MIS - Group 2 - Data Mining · DDaattaa MMiinniinngg MMIISS 330044 –– FFaallll 22001100 –– GGrroouupp ##22 Eduardo

DDDaaatttaaaMMMiiinnniiinnngggMMMIIISSS333000444–––FFFaaallllll222000111000–––GGGrrrooouuuppp###222

EduardoAsher,CharlesGeigner,JenniferHeth,JonathanMonjazi,ChristianSanchez

1

TableofContents

Overview.....................................................................................................................................2History.........................................................................................................................................4E­CommerceBusiness............................................................................................................7Phorm..........................................................................................................................................8GoogleAnalytics.................................................................................................................... 11Testimonials........................................................................................................................... 15CostcoTravel ................................................................................................................................... 15Vueling ............................................................................................................................................... 17DiscountTire ................................................................................................................................... 18AmericanCancerSociety ............................................................................................................. 19

FutureofDataMining ......................................................................................................... 20WorksCited ............................................................................................................................ 22

2

OverviewThe field of E‐commerce has transformed the way companies communicate with

theiremployeesandcustomers.Whatwasonceconsideredofficesuppliessuchasa

computer has transformed into a network of interdependent devices transferring

informationfromcustomerstobusinessestomanufacturerstosupplierstobrokers

toclients.

Inordertogainandsustainacompetitiveadvantage,acompanyrequiresasimple

and easy way to deliver and access information. The E‐commerce model has

changed dramatically and in our essay we will explore the application of data

mininginthemoderne‐commercebusiness.

Datamining is part of business intelligence. Business intelligence uses computer‐

basedmethodstofindandanalyzebusinessdata,suchsalesrevenueorunitsales.

There are different aspects of business intelligence, which includes technologies

that facilitatemethodssuchasanalytics,dataminingandpredictiveanalysis.Data

mining is the process of extracting data and using statistical techniques to find

relationshipsandpatterns,whichbusinessescanusetoincreasetheirbottomline.It

has been customary for businesses to contact prospective clients through such

mediums as: standardmail, e‐mail, or call centers. Thismethod has proven to be

ratherinefficientaswellascostlytomanycompaniesduetothefactthattheylack

theproperinformationtheyneedtoassessclientneeds.

3

Datamining’sbeginningisrootedinthreedifferentdisciplines.Theearliestofthese

three disciplines is classical statistics. Statistics is the main reason we have data

mining,sincestatisticsisthefoundationofthetechnologiesonwhichdataminingis

built.Instatistics,theconceptsofregressionanalysis,standarddeviation,variance,

and confidence intervals to name a few, are all used to study data and the

relationships data can represent. The second earliest discipline to influence data

mining isartificial intelligence.Artificial intelligence tries toapplyhuman thought

processestostatisticalproblems.Thethirdandfinaldisciplineismachinelearning.

Machine learning attempts to make computers “learn” about the data they are

mining by developing certain programs. These programsmake different bases on

thedifferentqualitiesrepresentedbythedata.Dataminingisbecomingincreasingly

popularinscienceandbusiness.Thesedisciplinesneedtoanalyzelargeamountsof

datasothattheymaydiscovertrendstheyotherwisecouldnotfind.

Sincethetopicofdataminingisquitelargeandintricate,theprimaryfocusofthis

paperwillbe Internetdataminingapplications. Inbusinesses,almosteverymajor

retailerintheUnitedStatesusesdatamining.Wal‐Martusesittostockitsinventory

and toanticipatecustomerneedsandAT&Tuses it to setupplans forcustomers.

Wewill show how companies such as Google and others can allow companies to

focus their energy on prospective clients increasing the likelihood in response to

salespitchesandproductmatchesonthetypesofproducttheysearch,clickon,and

buy.Dataminingcanbehelpful inmanywaysand isa toolheavilyusedbymany

Marketing departments. The information that data mining provides can also be

appliedtootherdepartmentssuchasHumanResourceswithregardstoidentifying

4

prospectiveemployees thatwillprovidevalue to firms.The sophistication indata

mining applications has increased dramatically allowing companies to broadly

targetcustomerswithproductstheywouldliketobuyaswellasitemsrelatedtothe

customer’s tastes and preferences. Businesses pursue data mining, because they

expecttoreceiveareturnoninvestment.

In order to ensure the proper application of datamining, the proper predicative

models must be used. The process of applying the appropriate model is where

business acumen comes into play. We will explore how these data mining

applicationsareusedE‐commercebusinessesandhowtheyaffecttheirbottomline,

thetypesofdataminingapplicationsoffered,andempiricalexamplesofcompanies

such asDiscountTireCompany andCostco that have it in use.Keep inmind that

dataminingisanewfield,andmanycompanieshavetakendifferentapproachesin

creatingdataminingtoolsaswellasusingthem.

