Business Intelligence Best Practices - BI-BestPractices.com

Collaboration. Communication. Community.

 
 
 Printer-friendly
 E-mail to friend
  Comments
ADVERTISEMENT
Beyond the Dashboard: Making Better Decisions

by Gabriel Fuchs
Charts and graphics are integral parts of modern business reporting. The reason is easy to understand: charts give a richer view of a specific business activity, often highlighting important trends.

Application charts and graphics can be divided into two groups. The first consists of tools that visually present specific business situations; they are commonly integrated in dashboards. In his article, “Dashboard or Scorecard: Which Should You Use?” author Wayne Eckerson points out that dashboards monitor and measure processes and can alert users “when performance against any number of metrics deviates from the norm.” In discussing dashboards and scorecards, Eckerson concludes that “the key is whether you can leverage these new tools to inform users with actionable information so they can more effectively execute the organization’s key strategic objectives, plans, and business processes.” (Eckerson, 2005)

A dashboard’s strength is also its limitation: it provides only a high-level business overview measured against predefined metrics. This prevents users from asking and answering new questions, and experiencing new relationships between different types of information. Often, users cannot drill down into data to see what is causing the top-line results. Additionally, decisions based on realtime dashboard information often exist in a black hole: they are made and executed, but the decision-making process is never passed down through the organization. The result is a need for something beyond the standard dashboard: the visual, interactive dashboard.

Charts and graphs are also derived from software that analyzes large and complex sets of data, commonly referred to as analytics, fueled by improved computing speeds and an ever-increasing business need to better understand and act on constantly growing volumes of business data. The resulting applications are often referred to as tools for visual interactive analytics and can be used for advanced data analyses.

Whereas both visual interactive analytics and visual interactive ashboards can add enormous value to the business when used efficiently, the integration between these two related activities is often neglected. There are usually few or no business processes describing how analyses should feed decisions that are then measured and fed back to the analysts. In his article “Avoid EIS Syndrome: MakeYour Dashboard Actionable,” Craig Schiff points out: “What is the point of looking at data that is either not important to you or is of interest but is not actionable?” (Schiff, 2005)

Dashboards alone are incapable of supporting the complexity of decision-making processes. Because analytics must be simplified to be deployed to business users, these requirements must be integrated into a single, seamless application accessible by all users. Only then can the user experience be truly scalable to the needs of users with a wide variety of skills.

In other words, visual interactive analytics and dashboards must be integrated—with guided activities for business users—to provide a single tool for decision making. This integrated solution will create a closed loop wherein the activities of the analyst are integrated with the business decisions, and where the outcome of these decisions can be fed back to inform and refine subsequent analyses. As a result of this feedback from business activities, a positive return on experience (ROE) is achieved, allowing the organization to continuously improve its decision-making process while accounting for the end-user experience.

Single versus Multiple Tools

A common trend for software suppliers, whether in the fields of nterprise resource planning (ERP), business intelligence (BI), or analytics, is to offer more functionality in a single tool for a more diverse set of users. This seems like a logical development, as these suppliers are driven to broaden their offering in accordance with new and changing client needs. Such an approach can be beneficial to clients, as it can simplify their IT administration, allowing the business user to work with a single and familiar interface and to keep down overall costs.

Nonetheless, there is an inherent risk with solutions that offer “all-in-one” tools. A supplier may have a solution that addresses a core business strength for a particular class of users. Risks arise when these solutions also produce added features that are not part of a supplier’s core knowledge. As a result, these added functionalities are often far from best-of-breed. As soon as a supplier steps out of its core business, it is worth reevaluating whether to buy the whole kit from them. Even though it may be temptingly convenient to buy a single tool from a single vendor, it may, in the end, be a disappointing decision.

On the other hand, vendors having a long track record for a specific kind of solution—even complex ones—can offer best-of-breed single-tool solutions that cover a vast range of business needs. The proven experience of a vendor is central to choosing a single tool and singlevendor solution. Without such an experience, it may instead be more interesting to choose several vendors in order to assure optimal business solutions.

Thus, at a high level, whether to choose a single tool or several tools depends on the needs of the organization coupled with the experience of the vendor behind a given solution. There is no universal answer. Having several tools can truly optimize different specific business activities, but if these tools are unable to communicate efficiently with each other, the whole organization will suffer.

Visual Interactive Dashboards

Dashboards attempt to give decision makers a complete overview of their business operations, and should allowfollow-up on important business decisions without a complicated interface. This is achieved by using charts that are relatively simple compared to some of the graphical displays that are often applied in more sophisticated analytic applications.

However, beyond the state of the dashboard today, decision makers must be able to ask and answer their own questions. When the decision maker needs to explore something of particular interest, that user should be able to get the detailed underlying data. No decision maker will be satisfied with red and green colors indicating good or poor performance; while useful as a quick overview, detailed information is needed to pinpoint the exact reason for a specific indicator.

Analysts need to be able to work interactively with data, as do decision makers with dashboards. Thus, the detailed data behind a dashboard should be available interactively so the decision-making process is not interrupted by lengthy response times or a development cycle involving the IT organization.

Figure 1 shows such a visual interactive dashboard, where simple dots are colored according to how sales compare to targets to quickly show how different regions are performing with different products. By clicking on the red dot (shown here as the darkest gray) at the intersection of the North region and product Civil Liability, the decision maker sees from detail at the bottom that the problem concerns the sales reps focused on developing the private customer segment.

AVisual,InteractiveDashboard

Figure 1. A visual, interactive dashboard

Thanks to the interactive link between the overview and the underlying data, the manager quickly recognizes a problem and its source, and can take the proper corrective actions.

