Is your business successful? Is it healthy? Will it continue to be both?
As business managers, we need to continuously ask these questions, rigorously interpret the answers and quickly identify and implement corrective actions. The answers to these questions all involve some form of measurement – some indication of current activity, a target for success or a goal for the future. The identification and capture of these measurements is commonly known as business performance measurement. The subsequent analysis of these measurements allows us to improve business processes. This also gives us the ability to identify if changes are having a positive or negative impact on the business. This stage of analysis is called business performance management.
Tracking business performance is a common practice today. Traditional performance measures include tracking revenue, cost, the number of units sold and the number of employees. We also use ratios and calculations to provide even more insight, with measures such as return on equity, return on assets and percentage profit. These measures are not only helping business managers better understand their performance, but they are also a standard means for investors and analysts to evaluate the health of an organization.
Identifying, capturing and communicating these measures is not a trivial task. Managers must be able to regularly measure the performance of the business and track the impact of changes to ensure that the enterprise is progressing towards its strategic goals. In recent years, performance measurement has focused primarily on financial measures. While this can be beneficial in many respects, it does not support a big-picture view of the business or accurately reflect the overall health of the organization. Management is now looking to broaden its analysis to include cross-functional business performance.
While management methods and computer applications are still advancing, several approaches have already been developed. These approaches offer solutions that allow executives to analyze and modify internal business processes in order to attain greater health and profitability. Many of these solutions are aimed at measuring the organization’s performance against stated corporate objectives. Each approach shares a common motive: to present a more coordinated, manageable and comprehensive vision of the enterprise.
The Balanced Scorecard
"If you can’t measure it, you can’t manage it" – Kaplan and Norton
The balanced scorecard is a strategic management system that was first proposed in an article, The balanced scorecard: Measures that Drive Performance. The article, authored by Robert Kaplan and David Norton, was published in the January/February 1992 edition of the Harvard Business Review. The authors define an organization’s vision and strategy in terms of specific measurable goals and objectives. The article also describes how to map these strategic objectives into a small set of key performance measures which are tracked by decision makers at all levels of the organization as they identify, measure and manage the elements that have been identified as critical to the success of the company’s vision.
With the balanced scorecard methodology, all of the measurement and management is done in the context of the objectives. These strategic objectives are grouped and measured from four perspectives: finance, customer satisfaction, internal processes and innovation/improvement. The underlying premise is that innovation and improvement within an organization, coupled with effective and efficient business processes, will generate customer satisfaction through quality products and service. The result is improved financial results and corporate health.
A balanced scorecard implementation must incorporate a communication mechanism. The goal: Enabling management to communicate strategic company vision and goals to the employees, who can also report their processes and progress back to senior executives. A well-designed balanced scorecard supports a clear strategy and performance goals. It can be used to determine future results. Executives are involved in defining the processes and targets that will support the company’s goals. By linking a business plan to performance, they can also introduce individual accountability.
For all its benefits, there is still an ongoing debate about the best way to implement and balance the four balanced scorecard perspectives, as well as the extent and speed with which an organization should embrace the Kaplan and Norton approach. The Hurwitz Group, an information technology research firm, believes that balanced scorecard applications are now the leading style of executive information systems (EIS), but most companies are looking for solutions that go well beyond the scope of EIS. They are seeking a scorecard-style approach, but to a broader and more coordinated enterprise of users. What they are really looking for is a solution that gives the right information to the right users in the way that best supports their decision-making style and requirements.
Distributed Scorecarding
The balanced scorecard methodology is only one approach to business performance management. To ensure consistent, effective management of business health across the enterprise, the balanced, multi-metric view that the balanced scorecard offers must live everywhere in the company – not just at the senior executive level. This requires a strong company-wide business performance management initiative. What’s needed is a distributed scorecard solution.
Business today is a truly multidimensional phenomenon. It is comprised of teams, departments, business units and individuals that measure and manage performance in ways that are often complex and frequently unique. To manage business performance effectively, a company must recognize which measures are unique to certain functions and which are common across groups, departments, business units and the entire enterprise. The balanced scorecard is the formalization of performance measures and any scorecard-style performance management solution requires it.
In any performance management situation, the company’s most important asset is the data. What becomes critical is first gaining access to this data, and then presenting it in a way that maximizes an understanding of the business. There is no one correct way to look at data. Just as a scientist can choose from a telescope, microscope, binoculars or magnifying glass to observe a particular event or subject more clearly, decision makers require a variety of ways to visualize data according to particular business requirements. Some users will want to simply view published reports. Others will want full exploration and analysis capabilities. Some will look immediately to the detailed numbers, while others will start with a graphical synopsis. Beyond individual preference, the solution must provide a cohesive view and consistent measurement that addresses all business perspectives. Only when all of these capabilities are brought together do we have an efficient distributed scorecard solution.
Decision Hierarchy of Distributed Scorecarding
Effective enterprise-wide business intelligence brings all of these distributed scorecarding requirements together into a single enterprise solution. It is fully integrated – decision-makers can effortlessly move from big-picture analyses to transaction-level details. This application optimizes performance at every level of the enterprise by providing a comprehensive software solution that ensures that every department is informed and healthy. The enterprise is facilitating better decision making through consistent business performance management.
