A week or so ago Gartner released the Magic Quadrant for Business Intelligence Platforms and commented on SaaS BI. Below is a relevant excerpt that needs some expounding:

“In the economic downturn, interest in SaaS solutions has increased in the past year, although it is still a small fraction of the overall market.” It continues “…. Moving BI off-premises may not suit all organizations and all use cases, especially those dealing with highly sensitive data. Many firms are evaluating hybrid options for deployments leveraging both private and public clouds, as well as a combination of on-premises and off-premises solutions. But firms that find the SaaS value proposition of more rapid, lower-cost deployments attractive should evaluate SaaS as an option.”

Let us look at a few of Gartner’s points that deserve a close look:

1. Moving BI off-premises may not suit all organizations and all use cases, especially those dealing with highly sensitive data

The above is true but it has less to do with security than with functional use cases. A SAS-70 Type II certified service – very different from and much more than a SAS 70 certified data center is widely accepted as a standard for sensitive data. PivotLink customers that are public companies and companies that are covered by HIPAA have guided PivotLink into achieving SAS 70 Type II certification. They feel comfortable using PivotLink with highly sensitive data because the entire service – data center, procedures and personnel are covered by an external audit.

2. Impact of  the economic downturn

There was an economic downturn during 2001-2003 and there have been many before.On-premise deployments is how IT has been delivered since the early 80s. During all the previous downturns, CEOs were supported by each department on scaling down – shutting down stores, cutting product lines, expenses. IT however was a problem child. This was for two reasons: One, data always grows, be it good economy or bad, so IT cannot really cut infrastructure. Two, during all the previous downturns CIOs did not have SaaS options. Due to a significant CAPEX in IT, IT can severely limit a CEOs option to free up cash during downturns. CIOs did not have many options in all previous downturns. There is only so much that can be outsourced.

Today, CIOs have plenty of options in almost every category of enterprise application software from Email to ERP.

With frequent ups and downs in business cycles, more CIOs will choose SaaS to support their CEOs to be more agile. CIOs know that the technology is there – Cisco has carpeted the earth with Mbps networks and will soon move to Gbps. Encryption technology, RSA Tokens, SAS 70 Audits all can make security adequately tight. SaaS companies now routinely provide over 99% availability. SaaS is a fundamental shift in how information technology will be delivered in this decade.

3. Firms that find rapid, lower cost deployment attractive should evaluate SaaS as an option.

Are there firms that don’t want rapid and lower cost deployments? What Gartner seems to be implying is that there are risks in going with SaaS.

These risks can easily cause anxiety in organizations evaluating SaaS BI, if not put in context. SaaS is not a new model. Every category of application source is moving to the cloud. CRM (Sales, Service and Marketing) has already moved, except in a few verticals. ERP, Finance, Expense Management, Talent Management. SaaS is designed to address the pain points of slow and high cost deployments. As many data sources move to the cloud a tipping point will reach when cloud BI will be the preferred option. SaaS is not new – Online Banking, Trading, e-Commerce are examples of SaaS. We all use them and trust the way they work.

The risks that are unique to SaaS are functionality fit and vendor viability. These risks can be mitigated via understanding the outcomes of the technology purchase and by evaluating the integration capabilities and financial health of the vendor.

Where does SaaS BI fit and where it doesn’t?

• Requirements around Large Data Volumes

If you have a data volume in the few hundred terabytes or petabytes range, there may not be a SaaS vendor today to handle such volume, if the expectation is to get ad-hoc reporting. For example, moving the enterprise data warehouses at companies like Safeway, Sears etc.  to SaaS is not feasible. However it might be feasible to give a sub-set of data to a SaaS vendor so that merchandisers in a specific division can do ad-hoc reporting.

• Requirements around Frequency of Data Updates

If your daily updates are in terabytes of data, the network and refresh latency may not meet your data availability needs. For example, if you are a company that is even 1/3rd the scale of an Amazon.com and want to collect click through data and want to share the results with suppliers multiple times daily, then although SaaS is a fit from a sharing model, it may not be a fit when it comes to transferring terabytes, multiple times a day to SaaS vendor.

• Functionality fit

If your company needs an MDM solution, a DW solution and a Data Mining solution all integrated into one, then the SaaS choices are limited to none. There are SaaS choices emerging in individual areas like MDM and Data Mining and DW but not under the same roof or integrated enough. Companies should think of BI not in terms of the suite definition imposed by self-serving large vendors like Oracle and SAP but in terms of the problem at hand the cost / benefit of the SaaS solution.

In summary, although security should be evaluated, it is no longer an unsolved problem. Organizations today must look at business agility, speed and empowering the business users for self-service when looking at SaaS vs. On-premise choices.