Hybrid Business Intelligence (BI) often involves coordinating the advantages of a centralized approach with that of a decentralized or distributed one. Although each of these methods has distinct benefits, a synthesis mitigates their respective drawbacks and forms a gestalt of their benefits.
Many organizations have realized there are a number of specific concerns that must be addressed when integrating these two forms of BI. In addition to aligning the expedience and agility that is essential for business users with the uniformity and standards that is required of IT and encouraged by upper-level management, companies must balance distributing BI hegemony with sensitive issues of communication and organizational politics. Additionally, the prominence of emerging technologies such as Mobile BI and Cloud BI reinforces the necessity of a hybridized approach.
The management of Hybrid BI presents a new set of considerations that are integral to the efficient use of actionable data. Conventional restrictions on standards and the decreased involvement of IT are part of a larger evolution in which multiple BI sources deliver alternate perspectives to inform decision-making via readily contrastable viewpoints. Subsequently, properly implemented Hybrid BI yields more information than ever before.
Distributed Meets Centralized
The deployment of multiple BI tools requires either a centralized or a distributed, silo-based approach – or a combination of the two. According to Forrester’s Boris Evelson, multiple tools are one of the drivers towards Hybrid BI solutions. The analyst said:
“…organizations now realize that consolidation and centralization are no panacea. Currently, Forrester client inquiries about how to live with multiple BI tools far exceed inquiries about platform consolidation. Why? There is simply no one vendor that provides all of the key agile BI capabilities. And this is unlikely to happen in the near future, so the need for multiple BI tools will be a pragmatic reality.”
The benefits of Agile BI also function as a key driver to facilitate hybrid environments. Iterations frequently require a specific, specialized approach based on user feedback that assists users according to business unit. Agile principles work well with a distributed BI deployment, which is highly user-oriented and provides expedient results at a pace viable for business professionals. Other boons of the distributed approach include an attractive economy of scale and flexibility that increases response time and the incorporation of data into decision-making.
But, without a centralized supporting structure, it becomes far too easy for decentralized BI data to become too user specific and not adhere to standards that are of use to the rest of the enterprise. A centralized approach is needed to provide uniform definitions and metrics, and to avoid the redundancy inherent in distributions. Effective data federation facilitates long-term operations and business strategy, which is a legitimate data use that is difficult to accomplish via strictly distributed approaches.
Yet, strictly centralized approaches may be even less effective than distributed ones. These methodologies frequently generate a backlog of IT requests to issue data to business users – who sit idly by while opportunities for action bypass them. Hybrid BI reduces these drawbacks and augments these strengths by utilizing a centralized architecture to regulate the terms and formatting of distributed data, which works with any BI tool. Such architecture provides end-users the swift implementation of distributed methodologies – without backlogging IT – while facilitating quick decision-making and Agile principles. According to a report from Gartner:
“Many BI programs have departmental roots with analytical resources embedded in the business. This model… lacks consistency in terms of data definitions and measures across an entire organization… such an overly centralized model lacks the agility and familiarity of the decentralized model. A hybrid delivery model enables greater consistency and economies of scale, more autonomy and faster turnaround times.”
Hybrid BI Implementation
The advance of Cloud-computing, mobile devices, and SOA has added a degree of complexity to the traditional centralized versus decentralized BI argument. Subsequently, enterprises have many more implementation options for achieving a balance between these two respective methodologies. One of the most commonly used approaches towards implementation is the 80/20 model, in which approximately 20 percent of data gained from distributed BI in respective departments is funneled into a central repository. This data is typically mission-critical and highly relevant to particular business processes. 80 percent of the data, however, is left to individual iterations of the departments that procured it. Ideally, this data conforms to enterprise wide architecture and best practices so that it can offer insight to others outside of specific business units
Organizations are only beginning to realize that despite the inherent value in the presence of a strong central architecture and standards that include dimensional schema, BI tool selection and more, it has become increasingly apparent that the agility and practicality of departmental or transactional data actually transcends standards in many cases. Universal standards should encompass every aspect of BI solutions, including not only definitions and naming conventions, but also the actual tools themselves. Expedience and business value derived from tools supersedes mere conformity. According to Evelson, “The best tool for each BI job trumps IT standards. BI has traditionally been ruled by over insistence on enterprise wide standards and a single version of the truth. These will continue to be important, but they won’t be the Holy Grail.”
This revelation doesn’t mean universal standards will be abandoned (especially since this is one of the benefits of Hybrid BI), but as paradoxical as it sounds, they must become more flexible. Certain features and tools work best for particular deployments and should be used. It is up to the individual organization to determine which standards should be mutable and which must be adhered to – the latter will generally pertain to clarity, transparency, and Data Governance.
The Final Frontier: Big Data and The Cloud
The advent of Big Data is another crucial driver affecting the prudent implementation of Hybrid BI. By virtue of its non-proprietorial, unstructured format, Big Data was originally accessed via business units in a silo format and was sequestered from other data within the corporate firewall. However, increased interest in the importance and potential of Big Data for achieving both business and mission critical objectives has resulted in a slew of new technologies that incorporate such data with conventional proprietary data. Some vendors, such as Teradata, Autonomy, and IBM (with Watson), are providing Big Data analytics and intelligence on hybrid data sources with self-service BI tools. Gartner’s Rita Sallam stated that “through 2015, organizations integrating high-value, diverse, new information types and sources into a coherent information management infrastructure will outperform their industry peers financially by more than 20%.”
While Big Data’s impact on hybrid sources and the intelligence required to glean insight from it are fairly established, Cloud BI offers an ideal form of hybridization that has yet to be realized. Forrester has indicated that user adoption of Cloud BI remains relatively low, especially for those that have made substantial investments in traditional enterprise-based BI. That fact has not stopped BI vendors from issuing a number of Cloud BI solutions in which data is transferred to the Cloud for reporting, dashboards, and analytics. Organizations that utilize some other aspect of the Cloud – such as data virtualization – are more likely to employ Cloud-based BI. Cloud BI presents a form of hybridization that combines offsite and on-premise intelligence. Positives associated with this form of BI include modest investment costs, flexibility of availability, and an ease on enterprise-based IT resources.
Wherein Lies the Truth
Hybrid BI has expanded beyond the conventional nexus of a distributed and centralized approach to encompass various aspects of mobile, Cloud, and Big Data technologies. Its true value lies in finding a delicate balance between a unified architecture typified by governance and standards consistency, and the celeritous implementation and iterations of business units utilizing tools for their own ends. Many of these concerns can be mitigated by vigilant architectural reviews, uniform metadata implementation, and diligent communication between departments. The 80/20 model provides a happy medium in which both agility and mission critical data storage is facilitated, enabling data to be shared throughout the entire enterprise. Whereas slower, centralized approaches provide a uniformed version of data on which to base decisions, Hybrid BI allows for the incorporation of more varieties and sources of data. It provides various shades of truth that may be relevant to some departments and processes more than others, while still yielding comprehensive, mission-critical information that assists all.