HistoryData mining has gone a long way since its beginnings in the late 1960's. Before

softwarewastheprimitive fileprocessingsystems, fromhierarchicalandnetwork

databasestothedevelopmentofrationaldatabasesystems,datamodelingtools,ad

indexing, and data organization techniques. The first data mining software or

packages where based on simple algorithms. With the introduction of Customer

RelationshipManagement(CRM)software,datamininggrewinpopularityanduse

among companies, especially those interested in E‐commerce (Mailvaganam). In

5

addition, users have developed flexible data access through query language, user

interfaces, optimized query processing, and transaction management. (Han &

Kamber,2001)

Themainreasondatamininggainedsomuchattentionintheworldmayhavebeen

duetothevastamountofdataandtheneedtoturnthisdataintoinformationthat

can be used. The information obtained from data mining can be used for many

applications such as “business management, production control, and market

analysis,toengineeringandscienceexploration.”(Han&Kamber,2001)

Inordertounderstanddatamininganditsmotivationaldrivingforceitisimportant

to understand that data mining evolved naturally through the process of

informationtechnology.“Anevolutionarypathhasbeenwitnessedinthedatabase

industry in the development of the following functionalities: data collection and

database creation, data managemet (including data storage and retrieval, and

database trasaction processing), and data analysis and understanding (involving

datawarehousinganddatamining)”(Han&Kamber,2001).

New advances in technology in the mid‐1980s gave way to powerful database

systems.These systemsuse advanceddatamodeling such as: extended‐relational,

object‐oriented,object‐rational,anddeductivemodels;andthemostwidelyknown

global database systems such as theWoldWideWeb (WWW)whichplays a vital

roleintheinformationindustry.

Due to all these innovations in the last fewdecades, databasesbegan to establish

their architecture, For example the data warehouse is “a repository of multiple

6

heterogeneous data sources, organized under a unified schema at a single site in

order to facilitate managment decision making” (Han & Kamber, 2001). Data

warehouse technology includes such systems as On‐Line Analytical Processing

(OLAP), amultidimensional tool that allows to extractmultiple sets of data from

differentpointsofview(Wikipediacontributors,2010).

Morerecentlystatisticalcollinearityhasestablishedtheimportanceofdatamining

anditsinterelationshipwithmostaspectsofbusinessmangement.Takingsubsetsof

datacouldnotdetermineadomainpatternasawhole,newdevelopmentinthefield

of human‐like decisionmodels have been employed alongside standard statistical

models creating a Choice Modeling system that experimentally mimics human

behaivior.Thissystemsallowforlargequanitiesofdataandcanresultinthousands

ofpossibleoutcomes.Sincethelate1990’sdatamininghasonseveralocasionsbeen

theobjectofbusinessandpoliticalpressureinattempttodefinestandardsfordata

miningsuchastheCrossIndustryStandardProcessforDataMining(CRISP‐DM1.0)

and more recently the 2004Java Data Miningstandard (JDM 1.0) thus creating

standard approaches that data miners could use to tackle problems that have

alreadybeensolvedbyotherusers.

Dataminingtoday isused ineveryaspectofour life, fromourmedicalrecords, to

our grocery coupons, andmail admarketing,whichwe usually throw away.Data

mininghasbecomeavitalroleineverybusinessandgovernment.Dataminingwill

continue to change through natural evolution of information technology. (Han &

Kamber,2001)

7

E‐CommerceBusinessData mining in the business world has created what the Stanford University

ProfessorAndreasS.WeigendPh.D.hascalled the “SocialDataRevolution (SDR)”.

Asbusinesscontinuetheirsearch foreffectivetools theyhavebecomeengulfed in

the data mining field trying to find a way to leverage data to create innovative

productsandbusinessmodels(WeigandPh.D.,2009).

InaretailmodelforE‐business,itisimportanttounderstandthecomponentsofthe

model.There are typical browser software such asWindows InternetExploreror

GoogleChrome,aswellasadatabasethatexists intheback‐endthatsupportsthe

architectureofthefront‐enddatabase.Insomeinstancesamiddlelayer(software)

may serve to convert the architecture of different browsers so that the back‐end

databasecanunderstandthedata.

