to implement. While vendors quote 30 days as the normal implementation time, "some software can still take up to seven months to set up."
Another assumption often made is that it is difficult or impossible to integrate SaaS with on-premises applications or data sources. Gartner advises that businesses need to remember that SaaS applications can be customized and are no longer only for basic functions and that data can be initially loaded to a SaaS application, then updated regularly or updated in real time using Web services.
Gartner took the top-five assumptions that users make and provided a fact check on their accuracy.
Assumption 1: SaaS is less expensive than on-premises software.
Fact Check: True during the first two years but may not be for a five-year TCO. SaaS applications will have lower total cost of ownership (TCO) for the first two years because SaaS applications do not require large capital investment for licenses or support infrastructure. However, in the third year and beyond, an on-premises deployment can become less expensive from an accounting perspective as the capital assets used for the on-premises deployment depreciate.
Assumption 2: SaaS is faster to implement than on-premises software.
Fact Check: True for simple-requirement SaaS, which will be faster, but growing complexity and other factors are coming into play. There is a danger in applying the general rule of SaaS being faster to implement for a specific deployment. Vendors often quote time frames of 30 days to implement but neglect to say that SaaS deployments can take seven months or longer. As the complexity of the business process and integration increases, the gap advantage between SaaS and on-premises deployment times will narrow because a larger percentage of the deployment time is associated with customization, configuration, and integration, which are equally difficult with both delivery models.
Assumption 3: SaaS is priced as a utility model.
Fact Check: False in the vast majority of cases. Many SaaS vendors state that they are utility-based providers, similar to electric companies, claiming that you're only charged for what you use. However, for most SaaS deployments, this is false. In the vast majority of cases, a company must commit to a predetermined contract independent of actual use. In some cases, the application lends itself to metered use - for example, an e-commerce application may have pricing based on order transaction processes - but for the most part, utility examples are in the minority
Assumption 4: SaaS does not integrate with on-premises application and/or data sources.
Fact Check: False. There are two primary methods of integrating SaaS offerings with on-premises applications and/or data sources. The first method is batch synchronization, which initially involves loading the SaaS application with data. Once this initial data load has been made, data can be incrementally synchronized on a scheduled basis. The second method is real-time integration using Web services. Another way to combine the two methods is by having a Web service trigger that is based on an event occurring in the SaaS service. Yet another method is emerging that involves integrating SaaS applications at the user-interface level through mashups.
Assumption 5: SaaS is only for simple, basic requirements.
Fact Check: False, but there are still limits. SaaS applications are highly configurable at the metadata level with many offering customization capabilities with platforms in the form of application platform as a service (APaas). There are industry examples in which complete custom applications have been built using SaaS APaas. However, some gaps remain for complex, end-to-end processes that require complex workflow or business process management capabilities.