Marketing Analysis:Targeting, Measurement and Analysis for Marketing [2]
论文作者:留学生论文论文属性:案例分析 Case Study登出时间:2011-01-20编辑:anterran点击率:15671
论文字数:5537论文编号:org201101200957272304语种:英语 English地区:美国价格:免费论文
附件:20110120095727228.pdf
关键词:CRM productTargetingMeasurement and AnalysisMarketing
sis for Marketing (2007) 15, 137 – 145. doi: 10.1057/palgrave.jt.5750045
van den Berg and Breur
138 Journal of Targeting, Measurement and Analysis for Marketing Vol. 15, 3, 137–145 © 2007 Palgrave Macmillan Ltd 0967-3237 $30.00rarely in place to begin with. Furthermore, it mayeasily be omitted in the short run, with poorresults for the long haul. Such short-sightednessposes a considerable risk as the use of datamining technology is not quite fully matureand established, yet.The hope for these grandiose solutions hasbeen fuelled, not in the least part, by vendorsof end-to-end CRM solutions. In one case, avendor ’ s claim suggested that implementation oftheir tool had caused response percentages to risefrom 1 to 30 per cent. And this was at a companywith ten years experience in direct marketing,notable for its sophisticated use of data miningmodels! Clearly such claims should foremost castdoubt on the expertise of vendors making suchunwarranted claims. But management is easilyimpressed by these claims, and expectations aboutthe possibilities of data mining do need to bemanaged. Is it really true that machines canreplace humans in this process? What are theadvantages and disadvantages of automating themodel building process? Is the current state ofaffairs advanced enough for this, yet? Which requirementsneed to be met to make this possible?The structure in which these thoughts arepresented in this paper will now be outlined.After discussing the advantages and disadvantagesof integrated solutions, we will have a look atautomated model building. Then we discussinteractive tree building, as an example ofinteractive model building, and explain why thisleads to more knowledge and better models. Ofcourse ‘ better ’ needs to be understood in relationto the application and the business context. In thenext paper, the authors go on to explain howinteractive tree building is done in practice, withdetailed practical
guidelines. In the conclusion, weexplain under what conditions one can typicallyexpect our proposed approach to be most useful.INTEGRATED AND AUTOMATEDDATA MININGIntegrated data mining toolsClearly, there are great advantages when analyticdata mining solutions can be seamlessly embeddedinto an operational environment. On the inputside of the data mining process, mapping of inputvariables tends to be a cumbersome and errorprone process. Often this pre-processing requiresextensive manual programming, making this anunreliable and brittle part of the
architecture.On the output side of the process, error-freedeployment is a challenge, for similar reasons.These advantages have greatly and legitimatelycontributed to the popularity of integratedCRM solutions.But there are disadvantages, too. When inputdata are fed into an integrated, automatic modelbuilding environment, one may be tempted to‘ forget ’ that scores on a variable are but anabstraction of reality — this link is easilyforgotten. Profound insight into the way actualcustomer behaviour translates into a mapping onan input variable can greatly enhance insight intopotential leverage points for infl uencing customerbehaviour. 1 This insight itself is not lost throughintegration, but for a lack of manual data captureand pre-processing, the quest for profound insightinto customer behaviour must be scheduledseparately. The workfl ow process in an integrateddata mining solution does not call for this anymore.Clearly there are big advantages to be gainedfrom integrated data mining workbench tool
本论文由英语论文网提供整理,提供论文代写,英语论文代写,代写论文,代写英语论文,代写留学生论文,代写英文论文,留学生论文代写相关核心关键词搜索。