1、毕业论文(设计)外文翻译外文原文SALESFORECASTSINCLOTHINGINDUSTRYTHEKEYSUCCESSFACTOROFTHESUPPLYCHAINMANAGEMENTLIKEMANYOTHERS,TEXTILEAPPARELCOMPANIESHAVETODEALWITHAVERYCOMPETITIVEENVIRONMENTANDHAVETOMANAGECONSUMERSWHICHBECOMEMOREDEMANDINGTHUS,TOSTAYCOMPETITIVE,COMPANIESRELYONSOPHISTICATEDINFORMATIONSYSTEMSANDLOGISTIC
2、SKILLS,ANDESPECIALLYACCURATEANDRELIABLEFORECASTINGSYSTEMSHOWEVER,FORECASTERSHAVETODEALWITHSOMESINGULARCONSTRAINTSOFTHETEXTILEAPPARELMARKETSUCHASFORINSTANCETHEVOLATILEDEMAND,THESTRONGSEASONALITYOFSALES,THEWIDENUMBEROFITEMSWITHSHORTLIFECYCLEORTHELACKOFHISTORICALDATATORESPONDTOTHESECONSTRAINTS,COMPANIE
3、SHAVEIMPLEMENTEDSPECIFICFORECASTINGSYSTEMSOFTENSIMPLEBUTROBUSTAFTERTHESTUDYOFEXISTINGPRACTICESINTHECLOTHINGINDUSTRY,WEPROPOSEDIFFERENTFORECASTINGMODELSWHICHPERFORMMOREACCURATEANDMORERELIABLESALESFORECASTSTHESEMODELSRELYONADVANCEDMETHODSSUCHASFUZZYLOGIC,NEURALNETWORKSANDDATAMININGINORDERTOEVALUATETHE
4、BENEFITSOFTHESEMETHODSFORTHESUPPLYCHAINANDMOREESPECIALLYFORTHEREDUCTIONOFTHEBULLWHIPEFFECT,ASIMULATIONBASEDONREALDATAOFSOURCINGANDFORECASTINGPROCESSESISPERFORMEDANDANALYZED1INTRODUCTIONINTHEWIDEAREAOFTHETEXTILE,THREEMARKETSAREGENERALLYDEFINEDAPPAREL,HOMETEXTILEANDTECHNICALTEXTILEINTHISARTICLE,WEINVE
5、STIGATETHEAPPARELMARKETWHICHISTHEMORESIGNIFICANTINTERMOFVOLUMEANDALSOVERYSPECIFICINDEED,IN2006THETRADEOFTEXTILEISESTIMATEDAT530BILLIONDOLLARSWITH319BILLIONDOLLARSFORTHECLOTHINGINDUSTRYTHECLOTHINGINDUSTRYINCLUDESMANYCOMPANIESFROMTHESPINNINGTOTHEDISTRIBUTIONWHICHAREREQUIREDTOTRANSFORMTHEFIBERINTOTHEFI
6、NALGARMENTFIG1THISLEADSTOAQUITELONGANDCOMPLEXPROCESSWITHMANYMANUFACTURINGSTEPSHOWEVER,THEMAINACTOROFTHISNETWORKISTHEDISTRIBUTORDOWNSTREAMOFTHEPROCESSITMAKESORDERSFORTHEUPSTREAMCOMPANIESANDSUPPLIESTHECONSUMERWITHTHEIRPRODUCTSITISTHEDRIVEROFWHOLEFLOWSINTHEPROCESSTHESEDIFFERENTSTAGESWITHQUITELONGANDFLU
