In my previous blog post, I showed that integrating “classic” BI tasks such as reporting and data analysis with planning and other performance management tasks is a highly relevant requirement for organizations. It is key to leveraging a company’s strategy and steering its operational processes. Other benefits range from cost reduction and improved data quality to increased agility.
In this blog post, I will present a holistic view on decision support and a framework that helps companies structure and relate their activities in this field using a closed-loop approach, which serves as a foundation for the effectiveness and efficiency of business intelligence and analytics on all levels.
Increasing need for an integrated approach to decision support
Growing demand for self-service BI, data discovery/visualization, Big Data, IoT, and predictive analytics is increasing organizations’ awareness of data as a key asset and analytics as the key capability to extract value from data. We also see increasing interest in developing data as a product and/or data-as-a-service offerings, brought to market both internally and externally.
But lines are becoming blurred. Traditional business intelligence requirements in reporting, analysis, and planning are complemented by explorative BI and operational BI use cases. Explorative BI needs a different approach from traditional BI, since processes (“fail fast”), architecture and tooling (advanced analytics) as well as the required skills (e.g., the data scientist) differ. Operational BI describes the trend of increasingly embedding BI into processes so that more information and analysis possibilities are available in the operational process. Also, process execution becomes the subject of detailed analysis in process mining.
A holistic view on all decision support tasks helps to correctly set up and assign tasks and projects and stay on top of the fast-moving world, where new data sources, data users, and analytical possibilities and ideas emerge regularly. The aspect of integration is crucial (as it often is in technology) to avoid breaks in data definitions, logic and output interfaces – breaks that are usually costly to fix and that lead to unhappy users.
Strategically, an integrated approach is also a prerequisite for a successful evolution towards becoming a data-driven company and using data more effectively and more actively.
Closed loops to manage an organization
To help companies plan, structure, prioritize, explain, and manage their initiatives and investments in decision support, I recommend an approach that has been helpful in many of our projects.
A simple definition of “management” is to plan, steer, and control an organization’s performance. The “BARC Decision Support Framework” shows how management tasks relate to each other, but also how to process execution and strategy management (see Figure 1).
Decisions are taken on many levels—strategically, tactically, and operationally—and include data about the past (e.g., in reporting) and about processes currently running (in activity monitoring), but increasingly also about the future (in planning, forecasting, and predictive analytics). Decisions should also be aligned with the organization’s vision and strategy as well as the objectives and targets of individuals and departments. The combination of these levels and the closed loop from using data about the past to predicting the future is shown in Figure 1:
Figure 1: The BARC Decision Support Framework
Let’s take a closer look at the three levels.
Business intelligence (BI) is the foundation of performance management in most organizations: Planning tries to anticipate and reflect future developments in an organization’s process goals and structures. Planning and controlling an organization’s performance relies on the analysis of the past outcomes of these processes – data analysis creates new information from available data.
Selected process results are usually reported once processes have finished executing—typically in daily, weekly, or monthly cadences—and aggregated to key performance indicators for further analysis. Supporting this closed loop, from reporting and analyzing process results to prediction and planning for adapting operational processes, has always been a key task for BI and performance management software. Even today, BI systems are still mostly used for tactical decision-making.
The rapidly increasing need for faster (up to real-time) decision-making has added emphasis to the steering of an organization’s performance, adding information on present process execution to the views on the past, in reporting and analysis, and on the future in planning. Enabling rapid reaction to current developments by monitoring and analyzing quality, time, and cost of processes is the goal of operational decision-making.
Monitoring processes using event monitoring and analysis to enable a rapid reaction constitutes the operational closed loop in the Decisions Support Framework. Popular methods to implement this include the integration of data visualization and analysis into operational systems, creating the capability to process and analyze events using data streaming, rules engines or complex event processing systems, and also the automation of decisions in processes using rules or predictive models.
Steering, controlling, and planning should be based on an integrated performance indicator framework that is aligned with the organization’s strategic objectives and the targets of decision makers. Defining this framework and aligning it with the organization’s goals is one of the key tasks of strategy management, adding direction and content to the business intelligence tasks and systems used for measuring the success of the execution of a strategy in operational processes.
How to get started
So, how to get started with achieving a more holistic approach to decision support?
Looking at integration is a key. Creating integrated systems for business intelligence and performance management tasks is a good starting point, especially between planning on the one hand, and reporting and analysis on the other.
Evaluating use cases and the strategic value of more informed or automated decisions on an operational level comes next. Setting up embedded business intelligence is often not only about monitoring process execution, also analyzing what is happening based on historical data. This creates the link between tactical business intelligence and operational BI.
Lastly, the questions “What is actually measured in BI systems?” and “How do KPIs relate to each other?” provide a good starting point for setting up integrated KPI models that can create a link between the organization’s vision and objectives with day-to-day decisions based on data.
For more insight on becoming a data-driven organization, see BI And Planning: Integrate It Or Suffer Forever.