You are here: Data Mining > Turning Organization Data Into Information

TURNING ORGANIZATIONAL DATA INTO INFORMATION

A franchise system generates volumes data every day from various sub-stages of the CSLC shown in Table 1.  Consider the “Managing the Franchise System” sub-stage as an example.  At the end of each business transaction, billing, customer tracking, inventory control, and labor all generate enormous amount of operational data.  At the end of the day, a sale report is sent through the telecommunication system to the franchiser headquarters to summarize the daily business transactions such as total sale, total cost of raw materials, and total cost of labor.  If the report is not received after a pre-determined time, a message is triggered to request prompt actions by the franchisee.  Once the daily sale reports are received from all the business outlets, they are converted into information using various analytical methods.  These statistical data analyses also help generate many business intelligence reports.  For example, a business outlet will typically receive its performance ranking report with respect to the franchise system along with the top 10 business outlets having the best sale reports.  A rewarding system can be built into the information generation process, e.g., the owner of the winning franchisee outlet may receive a free trip to Hawaii if he/she has been among the top 10 lists for a number of consecutive time periods.

The information generation process can be presented in four levels as is shown in Figure 1.  The process is adapted from Inmon’s (1996) four generic levels in the architected environment of data warehouse: (1) the data collection level, holding data collected from the CSLC model discussed in the previous section: online transaction processing  (OLTP) operational data, external data, and legacy data; (2) the reconciled data level, holding data warehouse data that are subject-oriented, integrated, time-variant, and nonvolatile (Inmon, 1996); (3) the derived data level, containing several data marts (e.g., departmental, regional, loyal customers, partners, and suppliers) derived from the data warehouse based on various customer-centered market segmentation; and (4) the analytical reporting level, producing various performance reports (e.g., business outlet periodical summary, financial, scorecards, and mysterious shopping) for the decision makers using the decision support systems (DSS) for their decision making.  To move from the data collection level to the reconciled data level, data integration is needed.  It is a very time consuming process that involves the activities such as cleansing, extracting, filtering, conditioning, scrubbing, and loading.  To move from the reconciled data level to the derived data level, data transformation is needed which involves the activities such as replication, propagation, summary, aggregate, and metadata. To move from the derived data level to the analytical reporting level, data analysis is needed which involves two major activities online analytical processing (OLAP) and data mining.

Typical OLAP analysis composes of pre-defined multi-dimensional queries (Kimball, 1996).  Shown below are some examples:

• Show the gross margin by product category and by franchise outlets from Thanksgiving to Christmas in the last five years.
• Which franchise outlets are increasing in sales and which are decreasing?
• Which kinds of customers place the same orders on a regular basis at certain franchise outlets?
• How many franchisees did we lose during the last quarter of 2001, compared to 2000, 1999, and 1998?
• How many new franchisees did we recruit during the last quarter of 2001, compared to 2000, 1999, and 1998?

Other OLAP activities also include spreadsheet analysis, data visualization, and a variety of statistical data modelling methods such as regression analysis, correlation analysis, time series analysis, forecasting, Pareto analysis, and quality assurance.  Data mining, on the other hand, is used to identify hidden patterns of the data residing in the data marts, which are derived based on the criteria such as frequent customers and products of high profits.  Typical data mining modelling techniques for analytical reporting include decision tree analysis, cluster analysis, market segmentation analysis, cross-sell analysis, association analysis, neural network, and genetic algorithm.  Table 2, adapted from Delmater and Hancock (2001), shows that data mining techniques can be used to help serve franchisees’ customers at the different stages of the CSLC model.

Due to the popularity of the Internet, franchising companies have adopted the Web-based reporting systems for their decision-making activities (Chen, Justis, and Watson, 2000).  Most recently, however, Mobile-based reporting systems, integrated with other DNS systems (e.g., email and groupware) for Empowerment and Collaboration Phase, are gaining the popularity.   By providing the “mobile” managers (e.g., field sales representatives, mysterious shopping auditors, and traveling senior executives) with hand-held, wireless information technologies, they will be able to access the critical and detailed information residing in the corporate Intranet in real time.  In some franchise industries, the use of Mobile-based reporting DNS is even extended from the managers to the front-line workers who are empowered to make timely and on-the-spot decisions when serving the customers.  For example, in the Hospitality industry some leading franchises are pushing the timely information, e.g., optimized room rates, into the hands of front-desk people to attract customers in the very competitive market.  

As we can see now, the Four Levels architecture depicted in Figure 1 shows how the franchise organizational information is deciphered from data analyses, i.e., OLAP analyses and data mining.  The valuable information contained in the business intelligence reports becomes the foundation upon which the working knowledge of the franchise system may be built.  Thus, a well-designed DNS in the Business Intelligence and Knowledge Management Phase shall empower the franchiser and the franchisees (e.g., though the use of mobile and wireless information technologies) to turn the valuable information into actions in real time.