Once the user or customer inputs information into the request form about the

information he/shemaywant to purchase the information is passed through the

middle layer into theOLTP systems and records it. Once the data is storedmany

algorithmscanbeappliedtothedatatotrytodeterminetheconsumerpotentialby

structure. (A Breif History of Data Mining, 2006). The data obtained through

customerinputhelpsthecompanydeterminewhethertheirwebsiteistargetingthe

right cluster of people, or if the site design is user friendly, attractive, and

professional. The basic questions is,will the hardwork payoff? This information

canalsodeterminewhetherothermarketingstrategiesofthecompanyareresulting

ininterestintheirgoodsorservices(Markov&Larose,2007).

8

AsubscriptionfeaturehasbecomecommonintheE‐businessmodel,becauseallthis

datathathasbeenminedandclusteredshouldbeeasilyaccessiblebythecompany

andalsoby the customerwhomhasmade inquires and/orpurchases in thepast.

This historical access to data makes the purpose of data mining relevant (Goel,

2006).

PhormPhormisdataminingcompanywhousestheircollecteddatatobenefitconsumers

and businesses. Phorm has offices located in “London, New York, Sao Paulo, and

Seoul (and is knownas) aDelaware,US incorporated company, publicly listed on

the London Stock Exchange’s Alternative Investment Market since 2004”

(Phorm.com).Phormofferstwodifferentplatformsfor itsdatamining.Thefirst is

known as the Open Internet Exchange (OIX) and the second is known as

PhormDiscover.

TheOpenInternetExchangeisusedforadvertisersandagencieswishingtobroaden

their reach and depth of acquiring new customers. According to the company’s

website, “the OIX connects buyers and sellers of online advertising media, with

pricesestablishedinacompetitive,auction‐basedmarketplace”.Alsoonthewebsite

isatutorialonhowtheOpenInternetExchangeactuallyworks.TheOIXworkswith

anonymous interest matching and relevant ad serving. Any website that is

participatingwithOIXwillplaceatagontheir“advertisinginventory”ofwhichads

willbeplacedtargetedforusersnotpages.Thoserelevantadswillbeauctionedoff

9

and the highest bidderwill win the advertising slot, gaining a relevant, potential

new customer. In comparison to other data mining companies in the industry,

Phorm’sOpenInternetExchangedoesnotstorenorretainpersonaldata,doesnot

have the possibility of accidental ormalicious disclosure, and offers the ability to

undopreviousparticipationandopt‐outoftheprogramatanytime(Phorm.com).

Phorm’s PhormDiscover is the consumer‐based product. According to the

company’swebsite,PhormDiscoverisa“freeserviceofferedbyyourISP,whichwill

bring a personalized and safer Internet experience”. Meaning that ISPs that are

partneredwithPhormwilloffer theircustomersanopt‐in toPhormDiscoverwith

thepromiseofmorerelevantadvertisingandmorerelevantinformationregarding

that person’s interests. If opted into, PhormDiscoverwill act as a security system

warning about potential fraudulent sites aswell as protecting from phishing and

malware.

AlloverPhorm’swebsiteitbecomesevidentthatthecompanybelievesitselftobe:

innovative, privacy‐concerned, and confidential. They say that have “adopted a

privacybydesignapproach”forseveraldifferentreasons.Thefirstbeingthatthey

giveconsumersthechoicetoopt‐inandoutofthesystemattheirdiscretion.They

havegiventhesystemtheabilitytobeshutdownat theuser level. Theyarealso

said to not store: personal data, browsing history, nor IP addresses, therefore

making the users anonymous. Their concern for privacy is what makes them

innovative as well, because users are assigned random numbers. Part of their

innovationcomesfromtheirownmarketingoftheirproduct.Theirwebsiteseems

10

tobeprimarilyfocusedontheconsumerinsteadoftheadvertisersorbusinesses.In

fact, their website offers no information regarding how much it would cost to

participateasanadvertisernordoes inoffertestimonials frombusinessthathave

orarecurrentlyusingPhormastheirprimarytechnologyfordatamining.

