Data mining has delivered profound results in various industries around the world by providing executives with business intelligence to assist in better decision-making, which consequently improves corporate performance. However, as successful as these analytical technologies have been, there is still an educational gap between what many executives believe data mining can deliver and the benefits they can produce on the shop floor.
One result of this lack of understanding is that many businesses run inferior data mining tools unable to meet their needs, or spend fortunes on sophisticated tools, while only using a fraction of their capabilities.
These failings can be spotted in many industries, but are more noticeable in B2B markets, where the common belief is that data mining is not a technology that can be applied in this complex segment.
"The technology to deliver on the promises of improved business intelligence is available in many top-class data mining products, but there is a general lack of implementation skills in SA to enable effective use of these tools," says Mervyn Mooi, director of Knowledge Integration Dynamics (KID). "This especially applies in B2B markets, where the prevailing opinion is that data mining can not deliver actionable intelligence because of the unique business models and processes used.
"This is an erroneous view. The benefits of data mining are as attainable in B2B organisations as any other if the appropriate skills and an in-depth understanding of the business are present. The problem is that too few business leaders understand the concept of data mining well enough to know where it can be applied effectively. Additionally, there are also far too many supposed experts unable to use data mining tools properly and gain the maximum effect from them."
In dealing with the demands of the B2B market, Mooi has found five data mining myths often accepted by business leaders, to the detriment of their companies.
1. Data mining for B2B is difficult: As stated above, there is no definitive market for data mining tools. If B2B companies use inferior methodologies without due consideration to the analysis behind the implementation, they will waste their money. If the implementation is planned and executed efficiently, data mining applications will deliver value irrespective of the industry.
In addition, B2B organisations should realise that what counts in successful data mining is the data and not the type of business model. Data mining can be successful on even a limited dataset, as long as the information is of high quality.
2. Huge volumes of data are required: Mooi says the success of mining tools is not dependent on the amount of data to be analysed. Even companies with a limited dataset can use mining tools with great success. Of course, the ideal when starting with a limited amount of data is to increase the information repository over time to improve the analysis capabilities as the business grows.
"The more data one uses in an analysis, the better predictive models can be designed," adds Mooi.
3. New models are needed for B2B mining: Mooi acknowledges that B2B business practices can be complex, however, he also believes that the greatest determinant of whether data mining tools are effective in B2B environments or not, will be the skills and understanding of the business technical staff have.
4. Only experts can perform data mining: "This statement is both true and false," states Mooi. "Expert skills are needed to construct the data models applicable to each company`s business practices, but any manager can run queries that use this complex architecture. Front-end tools have evolved to offer simple point-and-click functionality almost anyone can learn to use."
5. Data mining costs are too high: "Like most other IT tools, the costs depend on what the customer wants," observes Mooi. "A small investment can set companies on the right track to benefit from the intelligence data mining delivers."
To keep costs under control, he suggests making use of simple models and small datasets to start with and then expanding them to meet further needs. Management can then immediately identify the returns data mining delivers without the stress of spending too much before seeing value.
The B2B industry does not need specific data mining techniques unique from the rest of the world. B2B requirements may be somewhat different, but by using standard models and implementers with the appropriate expertise and experience, intelligence can easily be extracted and applied in these companies without excessive costs or having to remodel the organisation. In addition, methodologies aside, the foundation for any success in data mining, irrespective of the industry, is the data itself. Without reliable data, effective mining and analysis in any industry is simply not possible.
Share
Knowledge Integration Dynamics (KID) was formed in 1999 to address a clearly identified need in the South African corporate market for high-performance business intelligence solutions. The company aims to develop customers` business intelligence (BI) capability by providing application solutions, software components and comprehensive services to enable better business decisions. The company`s skills set spans multiple technologies while maintaining a focus on the business issues and deliverables, ensuring that the best technologies are deployed to support specific applications. KID`s skills embrace the full BI sphere, including data warehousing, data mining, business intelligence applications and information management. The company provides expert consulting in information management, including strategy development, capability development and realisation programmes. For further information, visit www.kid.co.za.
Editorial contacts