Rosella       Machine Intelligence & Data Mining

HomeDataMining & Machine LearningProducts & DownloadsSite MapContact |

The Balanced Scorecard (BSC)

The Balanced Scorecard (BSC) is a framework for managing business performance. Balanced scorecards provide concise, predictive and actionable information about how a company is performing and may perform in the future. BSC provides a framework for designing a set of measures for business activities as being the key drivers of the business or Key Performance Indicators (KPIs). KPIs are collected from CRM, ERP, Accounting, Personnel, Inventory, and so on. Original versions by Kaplan and Norton consist of four major perspectives: financial, customer, internal process and innovation and learning;

  • Financial perspective describes financial business drivers such as financial results, cash flow, revenues, and so on.
  • Customer perspective describes customer centric business drivers such as customer satisfaction, customer churns, customer loyalty, etc.
  • Internal process perspective is concerned with internal business activities created to deliver customer services. It may include numbers of transactions processed, accident ratios, equipment failures, etc.
  • Learning and growth perspective is concerned with drivers related to intangible business infrastructure such as employee training, investment, so on.

What are the good balanced scorecards?

Good balanced scorecards might be said to have good representation on good quality business drivers or KPIs. Qualities of good KPIs include;

  • Valid & agreed upon: drivers must be valid and agreed upon by stakeholders.
  • Specific & measurable: drivers must be specific and measurable systematically.
  • Reliable: information used as KPIs must be reliable.
  • Relevant: drivers must be relevant to business.
  • Achievable: targets assigned for drivers must be achievable. Otherwise drivers will be meaningless to include.
  • Easily understood: drivers should be easily understood by users. Complex and obscure drivers may not be useful.
  • Timely: drivers must use timely information obtained in a timely manner.

KPIs having these attributes alone do not produce good balanced scorecards (BSC). Good BSCs should further posses the following attributes;

  • Simple to see and interpret.
  • Support "drill-through" multiple dimensions and angles for critical business drivers.
  • Enable early detection of potential problems and successes.
  • Emphasize significant and success-determining business drivers.
  • Highlight leading predictive indicators as they are more useful than trailing indicators.

A study found that the full benefits of effective balanced scorecards are not being realized for more than 80 percent of typical companies examined by the survey. Most executives were not able to take the balanced scorecard from concept to reality. Primary reasons include too many business drivers and places far too much weight on historical performance and not enough emphasis on forward-looking measures. This exposes the problems with not-so-well-developed balanced scorecards. The following section describes how predictive inferential technology can lead to improvement. For the survey story, please read ...

Knowledge-Enhanced Predictive Balanced Scorecard improve business visibility!

Knowledge-enhanced Predictive Reports (KPRs) can improve business visibility harnessing BSC with predictive modeling and business logic using expert systems. KPRs can analyze changes in business drivers and co-inference them automatically to detect hidden patterns underneath complex numbers. KPRs incorporate predictive modeling with rule-based expert systems into report writing and charting systems;

  • Predictive analytics can be used to detect patterns and trends in business drivers automatically from hidden numbers, and to predict future directions. It is known that leading predictive indicators are more useful than trailing indicators. Directions and projections can be very useful information to have.
  • Rule-based expert systems can be used to leverage complexity of various business drivers and indicators. As the survey mentioned found, understanding too many drivers and complex numbers can be very daunting tasks for executives and business users. Expert systems based on business logic can take this task as an expert, making balanced scorecards more friendlier and easier to understand.
  • Web-based reporting & charting engines are essential in generating balanced scorecards in a timely real-time fashion so that executives and business users can recognize developing situation in real-time.

Incorporation of predictive analytics and rule-based expert systems into BSC provides a number of advantages. First of all, it will make BSC much easier to comprehend potential problems and successes. Trends developing can be detected early so that actions can be followed as early as possible. Complexity in interpreting KPIs is removed in real-time by embedded knowledge of business experts.

In addition, governance and compliance rules may be implemented into BSCs, making any potential violations automatically visible to stakeholders. Without complex rule processing engines, this task is almost impossible. Sophisticated predictive models may be employed.

The following figure shows a part of a predictive BSC dashboard. "Trending" and "Course" are based on time-series predictive analysis. Signals and figure colors are determined by expert systems which execute business logic. Clearly, this scorecard dashboard meets all the criteria for good balanced scorecards!

Predictive Balanced Scorcard Dashboard Example.

Trending shows recent figures (in seagreen) with near future predictions (in skyblue). (Note that near future figures are predicted using growth models.) Orange bars indicate target figures. "Course" indicates trend patterns. If positive patterns develop, it indicates to green. If negative patterns develop, it indicates to red. If no clear patterns emerge and fluctuations develop, it indicates to yellow for warning. Signals indicate overall health of KPIs. Monthly figures are colored based on pre-defined business rules on target figures.

Time-series predictive regression techniques are used in analyzing trends. Rule-based modeling (RME) and RME-EP Expert Systems engines are used to determine status and signals of indicators.

For more on predictive balanced scorecard, please write to us!