7、CTUATEDMANUFACTURINGTIMESINVOLVEAHIGHSENSITIVITYOFTHESUPPLYCHAINTOTHEBULLWHIPEFFECTSALESFORECASTSEMERGETHENASTHEKEYSUCCESSFACTOROFTHESUPPLYCHAINMANAGEMENTHOWEVER,THESPECIFICITIESOFCLOTHINGSALESMAKETHEFORECASTINGPROCESSVERYCOMPLEXINDEED,EXCEPTLONGANDINCOMPRESSIBLEMANUFACTURINGANDSHIPPINGLEADTIMESDUET
8、OCOMPLEXPROCESSOFCLOTHINGPRODUCTION,FORECASTINGSYSTEMSHAVEALSOTOTAKEINTOACCOUNTTHEPARTICULARITIESOFTHECLOTHINGITSELFSALESAREVERYSEASONALSINCEEACHKINDOFGARMENTISRELATEDWITHWEATHERCONDITIONSSEASONALDATAGIVEGENERALTRENDSBUTUNPREDICTABLEVARIATIONSOFWEATHERINVOLVESIGNIFICANTPEAKSORHOLLOWSMANYEXOGENOUSVAR
9、IABLESDISTURBSALESSUCHASENDOFSEASONSALE,SALESPROMOTION,PURCHASINGPOWEROFCONSUMERS,ETCSALESALSOSTRONGLYDEPENDOFFASHIONTRENDSTHISMEANSTHATDESIGNANDSTYLESHOULDBEALWAYSUPTODATEANDTHATMOSTOFITEMSARENOTRENEWEDFORTHENEXTCOLLECTIONCONSEQUENTLY,HISTORICALSALESAREOFTENNOTAVAILABLESINCEMOSTOFITEMSAREEPHEMERALI
10、TEMSAREDECLINEDINMANYCOLORALTERNATIVESWHICHISAVERYFASHIONEDATTRIBUTE,ANDINVARIOUSSIZESWHICHSHOULDMATCHWITHMORPHOLOGIESOFTHETARGETCONSUMERSALLTHESECONSTRAINTSMAKETHESALESFORECASTINGSYSTEMSFORAPPARELCOMPANIESVERYSPECIFICANDCOMPLEXTHEFIRSTSECTIONOFTHISPAPERDEALSWITHTHEDIFFERENTCHARACTERISTICSOFTHETEXTI
11、LEAPPARELMARKETITDESCRIBESTHESPECIFICITIESOFTHECLOTHINGSALESWHICHSHOULDABSOLUTELYBETAKENINTOACCOUNTINTHEFORECASTINGSYSTEMSTHISLEADSTOTHEBUILDINGOFDIFFERENTFORECASTINGSYSTEMSACCORDINGTOTHEREQUIREDHORIZONANDTHECONSIDEREDLEVELOFSALESAGGREGATIONTHESECONDSECTIONINVESTIGATESFIRSTTHEMETHODSUSEDBYCOMPANIEST
12、ORESPONDTOTHESECONSTRAINTSANDTHENSUGGESTSEXISTINGADVANCEDMETHODSTOPERFORMMOREACCURATEANDRELIABLESALESFORECASTSTHESEMETHODSINCLUDEFUZZYLOGIC,NEURALNETWORKSANDDATAMININGTINORDERTOESTIMATETHEBENEFITSTOTHESUPPLYCHAIN,MODELSBASEDONTHESETECHNIQUESAREIMPLEMENTEDONREALDATAFROMTHECLOTHINGDISTRIBUTIONTHELASTS
13、ECTIONFOCUSESONTHEIMPACTOFFORECASTERRORSWITHTHESIMULATIONASUPPLYPROCESSBASEDONREALSALESOF20ITEMSFINANCIALINDICATORSANDFLOWANALYSISENABLETHEEVALUATIONOFPERFORMANCESOFTHERETAILERANDTHEMANUFACTURERANDTRYTOHIGHLIGHTTHEBULLWHIPEFFECT2SPECIFICITIESOFCLOTHINGSALESANDRELATEDFORECASTSFACETOARELATIVELYUNIFORM
