SaaS: myths and reality

Growth of popularity of the software provided according to the SaaS model caused a number of the assumptions of new model, however for many companies it is quite difficult to separate the truth from fiction. Analysts of Gartner company [1] (file) carried out the analysis of five most frequent assumptions of the SaaS model and defined their justice. According to Robert Desisto (Robert DeSisto), the vice president and the analyst of Gartner company, growth of advertizing of the solutions provided according to the SaaS model is observed in recent years. As a result there was a large number positive and negative, responses and also the false assumptions around this technology. Thereof some companies made the decision on deployment of solution SaaS, based on these false premises.

Assumption 1. Solution SaaS are less expensive, than the "traditional" software (on-premises).

Result of check. These estimates are fair within the first two years when the SaaS-application really has rather low total cost of ownership (TCO) as they do not demand large investments on purchase of licenses and support of infrastructure. However, since third year and further traditional software, set on own servers of the client, can be less expensive, considering perspectives of further use.

Assumption 2. The software according to the SaaS model is quicker unrolled, than "traditional" software.

Result of check. It is right for a simple SaaS-applications, however the growing complexity of the SaaS-applications disproves this statement. Suppliers often promise to implement the SaaS-application in 30 days, however actually deployment according to the SaaS model can take seven months and more. As the complexity of business processes and their integration increases, the gap between solution SaaS and "traditional" software will be insignificant because the most part of time of implementation will be taken away under setup and integration, and these processes equally labor-consuming for both models.

Assumption 3. SaaS is estimated as service.

Result of check. Is not true in most cases. Many suppliers consider the SaaS model close to model of utility payments (for the electric power, water, etc.) as the companies pay only for what they use. However in most cases the companies should sign in advance prepared contract irrespective of the actual use of a system. In certain cases application of SaaS can be estimated proceeding from real load of a system (for example, in electronic trading - depending on number of tranaktion), however such examples minority.

Assumption 4. SaaS is not integrated with applications of on-premises and/or data sources.

Result of check. It's not true. There are two main methods of integration of SaaS into applications of on-premises and/or data sources. The first method is based on synchronization, i.e. on periodic loading of the changed data in the SaaS-applications. The second method consists in integration in real time using web services. These two methods can be combined, for example when using the trigger which at approach of a certain event will start procedures of synchronization. The same can be performed also on the instructions of the user.

Assumption 5. The SaaS model is intended only for automation of simple, basic requirements of the companies.

Result of check. It not absolutely so. A SaaS-application is configured at the level of metadata, has ample opportunities, assumes setup of the platform as service. There are examples of the full industry applications created according to the SaaS model. However there are difficult tasks, the management of which requires complete solutions.