Teradata Magazine Online

Tech2Tech

Tech Knowledge

Perceptions into Analytics

Perspective integration combines data and human feelings to enable business growth.

by Takehiko Ikeda, Ph.D.

We all know data is critical for making good decisions. But so is information regarding customers’ purchasing behavior and the intuitive knowledge of company employees. A new methodology, called perspective integration (PI), uses tradi­tional data mining along with people’s feelings and senses for analysis to help form strategies and aid in the decision making process. For example, since growth is imperative to a company’s sur­vival, organizations can use PI to analyze consumer and employee insights about a company, its products, sales and competi­tors’ businesses to develop a comprehen­sive growth strategy.

By using statistically verifying algo­rithms, PI can also monitor the gap between planned goals and reality. If necessary, the organization can modify its plan and manage the business process. In addition, the PI methodology brings strate­gic and operational intelligence together under one disciplined entity, then delivers that information to front-line employees, such as salespeople, so they can offer the best services to their customers.

Analytical Methodologies

PI uses traditional statistical methodolo­gies like time series analysis, multivari­ate statistical analysis and data mining, but what makes it unique is that it also incorporates people’s perceptions into algorithms that facilitate processing based on those feelings:

• A multiple-attribute decision making (MADM) algorithm is used when decision makers are required to choose between several options and each option has a variety of attributes. The algorithm also analyzes data derived from consumers’ senses to arrange consumers into groups.
• A multiple objective decision making (MODM) algorithm achieves more than one objective. It is normally associ­ated with mathematical programming.
• A group decision making (GDM) algorithm uses group ideas and analyzes the cause-effect relationship of complex problems.

Add Human Senses and Feelings

A major challenge for organizations is how to use their information systems for decision making. While business intel­ligence (BI) has long been acknowledged for enabling decisions, in many cases, decision makers don’t capitalize on the data they acquire through BI. Instead, they use information as reference materi­als and make decisions based on intuition and experience.

Click to enlarge

Instead of using consumer informa­tion and experienced employees’ percep­tions after data analytics are performed, PI feeds this information in advance to analytic algorithms that deal statistically with human senses to directly support decision makers.

Typically, decision makers must choose between or prioritize several options. PI, using people’s senses and feelings, chooses the best option or prioritizes them based on an algorithm.

STRATEGY

The figure shows a PI marketing strategy development structure. Strategic alternatives are derived by analyzing internal and external data, customers’ purchasing behavior, trend information and the percep­tions of decision makers. In this analysis, customers’ value is evaluated by five forecasting methodologies that are provided as the PI analytics knowledgebase.

Building sales forecasting models of customers is an essential part of a growth strategy. These models are based on:

• Past trends
• Consumers’ mindset
• Human senses
• Analyzing leading indicators
• Taking competitors’ market share

If the strategic alternatives are market segments, customers are grouped together according to similar characteristics. The groups are then evaluated and prioritized, once again using perceptions. Because the customers in each segment share attributes, the same development strategy can be used for the entire group.

TACTICS

Normally, characteristics of strategic alterna­tives and tactics development models are given in mathematical models. If the pur­pose of a project is to increase sales, a sales forecasting model with factors that impact sales as independent variables is defined for each strategic group.

The intuition of experienced people can be embedded in the models. The models for tactics, which are different from the strategic models, are ideally updated as frequently as possible, based on the most recent data. This is different from the traditional approach that explores past data and tells what actions are effective and their degree of influence.

MEET THE OBJECTIVE

Once the tactics models are well defined and updated, they can deliver action plans to the sales team. By following these plans for dealing with customers, salespeople can determine the most effective way to achieve the highest customer satisfaction.

PI maintains a paper trail for hypothesis verification. It keeps the original data in a mathematical form so others can easily check the decision makers’ hypothesis. PI also draws a step-by-step spiral chart to approach an objective. At the beginning of each cycle, strategic alternatives are priori­tized mathematically using the perceptions of decision makers and other people, such as promotion managers. The outcomes are evaluated through executions and data analytics. This process continues until the objective is obtained.

Reach Goals and Build Strategies

Organizations’ goals can be achieved if recommended actions, taken from the strat­egies and tactics models developed using PI methodologies, are executed as planned. This allows companies to use perceptions and human senses along with traditional data analysis to make better decisions. The PI approach lets the insights of employees who have different opinions and multiple objectives play a role in developing and executing business strategies.

By integrating data with people’s feelings and senses for analysis, the PI methodology provides deeper insight into the organiza­tion, its customers and their purchasing habits. This information can help compa­nies form strategies based on comprehen­sive analysis rather than the experience and intuition of employees. The results will be improved customer satisfaction, increased sales and business growth.

Takehiko Ikeda, Ph.D., is the president of Business Intelligent Solutions in Tokyo.