BI in the Enterprise
BI normally starts off as more of a passive, opportunistic activity, focused on the delivery of information via analytic applications used by “specialists” or “subject matter experts”. These analysts produce reports identifying high level problems and then focus on a specific problem to provide more insight. These reports are then provided to senior managers who are then expected to take corrective action. With this type of BI, the focus is generally more on maximizing efficiency and reducing cost. Analyses also tend to focus on either one department or one problem at a time, and tend to focus and rely on historical data and trend analysis techniques.
As senior managers start being held accountable for meeting proscribed performance targets, they in turn drive this down to middle managers who are expected to “make sure we look good”. Now more detailed information and timely analysis is needed. This is when a distinct cultural shift starts taking place, and instead of being reactive (get the report, then take action), users start wanting to be more proactive and start asking questions like, “What do I need to do to NOT have the report show bad numbers?” To meet those needs, BI activities start becoming integrated into tactical planning and operations management. The focus now shifts to process effectiveness, both within a department and among departments, as middle managers use the “intelligence” provided to modify or adapt the processes that they are responsible for.
As greater insight into the business becomes available, a shift also starts occurring at the strategic level. Business strategy and execution, normally defined at a very high level (often because that was the only level of information easily available) and working with only end-point metrics (e.g. total sales and cost of good sold), can now be defined and analyzed in much greater detail. Business activities can be modeled, and business assumptions and decisions tested, to determine the impact of a high level strategy to downstream processes and resources. For example, a comprehensive business model will be able to accept the strategy data of “increase sales of Product A by 15%” and determine the impact that would have on parts inventory, manufacturing capacity and activity levels, warehouse storage, distribution channels, and customer support. At this point, BI is now pervasive in the enterprise and provides a more holistic view of the organization and the interrelation of its activities.
Lastly the extended enterprise can be included in a BI initiative by gathering data from customers, suppliers, business partners, and other external sources, and using this data in the activity models to validate strategic assumptions and drive tactical planning. This is a practical BI “nirvana” where the organization is working proactively and collaboratively internally, as well as with its external value network, constantly using “business intelligence” to manage its processes and optimize its performance (and those of its business partners too).
Enterprise Architecture for Business Intelligence
BI is also one of the enabling technologies for Performance Management (PM). A diagram from Gartner (see Figure 1) perhaps best represents the relationship between BI and EPM. The foundation layer is the Information Management Infrastructure, which consists of all of an organization’s databases, servers, transactional applications, network, and workstations. The next two layers are both related to BI: the BI Platform, which represents the “back-end” technical platform for storing BI data; and Analytic Applications, the “front-end” applications that allow users to access, analyze, and report on the BI data.
Between the PM layer and the BI layers are People and Processes. The PM layer identifies the activities and metrics that should be monitored. It also defines the acceptable range of performance, which is used to identify performance exceptions (e.g., significant variances or deviations from the norm). These exceptions are brought to the attention of a Person who then makes decisions and takes action within the context of aProcess. Without competent people and adaptable processes, the value of BI and PM cannot be realized.
As always, all of a business’ activities should be driven by the Business Strategy—ensuring that all technologies, activities, and analyses support the business strategy is the key to successfully deploying and maintaining a healthy BI and PM initiative.
Starting a Business Intelligence Initiative
As the story in the first section described, many times BI initiatives will grow themselves as the demand for information and accountability for performance is driven down through the different levels and among the different departments of an organization. As this growth happens, the key is to focus on the cohesiveness and standardization of the BI architecture. The presentation of information, and the calculations and data it is based on, must be consistent for information to be used organization-wide. Remember the mantra “one version of the truth”—a goal that BI shares with data warehousing, one of its supporting technologies.
That leads to the last tip when starting a BI initiative: start small, but think big. Because the eventual goal is to have BI pervasive throughout the organization, even when starting with a small point project, the bigger picture and organization-wide BI vision should also be considered. This will help to ensure that the data structure is solid yet adaptable (as the data collected grows), the technologies chosen are scalable to support the whole organization, and the application functionality is flexible to be used in varying analyses.
The CFO's Call to Action
CFOs should start by working with business units to identify the information they need to support and monitor their business processes. This usually starts by looking at what information is needed to support decision making—where does the data come from and how is it used? Another perspective to look at is how are business processes monitored—what are the key metrics that should be monitored and what is the acceptable range for each metric? CFOs may also approach this from the internal audit perspective—what are the control points in a process, and what data conditions identify that an exception has occurred?
Once all of these requirements have been identified, CFOs should work with IT to develop a strategy for getting the data needed into the BI Platform. During this phase, CFOs should review the processes and algorithms used to gather, compile, and move data from operational data sources (i.e. transactional systems) into the BI Platform, ensuring that data quality (accuracy, correctness, and completeness) is maintained.
To enable the business units to use the data, the proper reports or alerts must be configured in the Analytic Applications. Most modern Analytic Applicationshave Microsoft Excel-like functionality making it easy for a non-techie to develop reports and wizard-type step-by-step configurators that support the setup of alerts so that CFOs can perform most of these tasks by themselves.
Lastly, CFOs should work with the business units to help them understand how to work with the reports or alerts that were setup, and also to understand the information that is being presented. This is the step most commonly forgotten but it is the most critical. Operational people may not have the background or knowledge to be able to understand what a report is telling them. So training them to understand what they’re reading and helping them to understand what effect their work has on numbers in the reports, is what really empowers them to take action. Remember that the endpoint goal is to enable people to make informed and proactive decisions quickly at all levels of the organization.
CFOs should also partner with the IT department to ensure that the risks from the BI technical architecture are properly managed. As decision makers come to rely on the information provided by BI processes, CFOs need to work with IT to ensure that systems are properly backed-up and that business continuity measures are put in place to prevent BI system downtime. And of course, CFOs should review the logical access controls over BI data since it often contains competitive intelligence that should only be accessed by those with proper authorization.
Successful BI initiatives require broad business perspective, high discipline, and strong analytical skills, all of which CFOs exemplify. By starting small and incrementing on success, CFOs can help empower decision making at all levels of the organization. Business Intelligence was #8 on the AICPA Top Technology Initiatives this year, but I’m sure it will continue to move upward as quick, accurate, and well informed decisions and actions become a cornerstone of business strategy execution.