14、OVERSUPPLYPROPOSEDBYTHEMANYBRANDSOFCLOTHING,CONSUMERSAREVERYUNFAITHFULANDGENERALLYDIRECTTHEIRCHOICESONCOMPETITIVEPRICESINORDERTOREPLYTOTHECONSUMERREQUIREMENTSANDDUETOTHECOMPETITIVEENVIRONMENT,COMPANIESANDESPECIALLYDISTRIBUTORSOFFASHIONITEMSSUCHASCLOTHING,HAVETOREDUCETHEIRPRODUCTIONCOSTSTOSTAYEFFICIE
15、NTTHUS,MOSTOFMANUFACTURINGPROCESSES,DETAILEDINFIG1,AREPERFORMEDINFARAWAYANDLOWCOSTCOUNTRIESINASIAORSOUTHAMERICATHISGLOBALIZATIONLEADSTOINCREASETHELEADTIMEANDLOTSIZEOFSUPPLIESCONSEQUENTLY,DISTRIBUTORSHAVETORELYONRIGOROUSSUPPLYCHAINMANAGEMENTTOAVOIDDELAY,OUTOFSTOCK,UNSOLDANDTOKEEPTHERIGHTINVENTORYLEVE
16、LINTHELASTDECADE,MANYSUPPLYCHAINMANAGEMENTTOOLSHAVEENABLEDDISTRIBUTORSTOIMPROVESCHEDULINGANDSYNCHRONIZINGOFMATERIALANDINFORMATIONFLOWSMOSTOFTHESETOOLSCANBECUSTOMIZEDTOTHESPECIFICCONSTRAINTSOFTHECLOTHINGRETAILING,HOWEVER,THEIREFFICIENCYISMOSTLYDEPENDENTOFTHEACCURACYOFSALESFORECASTSINORDERTODESIGNANAD
17、EQUATEANDEFFICIENTSALESFORECASTINGSYSTEM,ITISESSENTIALTOKNOWTHEPRODUCT,THESALESFEATURESANDHOWDISTRIBUTORSWILLUSETHEFORECASTSARMSTRONG,200121HORIZONOFFORECASTSTHEDECISIONRESULTINGFROMTHESALESFORECASTSORDERS,REPLENISHMENTS,SHOULDBECONSIDEREDINSUFFICIENTTIMETOALLOWTHEPRODUCTION,THESHIPMENTANDTHEQUALITY
18、CONTROLTHESUPPLYSTRATEGYOFCLOTHINGDISTRIBUTORSISGENERALLYBASEDONTWOSTEPS1AFIRSTORDERATTHEBEGINNINGOFTHESEASONTOENABLETHESUPPLYOFTHESTORESANDTOACHIEVETOARIGHTINVENTORYLEVELINWAREHOUSE2REPLENISHMENTSFORSOMEITEMSDURINGTHESEASONACCORDINGTOTHESCHEDULEOFDESIGN/PRODUCTION/DISTRIBUTIONOFCLOTHINGITEMSGIVENOF
19、FIG2,THISSTRATEGYINVOLVESTWOHORIZONSOFFORECASTALONGTERMHORIZON,IEONEYEAR,TOPLANTHESOURCINGANDTHEPRODUCTIONASHORTTERMHORIZON,IEFEWWEEKS,TOREPLENISHIFNECESSARYANDADJUSTTHEORDERSANDDELIVERIESOFLOCALSTORESCONSEQUENTLY,THEIMPLEMENTATIONOFTWODIFFERENTSALESFORECASTINGMETHODSACCORDINGTOTHEREQUIREDHORIZONISR
