Co-authored by Steve Miller and Dave Reinke
In a 1995 Harvard Business Review article, Joseph Bower and Clayton Christensen presented an analysis of disruptive technologies that still resonates more than a decade after its publication.1 The relevance of their work is especially remarkable given the recent successes of open source (OS) technologies. This article attempts to update Bower and Christensen's arguments to include the impact of the commercial open source (COS) business model, wherein for-profit companies support, enhance, integrate and indemnify OS software using various licensing options. Our focus is on business intelligence (COSBI), drawing on experiences with the two major platform vendors, Pentaho and JasperSoft, who have demonstrated early success in the marketplace.
For Bower and Christensen, the differences between "disruptive" and "sustaining" technologies revolve on two major axes: 1) disruptive technologies address a market that is not very attractive to the mainstream or "sustaining" vendors, often because of unappealing growth or financial characteristics; 2) sustaining technologies concentrate on existing customer needs, which may ultimately stifle innovation. In addition, if both sustaining and disruptive technology firms are enhancing product performance at trajectories ahead of market demand, disruptors with early success have a better opportunity to invade established markets.
Based on Bower and Christensen's conceptual foundation, and our own experiences with OS and BI, we believe the following dimensions are catalysts for the emergence of a disruptive business model such as COSBI:
- Unmet marketplace needs,
- An underserved market with willing early adopters,
- A compelling business model, and
- An innovation engine.
With unmet market needs, one or more underserved user communities and a more cost-effective approach to foster rapid technology innovation, chances are very high for a disruptive success. What follows considers each catalyst with respect to COSBI, relating experiences with the major COSBI vendors.
Unmet needs. Organizations have invested in BI for decades. Most BI implementations are now built on data warehouses and reporting databases that at a minimum support managed query and reporting, along with slice-and-dice analysis using online analytical processing (OLAP) technologies. The more sophisticated publish dashboards to measure business performance and provide an advanced analytics capability for statistical modeling. Unfortunately for today's increasingly demanding users, most of these deployments are both siloed from business operations and passive - requiring users to switch context from operations to analysis and back to operations to analyze data, make decisions and act. As a consequence, many organizations are now searching for quantum improvements in the efficiency and intelligence of their day-to-day processes, spawning new requirements to make the connection from analysis to action transparent - to seamlessly "operationalize" BI.
Operational BI has many definitions, but most include real-time reporting on operational data, transparency between transactional and analytical end-user contexts and automated decision-making driven from analytics. New requirements distilled from operational BI, however, pose problems for mainstream BI software vendors who followed the same enterprise application approach as enterprise resource planning (ERP) and customer relationship management (CRM) vendors and architected holistic, self-contained BI tools and applications. This black-box approach does not fit the new, extended architectural mantra of operational BI: componentize and embed.
COSBI vendors have seized on the opportunities surrounding operational BI. They espouse a fresh design perspective, shifting from enterprise applications to a service-oriented architecture (SOA) BI platform that delivers embeddable, intelligent components. Leveraging open technologies such as XML and J2EE, COSBI directly addresses the new needs of operational BI, enabling reports, OLAP views, analytical models and other intelligent components to be embedded in almost any standards-based application. In addition, these technologies change the way BI components are developed. Deeply rooted in SOA, the COSBI platforms introduce advanced modularity and a process orientation, enabling the creation and embedding of advanced composite BI components.
Underserved markets and early adopters. Historically, significant communities have been shut out of BI by excessive product costs and lack of accessibility. With the traditional BI architecture requiring database, extract, transform and load (ETL), reporting and OLAP software, the initial licensing fees from sustaining vendors can easily extend into six figures for a basic environment. For many businesses, this is simply a nonstarter. Instead, they have been relegated to the limited BI capabilities of office productivity software or the inflexible reporting modules of their core transactional applications. Now, though, these organizations can quickly download freely available OSBI report development, ETL and OLAP tools. The access to product is customer-driven, instantaneous and global. When the companies are ready for production, they can contract with the vendors for support subscriptions.
As newer OS technology, COSBI benefits from those OS siblings who preceded and succeeded. The adoption of established products, such as Linux, Apache, JBoss, Perl, Python, Ruby, MySQL and PostgreSQL, ensures an audience for COSBI in many IT departments. Having a senior-level OS champion in IT is considered a coup by vendors. Indeed, "has adopted other open source platforms" is often a preliminary deal qualifier for COSBI vendors. Fortunately for COSBI, OS technologies of many types have preceded into companies both large and small.
The ISV/OEM developer segment has made significant investments with COSBI for the architectural flexibilities that encourage embedded applications with reduced royalty payouts. The small to midsized business (SMB) market, long a BI nonplayer, is making enthusiastic overtures to COSBI. Emerging geographies, deprived by both price and access, are also establishing a presence with COSBI vendors. Finally, many mature, Fortune 2000-scale BI veterans are finding that projects which extend BI to their customers and partners are now feasible given favorable licensing provisions and the ease of embedding COSBI into existing extranet portals.