Phormhasdevelopedaninterestingrepertoirewiththepressandifonefollowsthe

progressiononPhorm in thenews itbecomesclear thatPhorm ison itsway toa

downward spiral. The progression of Phorm follows the path of anticipation,

hesitation, and finally withdrawal. In March 2008, theNew York Times wrote an

article stating the Phorm was “boasting that it will collect the most complete

informationofall”andthat“BT,CarphoneWarehouse,andVirginMediahavesigned

toworkwithPhorm”.InApril2008,theNewYorkTimeswroteanotherarticleabout

Phorm.This time theArticlewas titled, “Phorm’sAll‐SeeingParasiteCookie”. The

articleclaimsthatPhorm’scookie“couldtrackeverypageyouvisitontheInternet”

by piggybacking on other website’s cookies without permission. The anticipation

portionofPhorm’sarrivalisnotnecessarilyapositiveonefromthegetgo,butthese

articles are being written at about the same time as the article written on

MarketingCharts.com.Thisarticlestates,“threeoutoffourconsumersarewillingto

provide somemeaningful amountofpersonal information inexchange for amore

personalizedrelevantshoppingexperience”.Thisshowsthatdependingonwhowas

being talked to, their feelings towardsdataminingvaried. InMarch2009another

articlewaswrittenaboutPhorm,thistimeonSeobook.Thisarticlestatedthatthe

“behavioral ad targeting” that Phorm uses to gain information gave “commercial

incentiveforcompaniesorindividualstomisuse(theinformation)”. Also, in2009

11

theNewYorkTimes yet againwrote an article regardingPhorm.This time itwas

aboutBT choosing not to “adopt Phorm” because of “budget limitations” and this

factcausedPhorm’sstockto“plunge40percent”.ThemostrecentarticleonPhorm

was that by The Register. This article was written about the Crown Prosecution

Servicedecisiontoprosecuteanyoneover“BTandPhorm’ssecretinterceptionand

profiling of Internet traffic”. Apparently BT was using Phorm’s technology on its

clientswithoutaskingpermission.Withthehotwaterthattheyareinnow,itisno

wonderwhyPhorm’s currentwebsite stresses theutmost importanceon security

andprivacyas theyseemtohavebeengetting resistanceon thatmatter fromthe

start.

GoogleAnalyticsOnedatamining application that stands out among all others isGoogleAnalytics.

Google Analytics is a unique application that allows users to gather specific data

fromtheirwebsitesandanalyzethatdatain‐depththroughmanydifferentmethods.

Google markets this process as, “Enterprise‐class web analytics made smarter,

friendlier, and free” (GoogleAnalytics, 2010). The utilization of the analyzed data

can prove to be very beneficial for E‐commerce businesses and lead to an

exponentialincreaseinonlinesales.Thequickandsimplesetup,aswellasseamless

integrationwithotherGoogleApps,demonstratesthatGoogleAnalyticsisessential

toanyE‐commercebusiness.

12

The integration of Google Analytics onto a website is almost effortless. To begin

collecting vital website visitor information, a user must first go to the Google

Analyticswebsiteandcreateanaccount.Onceanaccountiscreatedtheusermust

direct the application to the desired website URL. In addition to creating the

account,apieceofJavascriptmustbeaddedintothecodeofeverysinglewebpage

thattheuserwishestocollectdatafrom.Oncetheaccountislinkedtoaspecificsite

and the script is added,Googlewill begin collecting basic information fromevery

singlewebsitevisitor.“GoogleAnalyticsissimpleenoughforbusinessesnewtoweb

analytics to get started quickly, and sophisticated enough for themost advanced

online marketers” (Harris, 2008). Initially the data gathered will range in its

usefulnessdependingonthewebsite’s traffic. If thewebsitehasa largeamountof

daily visitors the data gathered may be analyzed after only one full day. If the

website has a small amount of traffic the datamay be gathered for one to three

monthsbeforeanaccuratein‐depthanalysisiscomplete.

InvaliddataorlimiteddatamisusecanbedetrimentaltoanE‐commercebusiness.

Identicaltoanyotherformofdataanalysis,theactualdatainformationisessential

to acquiring legitimate beneficial data analysis. When considering using Google

Analytics to guide website management, users must be aware of monthly visitor

trendsanduserpreferences.Ifdataisgatheredintooshortofaperiodtoaccurately

demonstrate the random visitor variances then incorrect analysismay be carried

out. Incorrect analysis can lead towebsite layout reconstruction that can actually

hurt sales. In order to prevent these issues users must be fully aware of the

applicationtermsandformsofanalysis.

13

GoogleAnalyticsallowsuserstodissectvisitordataintomanydifferentformssuch

as:returnon investment,bouncerate,clickthroughrate,andconversionrate.For

an E‐commerce business this type of data analysis is essentialwhen determining

howtoincreasesales.Thebouncerateisknownasthepercentageofvisitorswho

immediatelyleavethewebsiteuponenteringit.Itiskeytoevaluatingonlinetarget

marketing and the overall website appearance. The conversion rate is the

percentage of peoplewho complete a target goal. It is very useful in determining

howmanyvisitorsactuallymakeanonlinepurchase.Theconversionrateistherate

at which data can be used to project sales based upon the number of expected

visitors in the next month or for a defined period. A typical conversion rate can

rangefromzeropercentuptofivepercentforwell‐designedandmarketedwebsites

(GoogleAnalytics,2010).Sincetheoverallamountofpeoplewhomakepurchasesis

so small, it is essential to make sure the majority of the site visitors are your

targetedvisitors.

