Five Ways to Improve the Success of Business Intelligence Projects

Image courtesy of Shutterstock

Numerous studies over the past few years characterize generally poor performance in the success rate of business intelligence projects. It appears there is a general acceptance that these failures are somehow an acceptable “cost of doing business.”

And yet, according to market research, the business intelligence market will reach over $17 billion in 2017. How can we reconcile this growth and market acceleration with such a high probability of failure?

The lack of robust business intelligence alternatives has driven a force-fit mentality where highly diverse projects follow a design pattern that hasn’t changed much since the 1980s; it is suitable for only distinct types of demand. The additional overhead drives unnecessary costs, time, risks and ultimately, poor business performance.

Now is the time to move business intelligence practices into the 21st century, or risk a continued cycle of overspending and underperformance.

Approaches that have plagued BI projects include:

  1. “Go Big or Go Home” – traditional business intelligence evolved from the belief that requirements were only addressable by first centralizing massive amounts of available data. This ensured a large margin of error to help avoid costly changes yet requires major upfront investments, which decimate ROI before even a penny of value is created.
  2. “Lost in the Translation” – with the emergence of increasingly specialized tools (and staff) to manage massive centralization efforts, the distance between business user and technology implementer grew. This widening gap meant added layers of translation and interpretation risk between logical need and physical implementation. As a result, far too many projects have ended up being technical marvels that simply don’t meet the business need.
  3. “It’s been a zero sum game” – major efforts have been made to perfect the operational processes and lower the execution risk of mass centralization strategies. Unfortunately, the incurred overhead to sustain execution excellence comes at the expense of poor agility. New business participants, an explosion in data, and new uses for business insight require rapid adaptation.

Fortunately, there is hope for firms faced with the complexity and opportunity of modern business. The next generation of business intelligence leverages a broader spectrum of strategies, empowered by emerging technologies to deliver better outcomes across the diverse set of business needs.

These five new strategies will help improve alignment between unique problem characteristics and specific strengths:

  1. Segmentation – it is critical to understand the diverse characteristics of emerging business problems to select the right response. Considerations like source system distribution and diversity, data volumes and variety, time to value and required agility must be addressed directly to drive success. Enterprises must build critical capabilities and apply them selectively to properly address specific segments of demand.
  2. Collaboration – modern enterprises must instill a culture of deep and ongoing collaboration between business users seeking value and technologists deploying solutions. These collaborations must apply established capabilities today, yet also plan for requirements only on the horizon.
  3. De-Layering – it is critical to maintain a high degree of transparency between underlying source systems and solution components. This alignment between logical design and physical deployment can dramatically reduce translational risk incurred when crossing multiple boundaries. For example, if you are using internal data from a customer survey and external data from a public data source like Wolfram Alpha, it is crucial that your system and users understand which source the information is coming from to avoid mistakes.
  4. Ability to Rapidly Prototype – gone (or nearly so) are the days when millions of dollars will be committed upfront for the mere promise of value. Business leaders want specifics on value potential, investment requirements and operational costs before committing funding. Development organizations must create the ability to rapidly prototype scalable analytical applications that demonstrate end-to-end value and also accurately project the total cost of ownership.
  5. Agility – development organizations must also demonstrate the ability to rapidly adapt to changes in all aspects of a solution – from data acquisition, processing, and delivery to user base, partnerships and requirements. Important concepts like non-invasive integration, late-binding, and dynamic scaling must be available to key segments of the portfolio to provide dramatic reduction in cycle time from change to value realization.

Key segments of the enterprise business intelligence portfolio are now being delivered in less than half the time and cost of traditional approaches. In particular, new platforms that support a highly distributed, non-invasive, and agile approach are being deployed as an alternative for rapidly growing segments of demand. These new platforms have been architected to target, extract, and process diverse solution elements and successfully deliver against challenging requirements. Importantly, these platforms not only co-exist with but leverage “legacy” investments to maintain maximum solution flexibility – ultimately bringing those solutions stuck in the eighties into today.

By embracing the 5 steps outlined above, business users, data practitioners and development organizations are together characterizing and addressing demand with the collective strength of multiple strategies. Most importantly, they are changing up what has been the norm in the business intelligence world for the past 30 years. Enterprise organizations would be well served to evaluate these alternatives and assess suitability for their demand profile.