What Delivers the Most Value for Enterprises New to BI -- Open Source or SaaS?

The question begs asking more questions. Does the enterprise get more economies of scale from the public cloud or from their own data center? Is the BI for a core business function or for non-core? Is the buyer looking for an analytical application or a platform to develop a solution?

Answering these questions can bring more focus into the leverage a company can have with open source vs. SaaS. Clearly, SMBs that lack development personnel and data center capabilities will look at SaaS as a great option. A large enterprise looking to build analytics to support a core business function might look to open source to get leverage. So the answer would seem to be that it depends.

I agree with Matt. It really comes down to the features of the application, since both approaches are about the same when considering costs. For instance, Sugar vs. Salesforce.com.

That draws more light on the Make or Buy question. Clearly, Open Source is increasingly relevant when developing solutions, but it assumes competent development teams. My recent experience with large enterprises points away from Make - most prefer to Buy if a solution exists, and would Make only if there's no alternative. The smaller the enterprise, the stronger the Buy preference. Relating back to the question here, this would weigh heavily in favor of SaaS.

In discussing issues like this, there are so many facets that need to be analyzed. Having developed a large-scale EII tool, I have learned that moving large volumes of data around in order to create an aggregate data set off which to provide analytics is still a very real hurdle, which ends up being a real limitation of the SaaS-based environments. Clearly, if moving to the cloud for analytics, MapReduce-based systems are on path to deliver value sooner than later, but it means that you will have integration between the Cloud and your internal environment.

Open source projects sometimes can be very powerful, but can lack the polish and out-of-box experience expected of many BI tools users. Remember, many of these users are not engineers who set up their own Linux boxes every day just for fun.

My vote is that in the short term, licensed BI will continue to thrive and as we get closer to large-scale analytics combined with Cloud functionality, SaaS will have a breakthrough opportunity.

I also agree with Matt, and see both open source and SaaS BI gaining momentum. Depending on the current business and IT infrastructure within an organization there will continue to be the need for both, as both structures can offer organizations value.

However, using open source solutions to make BI vs. buying a BI solution would mean that the perceived value of developing a solution in-house outweighs what already exists in the market. Although organizations want to cut costs overall, the trend that I see seems to be moving towards both mid-sized and enterprise organizations looking at SaaS based BI to provide their lower cost, easier to maintain applications.

The key to a successful BI deployment in an enterprise is to start BI on a handful of business processes and gradually expand the BI footprint while simultaneously building confidence in user's minds. Open Source with its minimal footprint and zero cost is ideal for enterprises that are new to BI as they can have a working system fast (with minimal features) and quickly demonstrate value to the users.

SaaS on the other hand, has an edge over Open Source by having feature rich tools as a part of its offering and still keep the total cost of ownership low. However migrating data to a cloud for SaaS, is a significant challenge requiring sign offs from different stakeholders in the enterprise.