InadditiontoGoogleAnalytics,therearetwootherkeyapplicationsthatcangreatly

benefit an E‐commerce website: Google AdWords and Google AdSense. Google

AdWords andGoogleAdSense allowusers of GoogleAnalytics to integrate all the

collecteddata informationwithonlinemarketing tools.AdSenseadsaredisplayed

toeverysingleGooglesearchengineusereverytimetheytypeinaspecifickeyword

orphrase.UsersofGoogleAnalyticsareabletomarkettheirwebsitethroughonline

advertisementsthatonlyshowupwhenpeoplesearchfortheirspecificallydefined

keywords. Keywords are chosen by the AdWords user and can be anything from

“SanDiegoPizzeria”to“SkiBoots”.Inadditiontojustbeinganorganictrafficsearch,

14

thespecifickeywordswhentypedinwilldisplayallappropriateandpre‐approved

advertisementsonthesideofthesearchengine.Whenakeywordadvertisementis

clickedonitisconsideredapay‐per‐clickadvertisementandwillcosttheownerof

the specific advertisement a varying amount of money. Every single person that

clicks on that specific advertisement will cost the owner money. The amount of

money varies depending on the popularity of the word and depending on what

othersarebiddingonforthatword.Pricescanrangefromfivecentstoovertwenty

dollarsaclickandcancosttheownersthousandsofdollarseveryday.Thetargeted

keyword advertisement data can be analyzed using the same criteria as website

visitorsandifanalyzedcorrectlyandusedproperlycansignificantlyincreaseonline

salesinashortperiodoftime.

InadditiontoAdWordsadvertisementsthatappearontheGooglesearchengine,E‐

commercemarketers can also use AdSense. Google AdSense is a tool designed to

worksimilarly toAdWords,except itdisplaysadvertisementsonactualwebpages.

AdSensewillfilterthroughallthetextdisplayedwithinawebpageandfindrelative

advertisementsbaseduponthecontentandposttheminalinkbox.Thismethodof

advertisement provides the web page owner with income generated from the

AdSense traffic andgives thewebpagevisitors advertisements that are specific to

theirwebpageinterests.AdSenseworkstheexactsamewayasAdWordswithpay‐

per‐clickadvertisementsandkeywords.

Theimpactthatdatamining,dataanalysis,andtargetmarketingcanhaveonanE‐

commerce business is phenomenal. Google Analytics and its advertising tools can

15

allow a small home business grow into amedium‐sized business over just a few

weeks.Analyzingwebcontentandvisitorsgrantsathoroughunderstandingofwhat

isactuallygoingonwhenvisitorsarriveatawebsite.Evenmajorbusinesseshave

started using Google Analytics, “Google Analytics has been a key facilitator in the

transformation of PBS online. Stakeholders are no longer focused on monthly

reports. Increasingly, theyareusingAnalytics to informactualbusinessdecisions”

(Tonkin,2009).Theeraofguessingandpredictingmarkettrendsandotheraspects

of a business is now evolving into a precise science through datamining. Google

Analytics allows a business to anatomically deconstruct website visitors and

reconstructthatinformationintoeffectiveandprofitabledata.

Testimonials

CostcoTravelCostcoTravel is a subsidiaryofCostcoWholesaleCorporation.Since2000,Costco

Travelhasprovideditscustomerswith:vacationpackages,cruises,hotels,andcar

rentals. In 2008, Costco Travel launched its online booking engine, allowing their

members topurchasepackagesover the Internet.Bycreating thiswebsite,Costco

Travel will be able to expand their offerings to their customers with more

destinationsandcruisepackages.

ThechallengethatCostcoTravelfacedwasthatsellingtravelpackagesonlineisnot

assimpleasitlooks.Thecustomerhastochoosefromahugevarietyofoptionsfor

example,buyingpackagesbasedon:thetimeofyear,differenthotels,severalflights,

combinations of flight and trains with different hotels, car rentals, different

16

destinations,etcetera.EachproducthasanIDforeachcombination,whichisadded

tothecustomer’sbasket,makingthecustomerconfused.“Asatravelbookingsite,

wetrackthousandsofuniquevariablesandparameterswhereusersspecifyhotels,

flights, dates, and locations” said PatMitchell, head ofWeb Analytics and Airline

PartnershipatCostcoTravel.BecauseofthisproblemCostcoTravelcreatedauser‐

friendlysystemthatreliesheavilyonAjaxcallsanddatabasemanagement.