Integrating Analytics and the Dashboard—Closing the Loop

It is essential that the measured results do not stay with the decision maker who looks at the dashboard. As more than one individual is usually involved in decisionmaking activities, each with varying skills, all should be clearly integrated into the decision-making process. The overall purpose is to turn the whole organization into a more efficient and focused entity, something achieved by implementing a closed loop where every business activity feeds another analysis or decision. Uniformity of this decision-making process is key to productivity.

Furthermore, all business users should get the maximum guidance to automate repetitive tasks, which act as barriers to participation as well as efficiency. The purpose is two-fold. First, the business users should be able to focus on adding value, not on spending time accessing data, constructing queries, and formatting reports in order to get their business-critical information. The more that can be automated and guided, the easier his or her tasks will be. This does not imply, of course, that the user is blocked from using his or her own initiative when necessary.

Given different user roles, the proper mix of freedom and guidance is essential to ensure an optimal decisionmaking process and yield confident business decisions.

The second important aspect with guided activities, whether fixed processes with inherent business rules or recommendations on how to navigate certain analytic procedures, is that they help the organization as a whole to better utilize the experience of its employees. Best practices can be put in place and shared with others.

This formalized ROE is essential to avoid repeating bad decisions. An organization may not be blamed for making a mistake, but it is stupid if it repeats the mistake. ROE also allows organization to better reap the benefits of good decisions, as they can be shared and thus easily refined further. Such a best practice should be considered a key driver to organizational success.

In practice, achieving and sustaining a healthy ROE through a closed-loop decision-making process requires going beyond the traditional dashboard. Only then can all users become involved and contribute their part to the decision-making process.

The Visual Interactive Closed Loop

Today, most organizations follow a well defined process when putting in place decision support systems. The business-critical key performance indicators are identified (another area for improvement at some organizations [Schiff, 2005]), followed by some serious and timeconsuming technical work (such as data modeling, data extraction, and query development). Once the data has been prepared, analyses and reports are distributed. Well defined as this procedure may appear, it does suffer from deficiencies, the major one being a lack of interconnection where each business activity feeds another. After analyses and report distribution, follow-up is often poor on how these reports are actually used and how they serve subsequent business decisions.

The result is that this “well defined” process often produces something akin to a black hole, where the value of the decision support systems clearly diminishes once the analyses and reports have been distributed. This situation will inevitably lead to poor decisions and acorrespondingly low ROE. Obviously, there should not be any black holes where the decision-making process abruptly terminates. To optimize the decision-making process and to make sure it adheres to the highest quality standards, the analyses on which the decisions are based must get formalized feedback on the results of these decisions. Only then can the analyses continually improve and thereby constantly advance the quality of the future business decisions.

With such a closed-loop decision-making system, the quality of both the decision and the underlying analyses is measured. Figure 2 illustrates this optimal decisionmaking system. The activities within this system should let the business users be guided to the largest possible extent to formalize and facilitate their work.

AClosedLoopDecision-MakingSystem

Figure 2. A closed loop decision-making system

In a way, two masters must be served to deliver integrated analytic capability across the organization. On the business side, these steps have to be communicated to a mixed audience with diverse expectations, meaning that the tool must be graphical, interactive, and scalable. Only then can the tool be agile and accessible enough to answer ad hoc questions. On the technical side, a single system is required for all activities in the closed loop decision-making process: analysis, guided decision making, dashboards, and analysis refinement. At every step, remember that interactivity is a crucial part of the process.

The result of such a system is the creation of a true visual, interactive, closed loop for decision making to support maximum ROE.

Without doubt, one of the major activities for an organization to manage (in order to thrive) is the continuous improvement of its business processes. Every decision implemented will test the quality of the organization’s business processes, and presents an opportunity to improve these processes. Consequently, every decision needs to be measured and clearly communicated so the quality of the decision and the quality of the process behind the decision can be validated and optimized.

What is required is integration between the different phases of decision making. The analysis that provides the basis for many decisions must remain closely connected to the outcome of these decisions. Only then can an organization fully understand and refine its analyses. At the same time, the decision makers must measure the results of their decisions, whether they be operational, tactical, or strategic, to determine the quality of those decisions and improve future decisions.

The best way to develop this closed loop and optimize the decision-making process is to guide business users as much as possible—just as it is done today for other operational business processes (such as order handling and customer service). The outcome of these guided and interactive steps shall always feed another step. By understanding and sharing the outcomes of these activities, ROE is maximized, which is essential to avoid repeating bad decisions.

The ultimate result will be a visual interactive closed loop, involving both visual interactive analytics and visual interactive dashboards, to constantly improve, in a measurable way, an organization’s decisions.

REFERENCES

Eckerson, W. “Dashboard or Scorecard: Which Should You Use?” TDWI Case Studies & Solutions electronic newsletter, January 11, 2005.

Schiff, C. “Avoid EIS Syndrome: Make Your Dashboard Actionable,” TDWI FlashPoint electronic newsletter, November 3, 2005.

Gabriel Fuchs -

Gabriel has more than ten years of experience in all aspects of the business intelligence value chain. He has witnessed all the hype, seen technologies come and go, and observes that basics still surprisingly often rule in the business intelligence world. Gabriel’s somewhat ironic writings are based on his own personal experience and imagination, and do not reflect the situation at any particular company. 

Gabriel is a renowned expert within strategic IT solutions, including business intelligence, performance management and business analytics. He has worked within a range of different industries and activities all over Europe, helping organizations align key operational activities with the strategic goals. His book, Dealing with Nasty Colleagues: The Art of Winning in Office Politics While Still Getting the Job Done, can be ordered at http://www.amazon.co.uk/. Gabriel can be reached at sgfuchs@bluewin.ch.