SQL Server 2005: Bringing Data Mining to the Mainstream
Data mining has become common property of highly competitive companies
If sales are not progressing as expected the decisive point is to identify the problem area. Is it the sales force, the suppliers, or maybe an entirely different area? To make sense out of and establish causalities from raw data to ascertain why things have developed the way they have is precisely what data mining is about. In December 2004, market research company IDC pointed to data mining as one of the essential tools for supporting Business Intelligence in a company.
2005 will see the launch of Microsoft SQL Server 2005 which will include data mining. TDC Solutions, Denmark’s largest telecom operator, was one of the first companies to be introduced to data mining in SQL Server 2005. In this article, the head of analysis explains how TDC is able to use data mining. Software company SPSS, which develops data mining tools, explains what impact SQL Server 2005 as a platform with data mining will have on future business.
From backward reporting to forward reporting
"Our goal at TDC Solutions is to move from backward reporting to forward reporting, and data mining will be part of supporting this goal. Data mining will become the analyst's right hand," says head of analysis at TDC Solutions, Peter Møllebjerg Andersen. TDC Solutions is part of a network for early adapters of SQL Server 2005, through which the company has just completed a course in data mining.
Jeppe Salmonsen, Account Manager for business sales, adds, "Data mining will for instance help us understand customer churn. When and why do customers leave TDC? Through SQL Server 2005 data mining, we may be able to pinpoint potential churners, look into why they leave, and what factors are involved. We may be able to map what customer segments are especially at risk. Thus, data mining will give us an understanding of how to target the market."
Today, Business Intelligence at TDC Solutions does not include data mining, since data mining is included in other systems. With the inclusion of SQL Server 2005, data mining becomes a part of TDC Solutions' Business Intelligence. The analysis department sees it as a welcome challenge for developers and analysts to work even closer together to build the best financial models, which will ensure optimum profits for the business. Both professional groups acknowledge and comprehend the potential of the product. Ingolf Benzen, analyst at TDC Solutions, who participated in the data-mining course, weighs it up as follows:
"The first hand impression from the course is that it is not possible to reach a deeper understanding of the product during a weekend. The environment is radically different from the one we usually work in. The change from SQL Server 2000 to SQL Server 2005 is a big one. There have been a huge number of upgrades. The general improvements of the 2005 version are significant. It will require a dedicated effort, but it will be worth every penny. I believe that data mining will be able to point out correlations, which will widen our horizon."
Peter Møllebjerg Andersen, head of analysis, concludes: "Yes, the project looks very promising. In reality it will probably be our own ideas for financial models that will determine how far we can go."
Stepping stone to the future
"For years data mining has been "nice to have", but not necessarily "need to have" – this has now changed. In our experience it is now much more the IT department and no longer just the marketing department that invests in data mining, since it is now considered a business necessity. As a part of Business Intelligence, data mining has become a business critical tool, which companies need to have in order to be competitive," says Morten Hansen, Marketing Manager of software company SPSS. Since 1993, the company has marketed the data-mining tool Clementine. It is now the market leader in data mining. Through Clementine, it is possible to work with the existing data mining models in SQL Server 2005. In future, the integration between future versions of Clementine and SQL Server will be even greater.
"In terms of performance SQL Server 2005 already surpasses or matches all other databases and data warehouses. It is a very strong platform, which is ideal for data mining. SPSS delivers a user-friendly data-mining tool and a number of final business solutions for optimising sales and marketing. For instance we were able to find 80% of a telecom operator's churners with our solutions.
"Data mining is also a stepping stone to the future, since it is the first step on the road to predictive analytics that combines data mining with decision optimisation. Predictive analytics gives company management control over daily decisions made in the company by presenting recommendations and optimising and automating decisions that will ensure the most profitable customer-directed marketing and handling of customers. A company will, however, not really gain from predictive analytics before it has a strong platform like SQL Server 2005 to collect and give business users easy and fast access to data," concludes Morten Hansen.
Fact box: Clarification of concepts
Business Intelligence is the process of monitoring, collecting, analysing, and presenting knowledge, which until now has been hidden in the company's vast amounts of data. Through Business Intelligence, the company may obtain an updated view of market development and its own processes – also across departments.
Data mining is a part of Business Intelligence. Data mining finds the causality of customers' actions based on the analysis of data, and this knowledge can then be used to target actions directed at the client and obtain better results.
In March 2005, American experts ran a course on SQL Server 2005 data mining for 23 Danish partners and customers. Microsoft Denmark is a special focus area within Business Intelligence and can therefore carry out activities at the very forefront of the latest Microsoft technology in the area.
Predictive Analytics is often differentiated from Data Mining by specifically attempting to predict the future on the basis of the present situation and trends. SQL Server 2005 Data Mining also provides abilities described by Predictive Analytics.
SQL Server 2005
During the summer of 2005, Microsoft will introduce SQL Server 2005, which is a major improvement on SQL Server 2000. SQL Server 2005 can for instance handle far greater amounts of data, which is especially important for large companies. On top of that, SQL Server 2005 is a complete Business Intelligence platform comprising all elements of data handling – including data mining.