In2008,after launchingtheirwebsite,CostcoTraveldecidedto implementGoogle

Analytics as their soleweb analytics software. After a few discussionswhether it

wasworking for them they finally decided to stickwithGoogle analytics but also

brought in Stratigient,which is a GoogleAnalytics Certified Partner, to perform a

taggingaudit.TheCostcoTravelwebsitewasdesignedtoworkwithanAjaxserver

request,whichmeantthattheURLwouldnotrefreshwhileacustomerwasutilizing

certain parts of the website. This came to play when Stratigient conducted their

audit.Theynoticedthattheyweresomeareasthatweren’tbeingtracked.Theygave

adviceonhowleverageE‐commercetaggingandeventtracking.TheCostcoTravel

teamalongwiththeStratigentteamworkedonaplantoredefinetaggingwiththe

intenttoenhancecontentreportsusingeasilyinterpretedidentifiers.

Since Costco Travel adopted Google Analytics their sales have grown over fifty

percentinvacationpackages.Theteamhasdiscoveredseveralareasofopportunity,

agreatexamplewasthe“what’snew” linkonthehomepage, theyplayedwiththe

placement and the design of it, andwhen they found the perfect spot, the results

werethatthereweremoretrafficcomingintothewebsiteandalotmoreclicks.

17

VuelingVuelingAirlinesisaSpanish‐basedcompanyfoundedin2004.Themainbasesarein

MadridandbeautifulBarcelona.Thisisalow‐costairline,sowithinitsfirsteighteen

monthsofoperation,theyexperiencedahugegrowth.Becauseofthedemand,they

movedfromaround100employeestoalmost600,thefleet increasedfromtwoto

fourteenplanes,whichflewmillionsofpeople.Vuelingoffersloscostticketswitha

quality service.Theirmainsales come from theirwebsiteand the lesserextentof

salescomefromtheircallcenter.

Vueling’scontroloveradvertisingbudgetisessentialtothem,andtheyhavealways

triedtocalculateROImetrics.Eventhough,thecompanyusesmetrics,theydecided

to implement Google Analytics, a tool that “measures not only online marketing

budgetperformance,butalsoanalysesin‐depthuserbehaviorandinteractionwith

thewebsite,identifyingwaystoimproveuserexperience”saysAnnaCaceres,Online

MarketingManagerforVueling.

The conversions (or clicks onwebsite ads)were raised twenty percent only two

weeks after the installation of Google Analytics, and therewas also a ten percent

reduction in the cost of the conversions.AlsowithAdWords,Vuelingmodified its

keywords in order to attract potential customers to thewebsite. GoogleAnalytics

was able to discover that many of the company’s current customers didn’t live

where they usually flew, so with this information Vueling launched advertising

campaign inmanyof these cities. “Analytics allowsVueling tounify all sourcesof

informationandeliminateduplicityinconversionaccountancy”(casestudy).

18

DiscountTireDiscountTireisacompanythatprovidesitscustomerswithawidevarietyoftires

andwheelsatalowcost.Thesetirescanbepurchasedthrougharetailstoreorthe

Internet. DiscountTireshasbeeninbusinessforoverfouryearandisthe largest

independenttiredealerinthenation.

The Internet isbecoming theirmost important sales channel, aheadof their retail

locations. Customers can interact with tires on the actual website by placing the

tiresonvehiclesandseeinghowtheyactuallylookontheircarbeforetheypurchase

them. Discount Tire uses AdWords in their website and it has been an effective

marketing tool, because it can target customer locations and send seasonal

messages.BecauseDiscount tire is a resultdriven company, theyusemetrics and

haveadepartmentthatisdedicatedtothosemetricsandanalytics.GoogleAnalytics

wasagreatacquisitionforthecompany,becausetheinstallationisverysimpleand

requiresnocustomization.

Google Analytics was able to help the company track all of their marketing

campaigns.Theywerealsoabletofollowsomepurchasingpatternsfromcustomers.