20、ECOMMENDEDFORCLOTHINGDISTRIBUTORS22LIFECYCLEOFITEMSTHELIFECYCLEOFITEMSDESCRIBESTHEEVOLUTIONOFSALESFROMTHELAUNCHUNTILTHEDISTRIBUTIONENDOFTHEPRODUCTFOURPHASESAREGENERALLYDEFINEDTHELAUNCHORTHEIMPLANTATION,THERISE,THEMATURATIONANDTHEDECLINEHOWEVERINTHECLOTHINGMARKET,DIFFERENTBEHAVIORSCOULDBEOBSERVEDACCO
21、RDINGTOTHEKINDOFITEMSBASICITEMSARESOLDONTHEWHOLEYEARFORINSTANCEDENIMSOREACHYEARFORINSTANCEBASICWHITETSHIRTFASHIONITEMS,ALSOCALLEDONESHOTITEMS,ARESOLDPUNCTUALLYINASHORTPERIODTHEYAREGENERALLYNOTREPLENISHEDBESTSELLINGITEMSARESOLDEACHYEARWITHSLIGHTMODIFICATIONSACCORDINGTOTHEFASHIONTRENDSANDCOULDBEREPLEN
22、ISHEDDURINGTHESEASONSALESFORECASTSAREREQUIREDFORBASICANDBESTSELLINGITEMS,WHILEFASHIONITEMSWITHONESHOTSUPPLYAREOFTENNOTTAKENINTOACCOUNTINTHEFORECASTINGPROCESS23AGGREGATIONOFSALESEACHGARMENTCANHAVEMANYVARIATIONSINSIZESANDCOLORSCONSEQUENTLYTHENUMBEROFSTOCKKEEPINGUNITSKUTHATTHEDISTRIBUTORHASTOMANAGE,ISV
23、ERYLARGEITISTHUSVERYDIFFICULTTOUSEEACHHISTORICALDATATOPERFORMTHESALESFORECASTSTHENUMBEROFSKUS,THESHORTLIFESPANANDTHEREFERENCECHANGINGFOREACHCOLLECTION,REQUIRETHEDISTRIBUTORTOAGGREGATETHEDATATHEMAINISSUEISTHENTOSELECTTHERIGHTLEVELANDCRITERIAOFAGGREGATIONCLASSIFICATIONMETHODSBASEDONQUANTITATIVEORQUALI
24、TATIVEATTRIBUTESCOULDBEIMPLEMENTEDSEESECTION322,BUTDISTRIBUTORSCOMMONLYPREFERTHEAGGREGATIONOFTHEIRDATAACCORDINGTOTHEHIERARCHICALCLASSIFICATIONGIVENINFIG3THEFAMILYLEVELISVERYSUITABLEFORSALESFORECASTSBASEDONTIMESERIESTECHNIQUESINDEED,THISLEVELENABLESTHEDISTRIBUTORTOGETHISTORICALDATAOFSEVERALYEARSSINCE
25、FAMILIESINCLUDEMANYITEMSEACHYEARINTHELOWERLEVELS,FORINSTANCETHESKULEVEL,DATAAREEPHEMERAL,NOHISTORICALDATAAREAVAILABLETHUS,OTHERTECHNIQUESBASEDONDATAMININGANDCLASSIFICATIONSHOULDBEIMPLEMENTED24SEASONALITYTHECHARACTERISTICSOFTHECLOTHINGITSELFMAKEITSSALESVERYSEASONALINDEED,THEFIRSTGOALOFTHEGARMENTSISTO
