One of the perils of working in the thought leadership world is that predictions can subsequently be evaluated against reality, perhaps unfavorably for the pundits. Three years ago, shortly after starting our BI services company, we put on our sage hats and looked out into the future of open source business intelligence (OSBI). In this column, we evaluate our soothsaying, while updating our hypotheses and predictions for OSBI going forward. Our “wisdom” is driven from work with the Jaspersoft and Pentaho BI platforms, the Postgres and MySQL relational databases, the Infobright and Vertica analytical databases, and the R platform for statistical computing.
The previous column’s point of departure was the 1995 Harvard Business Review article by Joseph Bower and Clayton Christensen, Disruptive Technologies: Catching the Wave, in which the authors distinguish disruptive from sustaining technologies. We conjectured OSBI as a disruptive technology, addressing markets which sustaining vendors did not find financially attractive. In turn, sustaining vendors would focus on existing customer needs – often to the detriment of their long term growth.
From the HBR article and our experience with BI and open source, we opined a number of hypotheses for OSBI, the most important of which are summarized here.
- OSBI adoption would be driven by the unmet needs of an underserved market.
- The ISV/OEM segment would comprise much of COSBI's early adoption.
- Commercial open source BI (COSBI) would benefit from successful open source technology lines such as operating systems, relational databases and web servers that preceded in the market.
- OSBI would fuel technology innovation; COSBI would benefit from this progress.
- The COSBI business model would be disruptive in its introduction of new ways to serve their markets and price their products. This disruption would, in turn, accelerate adoption.
- Functionality of COSBI and mainstream BI products would converge.
- Many traditional vendors would either accommodate OSBI or make the transition from license/maintenance to subscription/support.
- COSBI and Microsoft would square off in the SMB market.
We pretty much hit the target on 1) & 2), the unmet needs of underserved markets and prominence of OEM/ISV's. Many of the customers we've engaged over the last three years have been from previously neglected segments -- OEM/ISVs, SMB and start-ups -- an experience that's been affirmed by vendor partners. The size of adopting companies continues to increase, however, testimony to COSBI's growing acceptance in the marketplace. We see large company adoption especially for projects within new business/data domains, eCommerce/clickstream analytics and health care informatics being prime illustrations. For small and large prospects alike, successful experience with open source technology is so important that 3) is now an article of faith in the COSBI sales cycle. Indeed, the maturation and adoption of open source ETL has exceeded OS reporting, OLAP, and dashboarding to date, no doubt because of its infrastructure kinship.
Given the steep trajectory of COSBI product improvements over the past 3 years, 4) confirms our prediction of rapid technology innovation as well. And while COSBI vendors often employ or contract with the most important product developers, they continue to benefit from the goodwill of the open source community. Examples include the benefits to both the Pentaho and Talend ETL products of community-led contributions to transformation libraries , and the Pentaho and Jaspersoft largesse of community dashboarding and reporting frameworks. Finally, analytic database Infobright is built on the MySQL platform as a storage engine, leveraging MySQL’s extensible architecture to expedite its development.
The rapid expansion of COSBI product capabilities challenges established BI vendors, and though most proprietary products still lead in functionality, COSBI has closed the gap, 6), especially with the 80% of capabilities that matter most to customers. The cost/value tradeoffs are especially important to prudent buyers in the difficult current economic environment. We've also seen attempts by mainstream BI vendors to accommodate OSBI adoption, 7). Struggling database vendor Ingres transitioned from proprietary to open source. Oracle's acquisition of Sun, which in turn had acquired COSBI MySQL, is another case in point, as is MicroStrategy's decision to make a 100 user unsupported version of its platform available for free.
We haven't affirmed 8), the competition between COSBI and Microsoft at the smaller end of the BI market. For the most part, we've found that companies with a successful Microsoft history go with Microsoft for BI, while for those inclined towards open source (especially Java platform) solutions, Microsoft is not even a consideration. Open source vs Microsoft could well be in the works in other venues in coming years, however, as Oracle and Google enter the fray against Microsoft in its core product lines. Stay tuned.
The COSBI business model, while significantly different from traditional commercial licensing, has been more a dachshund than greyhound in sales performance, casting doubt on our disruptive prediction in 5). Most COSBI vendors have steadily grown revenue – in some cases approaching profitability -- but certainly not at a disruptive or chasm-crossing pace. The venture funders, while not despondent, aren't euphoric either, their aggressive, back-of-the-napkin adoption calculations of a few years ago too optimistic. The software subscription strategy that's been the foundation for growth simply hasn't trajected as steeply as planned.
Perhaps the COSBI community must rethink their business strategies as they progress to the next level of growth. And who better to help them examine alternatives than Wired editor Chris Anderson, author of New York Times bestseller, The Long Tail, and the just-released Free, The Future of a Radical Price. Anderson makes persuasive arguments for the acceleration of “free” business models driven by an online economy whose costs are fast approaching zero. Along with the challenges of the new imperative are opportunities for companies who can accommodate their business models to generate profit from free.