Seeing is Understanding
We have all heard the adage that "a picture is worth a thousand words." Just think of a weather map in the local newspaper as illustration. These weather maps are really examples of multidimensional data brought to life. They reveal the weather forecast for various regions of the country, state or province and they provide access to critical secondary measures. They give us insight into implications of time because they are forecasts. They give us insight into climate by region. They are suggestive of weather dynamics, of temperature and much more. Most important, all of this information is right before the viewer’s eyes. We do not have to read numbers first, then interpret the numbers to create a mental image of the same knowledge. This is the power of the picture: patterns, trends and occurrences over time are immediately intuitive.
This type of data visualization is critical to a distributed scorecarding solution. A map that groups states or provinces in bright red and then animates these groupings to reveal trends over time, is a far better indicator of a hot spot than any alphabetically sorted list. Effective business intelligence solutions need to bring this power of pictures to business decision-makers. By taking data from multiple sources and departments, an effective business intelligence program presents it all on a single screen utilizing advanced multi-metric displays. Managers get an immediate and intuitive view of what is important in their data. If their only interest is a high-level view of performance, they can present their results as a single aggregate value, a color swatch or a composite of both. The display quickly highlights areas of concern through the use of color, while still providing the ability to drill down and uncover what lies behind the summary.
Visualization Meets Business
While visualization technology has traditionally been used in sophisticated applications that require specialized graphic environments (such as geographical information systems) the robust visualization for business intelligence today brings this technology directly to the typical business user. In its traditional application, visualization technology has focused on creating visualizations that represent complex data. The goal: helping users understand the data.
Distributed scorecarding goes beyond just presenting the status of cross- functional enterprise health to a small number of executive-level decision makers – it enables this view to live everywhere inside the company. Distributed scorecarding centers the decision cycle around business performance management, which means that when a manager has uncovered an anomaly in company performance, he or she can begin to act on it immediately – by drilling down to the details, uncovering what lies beneath the issue and taking action. The end result for the company is a faster resolution of complex business issues and improved decision making.
To guarantee these results, every manger in the organization needs access to the right information. They must be able to understand local performance, and how it relates to targets, minimums/maximums and historical results. OLAP data cubes are an extremely effective way to collect, store and deploy this information. Their very structure matches both the multi dimensions of the data and the business, while opening up new analysis capabilities through online analytical processing (OLAP) analysis, data mining and visualization. Users can explore and analyze their data through interfaces designed for business use across a wide range of behaviors.
Key Performance Indicators: The Pulse of Your Business
A typical business hierarchy may start with strategic business units and divisions. It is then followed by departments, teams and individuals. Within each of these hierarchies there may also be multiple OLAP cubes to accommodate the multitude of key performance indicators (KPIs) which need to be measured. For example, accounts payable clerks will just want an accounts payable status report. However, the accounting manager will need one or more multi-dimensional scorecards for all the measures he or she is charged with tracking. At a higher level, the controller will want a scorecard that is a combination of all the measures from the two lower levels.
To accommodate this architecture, the scorecard can be structured as multiple, aggregated OLAP cubes, with each layer nesting the ones below. In general, the highest-level cube defines the measures and dimensions that apply to the corporate level. These cubes drill through to divisional and departmental cubes that contain more detailed information. This "treeing" of the cubes enables managers to disseminate the strategy throughout the organization, while performance is aggregated back up to the managers.
In addition to being an ideal vehicle to implement a hierarchical architecture, OLAP cubes allow both logical and physical data separation. Data can reside all in one cube where different views are offered to different users. It can also be segmented and maintained in completely separate cubes. In cases where the data is inconsistent in granularity, data allocation provides managers with more accurate and workable views for comparison and analysis.
Once an organization has fully developed its goals and strategies for success, it can use these same goals and strategies to identify the performance measures, which are also frequently referred to as key performance indicators (KPIs). For each measure, decision-makers must track more than just the current performance level. They need to identify and quantify the target or expected level of performance, minimum and maximum acceptable levels, as well as how the current performance relates to each of the other indicators.
These measures may not all be readily available. They may not exist in a company’s current system. Others may exist, but not in the same location. For example, operational performance numbers could be in an Informix OLTP database, historical values in an Oracle data warehouse and planned targets in an Excel spreadsheet. In addition, if the scorecard introduces the perspectives of customer satisfaction and internal process and innovation, it will most likely require new measures, many of which will be based on values in multiple locations. The challenge is making these performance measures available to the scorecard readership while minimizing the time and costs associated with restructuring data storage. Even those who have recently implemented data warehouses and data marts will need to investigate their performance measurement requirements. A distributed scorecard by its very nature presents data across disciplines and functional levels of an organization.
Don Campbell is IBM distinguished engineer and CTO of business intelligence and performance management at IBM. He can be contacted at don.campbell@ca.ibm.com. Don tweets at @IBMCognosCTO.
