DATA FUSION AND DATA MINING

Data Fusion And Data Mining

Data Fusion And Data Mining

Introduction

Direct marketing is not new in concept. In Venice in 1498 a book catalogue was published by Aldus Manutius - and William Lucas published a gardening catalogue in England in 1667. A variety of other mail order catalogues and clubs appeared, especially in Europe and the USA through the eighteenth and nineteenth centuries and in fact there was a significant growth in the USA in particular during the 1800s because of rising demand for goods from isolated communities which could be serviced by the improving distribution and postal systems (IDM, 1995).

Once created, the database has a variety of uses. It can clearly be used as a list from which to target customers via direct marketing activity. But in addition it can also provide a wealth of information on the market and on customers and potential customers within it. In this context, the database provides data for both planning and analysis purposes: the database can be analysed for most attractive segments, for campaign planning and predicting campaign response.(Chapman ,1999)Strategically, (Shaw and Stone, 1988a) the database can be used for:

changing the basis of competition;

strengthening customer relationships;

overcoming supplier problems;

building barriers against new entrants;

generating new products.

Data Mining

“Data mining is the use of automated facts and numbers analysis techniques to uncover before undetected connections amidst facts and numbers items. Data mining often engages the investigation of facts and numbers retained in a facts and numbers warehouse. Three of the foremost data mining methods are regression, classification and clustering.”(Klein,2004,51)

Data mining refers to the digging around in databases in a relatively unstructured way with the aim of discovering links between customer behaviour and almost any variable that might potentially be useful. There is a parallel with market research versus environmental scanning, because the former focuses on specific problems and the latter has a wider-ranging brief to identify anything in the marketing environment which might have a relevant impact on the marketing operation.(Noordwijk,1983)

Direct marketers have tried a variety of unusual or unexpected areas in which to mine. For example some have examined consumers' individual biorhythms and “star signs” as predictors of their purchasing patterns and others have linked their transactional and profile data with meteorological databases to predict, perhaps months ahead, what demand there might be for ice-cream or woolly sweaters.(Press,1992)

Another example of data mining is afforded by the linking with a GIS database to target as specifically as a newspaper round! This is usually around 150-200 households and, by linking transactional data with lifestyle, geodemographics and panel data, a very accurate picture of individual buying patterns emerges. The newspaper round - or milk round - can be used for door drops or direct mail as well as for local catchment area analysis. This has been formalised by Unigate which has advertised a door-step delivery service based on MOSAIC geodemographic profiles at local level. In addition to product delivery they offered a delivery service for samples and vouchers and a delivery and collection service for ...