The Sweet Spot to Corporate Productivity: Analytical Decision Support Systems
As the information economy continues to evolve with more recent trends involving a consolidation in the technology industry and prevailing economic slowdown around the globe, corporate leaders have increased their demands from providers of IT for tangible benefits from investments in this technology.
More specifically, corporations want to see increased productivity from investments in and deployment of various forms of IT. Over the past few years, many software providers have made claims regarding the potential productivity-enhancing capabilities of various products for companies across industry sectors. Given the robust economy of the period between 1995 and 2000, there was little concern as to the payoff of these investments as demand and profits increased steadily. However, this "golden era" came to an end at the dawn of the new millennium as U.S. GDP remains in a depressed state today. The result from this downturn is an increased requirement of state-of- the-art systems to more clearly illustrate tangible productivity-enhancing capabilities.
One such sector in the IT software spectrum that may be able to satisfy this requirement involves analytical technology. Analytical technology can encompass a wide variety of products in the software sector. For purposes of this article, analytical technology refers to products that enable decision-makers to perform simulations or sensitivity analysis corresponding to various business processes in an organization. This technology can range from advanced interrelated spreadsheets in its most crude form to the latest offerings in the data mining solution space.
Bottom-Line Analytical Approach
The spreadsheet approach is one of the most basic forms of analytical methodologies, as properly designed functions that incorporate relevant business drivers enable decision-makers to see the interaction effect among driving variables for a particular process. This interaction effect merely illustrates various outcomes in process measurements given changes in variables that describe them. For example, if I reduce labor in one functional area and pursue financing strategies in another, how does this affect corresponding costs?
The spreadsheet world can no doubt provide a tangible value add, but in today's corporate structure, there is an increasing requirement to provide spreadsheet capabilities along with mathematical and statistical capabilities to vast amounts of data across functional areas, and the capabilities need to be accessible by numerous users. A variety of software vendors has addressed this requirement in producing functionality to address a host of financial applications.
Analytical applications also involve the utilization of data mining methodologies, incorporating mathematical, statistical and visual functionality for data examination in order to augment the knowledge of decision-makers. This increased knowledge provides a greater understanding of the marketplace in which the decision-makers operate. Because of some prevailing misconceptions regarding the definition of data mining, we will provide a clear definition.
In the context of this article and for the commercial world in general, data mining includes methodologies such as segmentation, neural networking, clustering and regression. Data mining largely incorporates the use of mathematics and statistics to analyze business- related data. It differs from the spreadsheet world described earlier because it involves a bit more uncertainty regarding the measurement of a business process.
The spreadsheet world largely involves a series of interrelated equations that incorporate direct relationships between variables. In other words, if I reduce head count in one area by 50, I know that my costs will diminish by (x) amount. In the data mining spectrum, decision-makers seek to gain a greater understanding of particular processes that are not fully explained by variables in an equation. For example, if I launch a particular marketing campaign for product A, what can I expect as a response rate from consumers within an acceptable variance?
Although some mining methodologies do not enable users to perform simulations into the future, they do provide a more reliable explanation of what is driving a business process. This reliability is obtained through the use of mathematical functions, algorithms and statistics.
One of the true value adds of data mining lies in its ability to examine a variety of business processes. Various data mining methodologies can help increase the knowledge of decision-makers across functional areas and in a variety of industries. Common applications include the ever-important CRM space, marketing and advertising, manufacturing, demand forecasting, product pricing, human resources, fraud detection and e-commerce. In fact, when considering some of the more robust mining methodologies, (e.g., neural networking and regression), decision- makers can actually gain a better understanding of the macro economy (e.g., GDP) as well.
Don't Ignore the Macro Economy
One of the main causes of poor corporate performance over the past year has been the inability to predict the slowdown in the overall economy. Companies continued to expand facilities, produce products and hire employees to meet the expected continued strong demand that prevailed into the beginning of the year 2000. As a result, firms were caught with an overabundance in the face of decreased demand for products and services.
