The proponents of SaaS BI claim that it is a lot less expensive to use cloud-based BI products vs. traditional on-premise BI.  But is it?

The biggest advantage to SaaS BI is that it does not require IT to install and support BI or database software, which can save money, but is SaaS BI really less expensive than on-premise BI if everything is taken into consideration?

It depends on a several factors:

First, does the SaaS BI vendor offer a complete suite of BI tools and technology?  Or do they only offer dashboards or programmatic report generation?  If you depend on a SaaS BI vendor with limited functionality you might end up adding third party BI tools for ad-hoc queries, predictive analytics, OLAP, or data mining.

Second, does the SaaS BI vendor lock you into their cloud?  Do you have the flexibility to bring the BI applications in-house if the need arises, or would you be forced into a costly conversion?

Third, what is the staying power of the SaaS BI vendor?  We notice that despite the backing of deep-pocket venture capital, two SaaS BI vendors have folded.  Is your SaaS BI vendor backed by technology that is used by millions of users?

Fourth, does the SaaS BI vendor offer a simple, yet powerful data-loading tool (ETL) that can handle your requirements?  Does it provide a full suite of data transformations?  Is it equipped to deal with Master Data Management?  Can you use your existing data integration technology if you choose?  Can you access the data with standard SQL tools for database design and debugging?

Fifth, does the SaaS BI vendor give you the access and tools to manage the servers?  Or are you at the mercy of their support and help desk?