GoogleAnalyticswasable topinpointacoupleofbaddesignpatternsaswell.For

example,thecheckoutbuttonread,“Purchaseandmakeareservation”andGoogle

Analyticssuggestedchangingitto“checkoutandmakeareservation”,thischange

increase sales fourteen percent in just one week. Another problem that was not

taken into consideration was that after selecting their tires and seeing that they

wereunavailableattheirlocalstore,peoplewouldjustleavethewebsite.Whenthe

19

marketing departmentwasmade aware of this problem they found a solution by

reassuringtheircustomersthatbythetimetheygettotheirstore,thetireswillbe

waitingforthem.Thissolutionresultedinanincreaseofthirty‐sixpercentinonline

sales. “Every newpiece of informationwe get fromGoogle Analytics gives us ten

newideasthatcanhelpourcustomersfindwhattheyneedonline”,TravisUnwin,

ManagingDirectoratDiscountTire.

AmericanCancerSocietyThe American Cancer Societywas founded in 1913 by fifteen physicians and one

businessman. It is a community based voluntary organization dedicated to

eliminating cancer as a health problem. The society hasmillions of volunteers all

over the country in various research, education and advocacy programs. In 2005,

605 million dollars were funded for these programs, and over 250 million for

managementandfundraising.Thesocietyisusingthewebsite“cancer.org”tohelp

thefightagainstthisterribledisease.

TheAmericanCancerSocietyreallyvaluesmetricsanddatadrivendecisionstoget

the most out of their funding. The society uses many tools to fight cancer, from

research programs to campaigns against tobacco, and even a control diet. Social

networkingwebsites,wikis,chatrooms,contentpages,andopensourceprojectsare

allpartoftheonlinepresenceofthesocietyovertheInternet.Alloftheseneedtobe

evaluatedforperformance.

GoogleAnalyticswithAdWordshadbeenabletohelpthewebsiteinmanyways.For

example,tohighlightmanykeywordsthatcanhelpthesocietygetmoredonations

20

and benefits. The society was one of the first companies to use Google Analytics

whenitcameoutin2005,andhassoonastheyadopteditchangeswereseen,like

increasing their ROI (return on investment) over 600% using AdWords

performanceonTLCdirect.org.Alsothesocietybenefited,becausetheywereableto

tracethecancereducationprogramsthathelpleveragethelatesttechnologies.The

societyhasamobileversionoftheirwebsite,andwithGoogleAnalyticstheywere

abletotracethepeoplethatwentintothewebsiteanddialedtheircallcenter,that

waytheycanmonitortheeducationofcanceralloverthecountry.

FutureofDataMiningIntoday’sbusinesses,dataminingonthewebisvital. Just imaginethatthewebis

estimated to have 150million nodes (pages) and 1.7 billion edges (links). This is

aboutfourbillionpagesandwithapproximatelyonemillionaddedeveryday.It is

by far the largest,mostopen,democtraticpublishingsystemintheworld.Coulda

companyaffordnot to findaway toenhance itspublishingand takeadvantageof

thedatacollected?Thefutureofdataminingisveryprominent.Asthefieldofstudy

becomesofmoreinterest,theinformationnowcontainedsecretlywithincompany

wallswillbedevelopedandbecomeopen‐sourceforresearchers,scholars,andmid‐

sized companies to use. Thiswill create a need for far greater specialization and

morecomplexwaystoanalyzedatatodeterminethefuturebuyer.

Perhapsthefutureofdataminingdoesnotrelyonfindingyourpotentialcustomer,

butcouldshiftintoanalyzingdatatolocatethealphabuyers.Alphabuyersarethose

21

people who create popularity or trends in the market of goods and services. An

alphaconsumerwill“spreadtheword”aboutaninnovativeanduniqueproduct,this

personwill likely buy the product and serve to promote your product. This new

approach to data mining can result in lower marketing costs. It will prompt a

company toseekoutwhat thesealphaconsumersareshopping foranddevelopa

product that meets that need. Once this product is developed then it can be

marketed to the alpha consumer for consumption. The result of this practicewill

resultinanexpectedboomforgood.

Dataminingcouldbeshiftingitsdatasearchintoamoreaggressiveapproach.This

could result in higher R&D costs for a company and less deficits in surplus of

products. Itcouldeffectivelydeterminewhentheirproductisreachingtheend‐of‐

businesscycleandaccuratelychangetheirproductiontoanewproduct. All inan

efforttomaximizeshareholderwealththroughtheuseofsmartinformationmining.

Thefutureholdsdataminingtolocatealphacustomerandpredictfuturetrends.

22

WorksCitedA Brief History of Data Mining. (2006). In Data Mining Software. Retrieved December

7, 2010, from http://www.data-mining-software.com/data_mining_history.htm About Us, How it Works. (2010). In Phorm. Retrieved December 6, 2010, from

http://www.phorm.com/index.html Costco Uses Google Analytics to Grow Costco Travel. (2010). In Google Analytics.