26、PROTECTPEOPLEAGAINSTEXTERIORENVIRONMENTANDESPECIALLYTHECLIMATICSOMEITEMSARELOGICALLYMOREIMPACTEDBYSEASONALVARIATION,SUCHASSWIMWEARSORPULLOVERSTHESEASONALITYGIVESAGLOBALTRENDFORTHESALES,ANDTHISSHOULDABSOLUTELYBEINTEGRATEDINTOTHEFORECASTINGSYSTEMFORCLOTHINGSALES25EXPLANATORYVARIABLESTHECLOTHINGMARKETI
27、SSTRONGLYIMPACTEDBYNUMEROUSFACTORSTHESEFACTORS,CALLEDEXPLANATORYVARIABLES,AREOFTENNOTCONTROLLEDANDSOMETIMESUNKNOWNSOMEOFTHEMINVOLVEANINCREASEOFTHEPURCHASEDECISION,OTHERSMODIFYTHESTORETRAFFICLITTLE,1998THUS,ITISDIFFICULTTOEXACTLYIDENTIFYTHEMANDESPECIALLYTOQUANTIFYTHEIRIMPACTDETONIANDMENEGHETTI,2000IN
28、THETEXTILEAPPARELAREA,MARKETINGEXPERTSCOMMONLYTAKEINTOACCOUNTTHEFOLLOWINGVARIABLESNONEXHAUSTIVELISTFORTHEIRINFLUENCEONSTORETRAFFICAND/ORPURCHASEDECISIONLITTLE,1998FIG4WEATHERDATATEMPERATURE,RAINFALL,AREVERYINFLUENTONCLOTHINGSALESHOWEVER,THEYAREUNCONTROLLEDANDVERYFLUCTUATEDTHEFORECASTHORIZONOFWEATHER
29、DATAISALSOTOOSHORTFORANEFFICIENTUSEINTHEAPPARELINDUSTRYSIMILARLY,THEGEOGRAPHICLOCATIONOFSTORESISALSOAFACTORINFLUENCINGSALESCOMPETITIONBECOMESMOREANDMOREIMPORTANTINTHEGLOBALIZEDCLOTHINGENVIRONMENTANDITISTHUSANIMPORTANTFACTOROFVARIATIONSOFSALESCOMPANIESCANGENERALLYKNOWTHEVOLUMEOFSALES,TURNOVER,MARKETS
30、HARE,OFTHEIRCOMPETITORSBUTSOMEFACTORSSUCHASTHEINFLUENCEOFBRANDINGGUPTAETAL,1996ORADVERTISINGCAMPAIGNSOFCOMPETITORSAREDIFFICULTTOINTEGRATEINTOAFORECASTINGSYSTEMCALENDARDATA,SUCHASHOLIDAYS,OFTENINVOLVESIGNIFICANTFLUCTUATIONSINSALESTHESEVARIABLESAREGENERALLYKNOWNWELLINADVANCETOBEINTEGRATEDINTOTHEFORECA
31、STMARKETINGSTRATEGYINCLUDESPROMOTIONALANDADVERTISINGACTIONSWITHORWITHOUTDECREASEOFPRICETHESEACTIONSAREGENERATEDBYTHEDISTRIBUTORSANDTHEYLOGICALLYSHOULDBETAKENINTOACCOUNTINTHEFORECASTPOLITICALOFRETAILCORRESPONDSTOTHEMODEOFDISTRIBUTIONSUPERMARKETS,SPECIALIZEDSTORES,MAILORDER,THENUMBEROFSTORESINTHENETWO
