Data mining can play a key role in developing business strategy and is a fundamental and compelling factor for justifying, designing and building intelligent data warehouses. The transition from data mining buzz to practical implementation, however, has been slow to come although it is quickly gaining momentum. There are many reasons why it has taken time for the technology to take off. First, it is extremely powerful, and companies are reticent to adopt a technology without the core competency to utilize it. Second, there is an enormous cultural shift that comes with automating and integrating application intelligence into the decision-making process. Third, the technology has been somewhat raw and geared toward analytics researchers and experts, not necessarily toward mainstream business. In the last few years, companies have been quietly adopting data mining with success. What is the essential difference that has caused acceptance and usage of the technology? First, vendors have become proactive in teaming up with universities and consulting groups to provide the massive education required to bring business and IT up to speed. Second, vendors are tuning into the natural marriage of warehousing and mining and are developing more robust architectures that take advantage of warehousing as the enabling infrastructure for enterprise mining. Third, both vendors and consulting groups have been packaging data mining into software and processes, respectively, enshrouding the raw horsepower with sleek, application-oriented, user-friendly systems. Fourth, the competitive urgency is overshadowing the tendency to watch and wait.

With company mission statements focused on "customer- centric" objectives, it is becoming critical to organize and manage processes from the customer perspective. Data mining is a powerful ally in aligning business strategies to acquiring, retaining and growing customers. The best asset available to organizations in realigning business processes to the customer perspective is internal historical data integrated with external demographic and psychographic data. Building warehouses with mining at the heart brings a streamlined focus to design and content and delivers immediate exploitative capability to warehouse users.

A customer-centric warehouse can be mined and analyzed by taking information about consumers and customers that is derived from POS transactions and primary research. For example:

  • Information about the public at large is mined to create segments around which businesses position their products and services.
  • Key buying behavior and physical/lifestyle attributes of existing customers are used to learn a more effective means of targeting and acquiring new customers.
  • The same information is used to spot potential customer defectors before they leave.
  • Individual transactions are analyzed to spot product assortments, giving better understanding of what products customers buy in groups.
  • The same information is used to design product bundles and cross-sell strategies that will ensure customer loyalty through time.
  • Analysis is performed to understand profitability segments so that promotions and services can be targeted to those that are most profitable.

Further analysis feeds everything from promotion planning to shelf-space management. Data mining supplies information for developing strategies that foster customer growth and satisfaction and discovers those factors that predict bottom-line profitability.
In today's world, allocating capital and expense is grounded in experience and an in-depth knowledge of functional operations. Processes and methods for forecasting and allocation have been proven time and again. For the most part, companies make these decisions from a well-informed basis and have a solid analytical footing to judge where the investments should be allocated. The marketing function, however, is often an exception.

There is a culture clash that is taking place between traditional marketers and database marketers that centers around the usage of data mining. Whereas traditional marketers rely on mass communication campaigns geared at acquiring new customers, database marketers have imbedded data mining techniques into their analysis processes, squeezing the most out of the current customer base. Database marketers clearly have more proof-of-performance numbers on their side. They are proficient at using data mining to make management comfortable by providing direct financial and economic return results from their promotional campaigns. The end result is that decision making within the marketing discipline is split between informed and uninformed resource allocation decisions.

The irony is that the answer to providing the right metrics to judge performance of all marketing functions lies within the existing marketing database. It's just that many companies never bother to look. Database marketers are accustomed to embedding data mining tools directly into their business processes, yet the remainder of the organization chooses not to. A clear competitive advantage exists for companies willing to make this change.

Integrating Data Mining into the Decision Making Process

It should be clear by now that data mining is a business function and can provide a strategic advantage in developing, defining and deploying competitive business strategies. There are two areas to consider in successfully ushering data mining into an information environment: skill sets and technology.

Skill sets will vary by the data mining stakeholders in your organization. The skill sets for each of the stakeholders include:

Stakeholder Skill set
Miner Analytics, model building, statistics, neural net development, research
Domain Expert Intensive business and data knowledge, experience, decision maker
Business User Understand business and data, decision maker, user of mining results
IT Support analytic environment, data model for new DM components, integrate DM (tools, processes, results, models) into DW
Technology integration points include communication links between data mining software and both data and application domains. Data links include sourcing, transformation and loading of input and output variables and result sets. Application links include the ability to invoke data mining tools from the DW environment, access to mining results and visualizations, and ability to invoke analytic models for prediction and description from operational and decision support systems, both back end and front end.

Data mining has evolved from manual statistical methods to desktop mining to enterprise mining. With appropriate skill sets, the right team, a warehousing infrastructure and data mining tools, companies can transition into agile competitors who maneuver quickly with the global demands of the marketplace.

Register or login for access to this item and much more

All Information Management content is archived after seven days.

Community members receive:
  • All recent and archived articles
  • Conference offers and updates
  • A full menu of enewsletter options
  • Web seminars, white papers, ebooks

Don't have an account? Register for Free Unlimited Access