Retrieved December 1, 2010, from http://www.google.com/intl/en/analytics/case_study_costco.html

Data Mining Software. (n.d.). In Data Mining History. Retrieved December 5, 2010,

from http://www.data-mining-software.com/data_mining_history.htm Discount Tire Increased Online Sales by 14% in the First Week Using Google Analytics.

(2010). In Google Analytics. Retrieved December 1, 2010, from http://www.google.com/intl/en/analytics/case_study_discount_tire.html

Enterprise Class Web Analytics Made Smarter, Friendlier, and Free. (2010). In Google

Analytics. Retrieved December 3, 2010, from http://www.google.com/analytics/index.html

Frand, J. (n.d.). Data Mining: What is Data Mining?. In UCLA Anderson School of

Management. Retrieved December 5, 2010, from http://www.anderson.ucla.edu/faculty/jason.frand/teacher/technologies/palace/datamining.htm

Goel, S. (2006). Data Mining in Ebusiness. In Impact Solutions. Retrieved December 6,

2010, from http://www.impact-sol.com/datamining.html Graham, J. (2005, October 3). Google's Adsense A Bonanza for Some Websites. In USA

Today. Retrieved December 4, 2010, from http://www.usatoday.com/tech/news/2005-03-10-google-ads-usat_x.htm

Han,J.,&Kamber,M.(2001).DataMining:ConceptsandTechniques.SanFrancisco:

MorganKaufmannPublishers. Hansell, S. (2008, April 7). Phorm's All-Seeing Parasite Cookie. In New York Times.

Retrieved December 6, 2010, from http://bits.blogs.nytimes.com/2008/04/07/phorms-all-seeing-parasite-cookie/

Harris, C. (2008, August 22). Google Analytics. In ZDNet. Retrieved December 3, 2010,

from http://www.zdnet.com.au/google-analytics-339291469.htm

23

Mailvaganam, H. (n.d.). Future of Data Mining. In Data Warehousing Review. Retrieved

December 5, 2010, from http://www.dwreview.com/Data_mining/Future_data_mining.html

Markov,Z.,&Larose,D.T.(2007).DataMiningTheWeb:UncoveringPatternsinWeb

Content,StructureadUsage.NewBritain,CT:JohnWiley&Sons,Inc.Moss, L. (2003, December 15). Defining Data Mining. In Business Intelligence.

Retrieved December 5, 2010, from http://businessintelligence.com/article/64 Pfanner, E. (2009, July 6). BT Decides Not to Adopt Internet-Based Ad System. In New

York Times. Retrieved December 6, 2010, from http://www.nytimes.com/2009/07/07/technology/internet/07private.html?scp=2&sq=phorm&st=nyt

Phorm/Google Behavioral Ad Targeting. (2009, March 11). In Seobook. Retrieved

December 6, 2010, from http://www.seobook.com/google-phorm-behavioral-ad-targeting-based-your-browsing-data

Story, L. (2008, March 20). A Company Promises Deepest Data Mining Yet. In New

York Times. Retrieved December 6, 2010, from http://www.nytimes.com/2008/03/20/business/media/20adcoside.html?_r=1&scp=2&sq=phorm&st=cse

The American Cancer Society Uses Google Analytics. (2010). In Google Analytics.

Retrieved December 1, 2010, from http://www.google.com/intl/en/analytics/case_study_acs.html

Tonkin, S. (2009, July 14). Teach People to Fish. In Public Broadcasting Service.

Retrieved December 4, 2010, from http://analytics.blogspot.com/2009/07/public-broadcasting-service-pbs-teach.html Google Analytics

Vuelings Flight Tickets Sales Increased by 20% with Google Analytics. (2010). In

Google Analytics. Retrieved December 1, 2010, from http://www.google.com/intl/en/analytics/case_study_vueling.html

WeigandPh.D.,A.(Composer).(2009).DataMiningandElectronicBusiness.San

Francisco,CA,USA. What Online Consumers Want: A personalized Experience. (2008, July 23). In Marketing

Charts. Retrieved December 6, 2010, from http://www.marketingcharts.com/direct/what-online-consumers-want-a-personalized-experience-5022/

24

Wikipedia Contributors. (2010, December 5). Globalization. In Wikipedia. Retrieved December 6, 2010, from http://en.wikipedia.org/w/index.php?title=Globalization&oldid=400718003

Wikipedia Contributors. (2010, December 3). Online Analytical Processing. In

Wikipedia. Retrieved December 4, 2010, from http://en.wikipedia.org/w/index.php?title=Online_analytical_processing&oldid=400398945