IT executives know that empowering employees with information and insights can streamline business processes, reduce costs and improve decision making and planning. The only problem is that most executives find it extremely challenging to transform business intelligence (BI) from a departmental tool to an enterprise-wide resource.
The irony is that most organizations have more BI tools than they need. On average, organizations have almost 13 tools spanning five major categories of BI. (See Table 1.) This proliferation has CXOs seeing red; they can't understand why they can't buy one BI tool for everybody, which would reduce overhead, save money and streamline suppliers.
|Type of tool||Number of distinct tools, built or bought|
|Online analytical processing||2.8|
Most organizations are eager to clean up this mess and standardize on one enterprise BI tool per category, ideally sourced from one vendor. Unfortunately, few organizations have achieved this goal for a variety of reasons:
- Politics: Many organizations have pursued decentralization strategies, giving departments and divisions greater autonomy to deploy their own BI tools and systems. This undermines enterprise initiatives.
- Lack of available tools: Until recently, it was nearly impossible to purchase a complete set of BI tools from a single vendor. Some would argue it's still impossible.
- Culture: One BI architect told me that "it has been more difficult to change a legacy BI tool than consolidate all their data marts." People get attached to their BI tools and often are reluctant to give them up.
- Costs: Deploying a new enterprise BI tool can cost millions of dollars in license and maintenance fees. In addition, the cost of configuring, customizing and testing the software, plus training developers and end users and supporting dual BI environments during the interim period can make standardizing on a single enterprise BI tool prohibitively expensive.
- Business volatility: Organizations that succeed in standardizing on an enterprise BI tool often find that business events undermine their plans. Mergers, acquisitions, reorganizations and outsourcing arrangements with third parties quickly dilute the corporate standard with a bevy of third-party alternatives.
Fortunately, a few gallant companies have addressed these challenges and have successfully deployed BI on an enterprise basis. According to TDWI research, 17% of companies have already deployed BI on an enterprise basis and a whopping 66% are in the process. Only 17% have no plans for enterprise BI.
There are several keys to moving to enterprise BI. Here's some advice:
EBI platforms. On a technology front, leading BI vendors such as Business Objects, Cognos, MicroStrategy and others are beginning to offer next-generation products that are BI platforms rather than BI tools. These enterprise BI platforms support a variety of BI operations for a multiplicity of different users on a single integrated architecture that is highly scalable, extensible and flexible. Once these enterprise BI platforms mature, organizations will at least be able to buy a full suite of BI functionality from a single vendor.
Patience. To avoid the political challenges inherent in enterprise BI, most companies grandfather existing BI applications and toolsets and allow these groups to convert at their own initiative.
Executive mandate. To deliver enterprise BI, organizations need a high ranking executive, preferably the CEO, to mandate the BI standard. This sets the tone and helps open doors and minds to change.
Sales mind-set. With an executive mandate in hand, BI managers must then sell the new environment aggressively and continuously. They need to first create an inventory of existing BI tools and applications and create a strategy to convert each one, starting with "low-hanging fruit." To persuade more entrenched groups, managers need to make sure the BI environment delivers the same or greater BI functionality, subject areas and data quality.
Monitor usage. Finally, BI managers need to continuously monitor query and report activity. When activity appears to slow down, they need to investigate the reason. This may entail scheduling time at departmental meetings to discuss user satisfaction with the enterprise BI environment and perhaps scheduling additional training or development to fix problems.
Wayne Eckerson is the director of research and services for The Data Warehousing Institute, a worldwide association of business intelligence and data warehousing professionals. He can be reached at firstname.lastname@example.org.