32、RK,THELOCATIONOFSTORES,THEMERCHANDISINGTECHNIQUESTHEEVOLUTIONOFTHESEFACTORSFROMONEYEARTOANOTHERSHOULDBETAKENINTOACCOUNTINTHEFORECASTFEATURESOFITEMSSUCHASTHENUMBEROFCOLORSANDSIZES,THESTYLE,ANDTHEIRMATCHWITHTHEFASHIONTREND,HAVEALSOANIMPACTONSALESHOWEVER,ITMAYBEDIFFICULTTOQUANTIFYTHEIRINFLUENCEMACROECO
33、NOMICANDTRENDDATAPURCHASINGPOWER,UNEMPLOYMENTRATECANHAVEANIMPACTONGLOBALVOLUMEOFSALESITISNOTICEABLETHATTHESEVARIABLESCOULDBEDISTINGUISHEDACCORDINGTOTHEIRIMPACTONSALESINDEED,SOMEOFTHEMGENERATEAPUNCTUALFLUCTUATIONWITHOUTSIGNIFICANTLYIMPACTINGTHEOVERALLVOLUMEOFSALES,FORINSTANCEHOLIDAYSORENDOFSEASONSALE
34、OTHERSIMPACTMOREGLOBALLYTHESALESSUCHASMACROECONOMICDATAORSTRATEGYOFRETAILFIG5REGARDINGTOTHEPREVIOUSREMARKS,ANDBEFORETOBUILDASUITABLEFORECASTSYSTEM,PRACTITIONERSHAVETOKEEPINMINDTHATEXPLANATORYVARIABLESAREESSENTIALTOMODELTHECLOTHINGSALESANDIFPOSSIBLETHEMOSTRELEVANTONESHAVETOBEINTEGRATEDINTHECOMPUTINGO
35、FTHEFORECASTTHESEVARIABLESAREMANYANDVARIEDANDITISNOTPOSSIBLETOESTABLISHANEXHAUSTIVELISTTHEIMPACTOFEACHOFTHESEVARIABLESISPARTICULARLYDIFFICULTTOESTIMATEANDITISNOTCONSTANTOVERTIMETHESEVARIABLESCANBECORRELATEDONTHEMTHISSTRONGLYCOMPLICATESTHEUNDERSTANDINGANDTHEMODELINGOFTHEIRIMPACTONSALESSOMEVARIABLESAR
36、ENOTAVAILABLEIECOMPETITORDATAORPREDICTABLEIEWEATHERDATAANDTHUSCANNOTBEINTEGRATEDINTHEFORECASTINGSYSTEM译文制衣业的销售预测供应链管理的关键成功因素和许多企业一样,纺织服装企业不得不面对一个非常有竞争力的环境,也不得不管理需求越来越高的消费者。因此,为了保持竞争力,公司特别依赖于先进的信息系统和物流技术,尤其是准确可靠的预测系统。然而,预测还要受到纺织服装交易市场的一些奇异限制,例如挥发性需求,销售的强季节性,为数众多的短生命周期项目,或缺乏历史数据。为了应对这些制约因素,公司已经实施的具体预报
37、系统往往比较简单,但强劲。在研究服装行业的现行做法后,我们提出不同的预测模型进行更准确,更可靠的销售预测。这些模型依赖于诸如模糊逻辑,神经网络和数据挖掘的先进方法。为了评价这些方法对供应链的特别是对牛鞭效应减少的好处,我们对采购和预测过程在实际数据的基础上进行了模拟和分析。1简介在宽广的纺织领域里,按一般定义可分为三个市场服装,家用纺织品及技术纺织品。在这篇文章中,我们调查了在数量上非常重要和明确的服装市场。事实上,在2006年的纺织品贸易额估计在530亿美元,其中服装产业就占到319亿美元。服装行业包括许多被分配给纺纱的公司,这些纺纱都被要求从纤维转变成最终的服装(图1)。这导致了制造步骤有
38、一个相当长而复杂的过程。然而,这个网络的主要成员是这个过程的下游分销商。它为上游企业提供订单,为消费者提供产品这是整个流动过程中的驱动程。这些有着相当长和波动生产时间的不同阶段涉及到牛鞭效应中供应链的高灵敏度。然后销售预测作为供应链管理的关键成功因素出现了。然而,服装销售的特殊性使预测过程变得非常复杂。事实上,服装生产的复杂过程除了时间长,它的不可制造和运输交货时间使预报系统也考虑到了服装本身的特殊性销售是非常季节性的,因为每个服装类都与气象条件有关。季度数据发出一般趋势,但天有不测风云的变化使得数据出现重大波峰或凹陷。许多外生变数干扰销售,如最终反季节销售,促销,消费者的购买能力等销售也强烈
