What is data mining and predictive analytics used for?In order for you to better understand what data mining is about and why it is a valuable thing, we describe a number of applications where data mining and predictive analytics is a good option to consider. Please feel free to contact us with any type of comments or questions.
Direct marketingYou probably have heard the marketing manager phrase “I know that half of my marketing budget is wasted, the only question is what half?”
The challenge of marketing is that while there are constantly more and more competing offers, the number of channels (ways of communication) available to communicate with the buyer also increases. In addition to traditional direct marketing means such as direct mailing, advertising in newspapers, TV and other media, new means of communication such as the Internet, not only has introduced a large number of new channels for direct marketing, it has also brought measurability of the response to a whole new level. Today, there is a large amount of data being generated not only in internal customer data bases, but also related to the response of the audience to marketing campaigns. This is where predictive data mining comes in. By applying predictive data mining to historical data, such as customer response for the various channels, demographic, geographic, sales history etc, it is possible to significantly improve the odds of directing a campaign towards the right audience.
By successfully applying predictive data mining, you not only will be able to target the right audience, thereby increasing return on invested marketing money. In addition to this, you will also get to know them better, and by adapting the message to the preferences of your audience you will be able to communicate more effectively.CollectionAll companies with a large customer base have a number of customers who do not pay their dues on time. Collecting these payments from the debtors requires a great deal of resources, and a large proportion of this work is wasted on customers that are difficult or impossible to recover. By applying predictive data mining to historical customer debt data, the collection procedure can be optimized by identifying the debtors most likely to pay and finding the most effective contact methods or legal actions for each debtor.
By successfully applying predictive data mining within collection, you will recover more money while reducing collection costs.
Scientific applicationsIn pharmaceutical companies, chemistry is one of the most resource intensive areas within the research and development (R&D) activities. The whole purpose of the pharmaceutical company’s R&D is to produce new chemical entities (NCE) that will make it all the way through clinical trials to the market as new drugs. The search for new chemical compounds is in essence a trial-and-error process. The job of the R&D chemist is to synthesise (produce) new compounds for testing in the laboratory.
One such chemist may spend up to three weeks for making just one such compound.
Despite the expensive production of new compounds, pharmaceutical companies test very many compounds in their R&D activities and stores all the results in large databases (there is also a fairly large industry selling chemical compounds of great diversity). By applying predictive data mining to such historical laboratory test data, it is possible to reasonably predict the outcome of the laboratory tests without having to synthesise the compounds. This means that the chemist can find out the most likely properties of the alternative compounds and choose to work on the most promising ones, before spending the next three weeks in the laboratory, thereby increasing the quality of the resulting compounds.
Recommendation systemsAll companies want to sell more to existing customers. This is often the most effective way of increasing the profitability. For companies that sell many different products to large customer bases and that keep records of sales transactions for their customers, it is possible to apply predictive data mining to identify sales opportunities as products likely to appeal to a particular customer who has not yet bought them. This type of application is also commonly referred to as cross-selling, and some of the most notable examples of companies using it are Amazon.com where you will get relevant books recommended, and the DVD rental site Netflix, who now even has arranged a $1Million prize money competition for the best improvement of their recommendation system.
Other applications Other applications of predictive data mining include fraud detection (e.g. within credit card transactions, taxation, telephony, and insurance industry) and risk management (e.g. for determining insurance policy rates or managing credit applications).
|
|
|
|
Rule Discovery System™
