It’s clear that the Business Intelligence community is trying to better understand SaaS BI. Case in point, we held a webinar with Wayne Eckerson (When SaaS BI and On-premise Worlds Collide, Nov 19), and another with Claudia Imhoff (SaaS BI – The New Framework for End-user & IT Success, Oct 29), both of which showed heavy attendance. Wayne and Claudia started their respective webinars by defining SaaS BI to clear the confusion around terminology, applications, strengths and weaknesses. As if to reinforce this, there were many questions from the audience during Q&A – many more than you would typically have for a webinar.
The analysts’ approach reflects the mind-set of the business intelligence community. Yet, Claudia and Wayne were each followed by a case study presentation where the VP of Data Management at a distribution company (DMA) and the CIO of a consumer products company (Shaklee) discussed how effectively they have been using SaaS BI for several years. Their passion is undeniable. It was this very dichotomy that set off the avalanche of questions.
The Business Intelligence community does not mess around. Their questions went to the heart of the matter in a hurry. Does it scale? Is it secure? How is data integrated? And so on. I will touch on the first three topics. In future posts, I will dive a little deeper into each topic and work through the 20+ other questions that were asked.
Does SaaS BI Scale?
SaaS BI can scale across 3 dimensions: Users, Data and Sources. Deployments with thousands of users all with ad-hoc reporting capabilities are becoming more common. Deployments are often multi-source spanning hundreds of tables, complex joins and relationships in individual customer deployments. With SaaS BI, users have the ability to analyze billions of rows of data with wide dimensionality. There are different flavors of SaaS BI that may not cater to the above scalability requirements, and instead focus on single data sources that can be instantly uploaded, etc. We will address the SaaS BI landscape in another post.
Is it Secure?
SaaS BI security should be either at par or better than your security. Security needs to be beyond the data center and cover the entire service and data handling processes. SAS-70 Type II certification is a good benchmark when evaluating SaaS vendors (not just SaaS BI). This meets the needs of public companies who analyze financial data and companies with data covered by HIPAA. There is a SAS-70 Type I certification, that may or may not be robust enough for you. More on that later.
How does Data Integration work?
It is now more common for SaaS BI providers to provide you with access to application connectors for the likes of SAP, ORCL, Salesforce.com, Netsuite, databases and flat files. There can be 50+ of these ready to go. The extraction process is typically automated to send data to the SaaS provider over secure protocols where the data is joined and combined for complex analytical calculations. Data refresh once or several times a day is reasonable. The manner in which the data is stored and processed will differ by vendor. Columnar storage, in-memory processing vs. fixed schema’s is something we will cover later in more detail.
Please check back in the coming days as we work to respond to the many questions we have received.
