The Pros and Cons of Open Source Business Intelligence

Though its value-adding capabilities are widely recognized, business intelligence (BI) is still considered a high-cost endeavor. BI projects are indeed costly, and important investments in expensive software licenses are usually required up front. In this context, the promise of open source BI (OS BI) as a low-cost alternative to commercial BI is worth considering. OS BI products are not entirely free, but there is a strong sense of “freeness” associated with them. Unfortunately, this notion of freeness can bring about negative perceptions that should be rectified, such as that OS BI may not meet quality expectations, or that it is poorly tested, or that it suffers from a lack of support, and so on.

Open source BI products have clear advantages over commercial offerings. The most important advantage is that they are available to download and evaluate for free, without having to deal with a software provider. They can also be used as a quick and low-cost solution for the implementation of a proof of concept or prototype, and they allow corporations to lower the complexity of license cost management. Several other aspects are, of course, worth mentioning, such as the extensive online support community involving tens of thousands of active forum users, not to mention the proliferation of white papers and third-party books on the subject. OS BI tools are usually more accommodating to custom environments and offer more integration possibilities and flexibility than commercial products. Moreover, a corporation can integrate OS BI to its own infrastructure without having to acquire rights from third parties. OS BI is also especially geared towards agile environments where it is useful for tools to fit on standard laptops from which they can be scaled up to full-blown enterprise platforms. 


Image via Shutterstock

However, as with any issue, the other side of the story should be considered. Open source solutions also have disadvantages when compared to their commercial counterparts. The most notable one is the frequent lack of centralized and integrated metadata management frameworks. OS BI products often come with metadata maintained in heterogeneous XML structures, which complicates integration to enterprise metadata management frameworks. Also, OS BI products are more difficult to use than commercial ones, which makes for a poorer user experience. The positive side of open source code, whereby code can be modified locally to address the particular needs of a corporation, also has its drawbacks such as when the time comes to upgrade the product to an official version. This requires extensive technical expertise not always readily available in-house – something that might incur significant costs in the long run.

An important factor for any enterprise is the question of legal liability, which is nonexistent or minimal on the part of open source tool development and support scores. In any case, it is usually lower than with commercial products. This is something to be careful about when going forward with a solution implemented using these packages. A last point is that, to date, OS BI tools have been mostly used in the context of small to mid-sized firms; hence very few large corporations have adopted them as an enterprise strategy. It is therefore still very hard to assess the viability of OS BI in this context due to its low adoption rate.

To conclude, several factors should be considered before adopting an OS BI product. First, if the goal is to acquire a comprehensive BI suite, the one with the best modules to address the specific needs of the corporation should be selected, not the one that has the best overall suite. Since most corporations only exploit a subset of OS BI functionality, only those modules or functions that will bring the most value should be evaluated in depth. Second, it is also important to consider the formal selection process. If the executives consider it important to have a commercial entity standing behind an acquired product, enterprise editions should top the list. The third point is to consider the corporate infrastructure: there is a cost implied in integrating new tools to a mature environment, therefore legacy applications should always be considered. On a final note, a good assessment of business requirements should be done before considering OS BI as well as a serious review of the companies behind the OS BI solutions under consideration.

For more details on evaluating Open Source (News - Alert) BI, read the InEdge white paper A Review of Open Source Business Intelligence.

Jean-Stephane Faubert is a Senior Solutions Architect at InEdge (News - Alert). Jean-Stephane has more than 20 years of experience designing and implementing information technology solutions. For more than half of his career, his focus has been in the design and implementation of Analytics solutions. Over the years, Jean-Stephane has designed Analytics solutions for clients in several industries, such as Insurance, Banking, Brokerage, Health, and others. His expertise covers the whole spectrum of Analytics project-specific tasks ranging from strategic planning, solutions architecture, project management, and assessment, all the way down to data modeling and detailed design. Being pragmatic, he focuses on delivering added value to InEdge’s clients to answer their specific business challenges.