Introduction
Traditionally, organizations use data tactically - to manage operations. To gain a competitive edge, strong organizations with strategic data - to expand the business to improve profitability, reduce costs and better marketing. Data Mining (DM) creates information assets that can make use of the organization of these strategic objectives.
In this article we will see some of the key questions executives have about data mining. These include:
* What is data mining?
What it's for my organization?
* How can my organization start?
Definition of data mining business
Data mining is a new component in an enterprise decision support system (DSS) architecture. It complements and interlocks with other DSS features such as query and reporting, online analytical processing (OLAP), data visualization and traditional statistical analysis. These other DSS technologies are generally retrospective. These reports, graphs and diagrams, which in the past. A user can search for specific topics white answer, such as ; How many new accounts in the Midwest were opened in the last quarter , What business had the largest change in revenue compared to the same month last year , or Did we meet our goal of a holiday than ten percent of sales?
Define data mining as data discovery and modeling of hidden patterns in large amounts of data , data mining is different from the back of technology, because it produces models - models that capture and represent the hidden patterns in data. With it, a user can discover patterns and build models automatically, without knowing exactly what they want. The models are descriptive and prospective customers. Relate why it happened and what is likely to happen next. A user can pose ; what-if ask a data mining model that can be directly in the database or warehouse. Some examples are: What is the expected lifetime value of each customer account , What customers are likely to be an open money market account , or Will this customer cancel the service if we introduce fees ?
information technology is associated with DM neural networks, genetic algorithms, fuzzy logic and rule induction. It is produced in the scope of this article for all of these technologies. Instead, we will be able to business needs and how data mining solutions to these care needs are translated into dollars.
Allocation of business needs to provide solutions and benefits
What can data mining for your company? In the introduction, which describes various strategic options for an organization to use the data to obtain an advantage: business growth, profitability, cost reduction, sales and marketing. Consider these possibilities very clearly by several examples where companies are successfully applied DM
Expand your business, Keystone Financial Services Williamsport, PA wanted to expand its customer base and offer new accounts LoanCheck. To initiate a loan, a recipient had to go on one floor and money LoanCheck Keystone. Keystone filed the $ 5000 mailed a promotion for existing customers LoanCheck.
The Keystone database tracks more than 300 characteristics of each customer. These characteristics include whether the person has open loans in the past two years, the number of active credit cards, the levels of balance on these cards, and, finally, if you reply to the offer of $ 5000 LoanCheck. Cornerstone of data mining to sift through the 300 characteristics of clients, where the most important, and build a model of the response to the offer LoanCheck. Then apply the model to obtain a list of 400,000 potential customers in a credit bureau.
Through direct mail to potential customers more value determined by the DM model, generated U.S. $ 1.6 million in additional net income of 12,000 new customers trapezoidal.
Cost reduction: Empire Blue Cross Blue Shield of New York States the largest health insurer. To compete with other health organizations must provide high quality services empire of high quality and minimize costs. Attacking costs in the form of fraud and abuse is a cornerstone of the rule s strategy, and it requires great skill and technology of sophisticated information research.
The latter includes a data mining application based on profiles of each physician in the network of Empire to demand records of patients in their database. From the profile, the application detects subtle variations in physician behavior toward their peer group. These deviations must be fraud investigators as reported index of suspicion . A doctor who performs a large number of procedures per visit, rates would be 40% more per patient, or go to many patients over the weekend will be immediately marked by the index of suspicion.
What this effort MS was again in the Empire? In the first three years, they realized the fraud and abuse savings of $ 29M, $ 36M and $ 39M each.
The improvement in sales and profitability: sales representatives with a wide range of tools to promote products to doctors. These tools include the clinical literature, product samples, dinner meetings, telephone conferences, golf outings and more. To know what actions will be most effective, with doctors is very valuable, as wrong decisions, the company can call hundreds of dollars for the sale and cost more in lost revenue.
Representatives of a major pharmaceutical company tens of thousands of sales calls. A pharmaceutical company has six months promotional activities with the sales figures in a database used to construct a predictive model for each doctor. The data mining models reveals, for example, that among the six different advertising options, only two had a significant influence on the prescribing behavior of physicians. With all the knowledge embedded in data mining models, the promotional mix for each individual doctor was to maximize ROI.
Despite this new program was expanded recently filed an early response, the drug maker more than 1.4 million U.S. dollars increased sales originally anticipated. Since this increase does not create advertising revenue expenditure is expected to raise a similar amount.
Looking back at these examples, we must ask: , Why data mining is necessary? For Keystone has to offer in response to the loan from the database of the agency credit for 400,000 new potential customers. The model predicts the response of the client features available. For Empire, the suspicion index quantifies the differences between medical practices and peer (model) behavior. The doctor was proper behavior a number of multivariate data mining product - not again be available in the database. For pharmaceutical companies, including the promotion and sale of databases, historical records of the activity. A process of automated data mining was necessary for each medical model and determine the best combination of actions to improve future sales.
Getting Started
In all cases above data mining led to significant benefits for the company. Some were the results of first-line revenue increases or expands the customer base. Others were the bottom-line improvements, cost savings and increased productivity due. The next question, of course: How can my organization get started and begin to realize the competitive advantages of the MS ?
In our experience, pilot projects are the most effective vehicles for the introduction of data mining. A pilot project is an effort to short, well-planned by an organization that in DM. Successful pilot projects will focus on a very specific business need, and include business users before and during the entire project. The duration of a typical month pilot 2:59, and usually requires 4 to 10 people part time.
The role of leadership in the pilot projects is twofold. At the beginning of the Board, in defining strategic objectives for the project. During deployment of the project before making the executive branch by controlling the measurement and interpretation of results. Lack of executive sponsorship to participate and the users of its failure are two main reasons for the post of DM initiatives or less.
By reading this article, perhaps you have developed a vision and want to go to a pressing business problem by sponsoring a pilot of the address data mining project. Twisting say the old adage, that , just because it does not mean that you should be able to. Note that you have a skills assessment will be an integral part of a pilot project should DM. The assessment takes a critical view of data and data access, human resources and skills, equipment and software. Organizations often underestimate the impact of data mining (and information technology in general) for its people, its processes, its corporate culture. The pilot project offers a relatively high reward and low-cost and low-risk opportunity to quantify the potential impact of MS.
Another obstacle to an organization is to postpone the election of mining activities to a data warehouse is built. Our experience shows that, often, the DM can and should come first. The purpose of the data warehouse is to provide users with the opportunity, customers and market trends to consider both retrospectively and prospectively. A pilot project of data mining can provide important insights into the fields and equipment, and shall be designed in store for truly valuable. In addition, cost savings or revenue generation of the MS boot funding for a data warehouse or initiatives provided.
In closing, this article has addressed the executives of the key questions about data mining - that is, what are the benefits and how to begin. Armed with this knowledge, to start with a pilot project. From there you can take to reduce the development of data mining capabilities in your organization to expand its business, increase profitability, costs and benefits of better market their products.