Some of the more robust mining methodologies can help decrease the uncertainty of aggregate demand forecasts through the use of macroeconomic, time-series models. It must be noted, however, that this latter application delves into the space of econometric modeling.
What's the Connection?
The connection between analytical applications and corporate productivity lies in the issue of knowledge creation and reduced uncertainty in the marketplace. The more basic form of analytics (e.g., the ability to conduct what-if simulations from interrelated equations) enables decision-makers to fine-tune business strategies rather than merely base actions on ad hoc assumptions. These applications provide tangible, direct answers to business simulations that help decision-makers augment corresponding strategies.
Results from data mining analytical analysis provide another value add to decision-makers. The various methodologies enable analysts to leverage vast amounts of data that is at the disposal of companies in all industries. In a nutshell, data mining gives analysts the ability to extract information from data that helps create and augment knowledge of many different business processes. This enhanced knowledge helps reduce the uncertainty as to what drives these processes, thereby providing powerful and tangible benefits to decision-makers.
Through the use of segmentation and decision trees, analysts can quickly determine paths of success or failure in business processes and highlight threshold points among variables that drive a process. Neural networks and regression give users a more detailed vision of how particular variables affect target or performance measures. Resulting models also permit decision-makers to perform simulations to test proposed strategies.
Common efficiency enhancements from data mining include:
- More accurate identification of target markets.
- Appropriate pricing, marketing and advertising activities in bringing a product or service to market.
- Optimization of production processes and inventory based on demand forecasting.
- Reduced variance or error rates in production processes (pursuing Six Sigma).
- More effective motivation and compensation for employees.
- Optimization of Web sites and e-commerce strategies to more efficiently accommodate consumers.
- Mitigation of fraudulent activities that erode operating margins.
As powerful as analytic applications may be, it is imperative for decision-makers to monitor the performance of these applications. Analytical applications increase knowledge and reduce uncertainty of business processes, which helps enhance the effectiveness of strategies. However, in order to attain a measure of real efficiency, decision-makers must examine the success of these devised strategies in the marketplace.
Efficiency and productivity are augmented when enhanced knowledge enables firms to devise and implement strategies that achieve:
- Consistent desired output or goals with a lesser amount of input resources, or
- Increased output or target measure with the same amount of input or productive resources.
Increased productivity and efficiency can only be measured by examining post-strategy results which provide the tangible evidence as to the value add of analytical applications.
When results in the marketplace are close to what was expected by decision-makers, the value of the methodologies that were used to help devise those goals is high. When the difference between actual results and what was expected is large, the value of those methodologies used to devise those goals is diminished.
The key to achieving increased productivity or efficiency is the reduction of the operating variance (i.e., established goals from implemented strategy minus actual results) mentioned earlier. A low variance means that results are within a reasonable range of established goals. Achieving a low variance drives productivity because operational resources have
been adjusted to facilitate the level of desired or expected target measures.
In other words, if decision-makers use analytical methodologies to determine response rates for a particular marketing or advertising campaign, they adjust resources (e.g., designated employees and inventory) to meet that demand. If the operating variance is large, there will be an under- or over-allocation of firm resources, which reduces productivity. Conversely, a low operating variance means that the firm is utilizing designated resources in a more efficient manner.
The Final Step
In order to establish and maintain productivity, decision-makers must constantly monitor operating variances. If they have been able to achieve acceptable variances, future activities require them to improve or maintain variances. If variances are too large, decision-makers must more closely examine and identify the potential causes of the breakdown in strategy. This may require adjustment of existing models that provided the baseline information used to devise corresponding strategies. The adjustment process may include the addition of more data and a closer look at the activities of competitors.
Today's information age involves the collection of vast amounts of data by firms across industry sectors. The key to increasing productivity for firms is to more fully extract information from existing data. The resulting information helps create and increase knowledge of the processes that are crucial to a firm's performance. A variety of technologies help store, access and manipulate data. However, one of the greatest value adds to creating knowledge is to implement analytical methodologies that help identify reliable relationships and interactions between variables that drive these processes. The result should be a reduction in operating variances and an increase in productivity.