39、依赖于流行趋势。这意味着,设计和风格,应始终保持最新,而大多数项目都没有为下一个集合更新。因此,历史销售往往无法获得,因为大多数项目都是短暂的。许多项目在非常老式属性的颜色选择性上被拒绝,当然也会输在配合目标消费者形态的各种规格上。上述种种限制,使销售预测成为服装公司非常具体和复杂的系统。本文的第一部分论述了纺织服装市场的不同特点。它描述了绝对应该考虑在预测系统内的服装销售的具体情况。这导致了不同预测系统根据所需的视野和销售聚集的考虑水平逐一建立起来。第二部分首先调查响应这些限制的公司所采用的方法,然后提出现有的先进方法以执行更准确可靠的销售预测。这里的方法包括模糊逻辑,神经网络和数据挖掘。为
40、了估计供应链所带来的利益,以这些技术为基础的模型根据服装分布的真实数据而实施起来。最后一节的重点集中在以20个项目的实际销售业绩为基础的模拟供应工艺的预测误差的影响。财务指标和流量的分析使零售商和制造商得以进行绩效评估,并尝试突出牛鞭效应。2服装销售及相关预测的具体情况面对相对统一的供大于求的众多服装品牌,消费者对其是不忠诚的,他们一般都直接选择有竞争力的价格。为了满足消费者的要求,并且适应竞争激烈的环境,企业尤其是服装等时尚用品经销商,必须降低其生产成本以保持效率。因此,图1中详细介绍的大部分制造工艺,都在亚洲或南美洲比较偏远和成本低的国家完成。这种全球化导致导致供应的时间和物资的大小增加。
41、因此,经销商必须依靠严格的供应链管理,避免延误,缺货,未售出,并保持正确的库存水平。在过去的十年中,许多供应链管理工具使经销商提高调度,保证物质和信息流的同步。面对服装零售业的具体限制,这些工具中的大多数可以定制解决,然而,他们的效率主要依据销售预测的准确性而定。为了设计一个适当和有效的销售预测系统,关键是要了解产品,销售的特点,以及经销商将如何使用预测(阿姆斯特朗,2001年)。21预测期由销售预测(订单,增资,)产生的决定应用充分的时间考虑,以使生产,运输和质量控制得以进行。服装经销商的供应战略一般是基于以下两个步骤3在赛季开始的第一个订单是为了满足商店的供应,实现正确的仓库库存水平。4为
42、赛季中的一些项目补货。根据图2给出的设计/生产/衣服物品分配的进度,该战略包括两个预测期一个长预测期,即一年,计划采购和生产。一个短预测期,也就是几个星期。如果有必要就补充,调整订单,送货到当地商店。因此,按规定的预测期,我们建议服装经销商执行两种不同的销售预测方法。22项目的生命周期项目的生命周期描述了销售从启动到经销商结束产品的演变。四个阶段的一般定义启动,上升,成熟和衰落。然而在服装市场,按项目类型可以观察到不同的行为基本项目全年(例如牛仔布)或每年(比如基本的白色T恤)均有销售。时尚物品,也被称为“ONESHORT”项目,主要准时在很短的时期销售。他们一般不补货。最畅销的物品的价格,根
43、据流行趋势,每年有轻微的修改,并可能在赛季里的补货。基本项目和最畅销项目要求销售预测,而时尚物品“ONESHORT”往往没有考虑到预测过程。23聚集销售每件衣服在大小和颜色上可以有很多变化。因此代理商管理的股票数量单位(SKU)是非常大的。因此,这很难用各个历史数据进行销售预测。SKU的数量,短寿命和每个集合的参考变化,要求分销商对数据进行聚合。主要问题是,选择合适的水平和聚集条件。以定量或定性的属性为基础的分类方法可以执行(见322节),但经销商普遍喜欢图3给定的按照分层分类的数据集合。科级水平对以时间序列技术为基础的销售预测非常合适。事实上,这一水平使得分销商获得数年的历史数据,因为每一年
44、科级都包括很多项目。在较低的水平,比如SKU的水平,数据是短暂的,没有可用的历史数据。因此,以数据挖掘和分类为基础的其他技术应该得到执行。24季节性服装本身的特性使得它的销售非常具有季节性。事实上,服装的首要目标是保护人们抵御外部环境,尤其是气候。有些项目在逻辑上更受季节变化影响,如游衣或划船用品。季节性因素给出了一个全球性的销售趋势,这绝对应该使服装销售预测系统集成。25说明变量服装市场受到众多因素的强烈影响。这些所谓的说明变量因素,往往不被控制,有时不明。其中一些涉及对购买决策的增加,其他修改存储流量(小,1998)。因此,很难准确识别他们,尤其是量化其影响(DETONI和MENEGHET
45、TI,2000)。在纺织服装领域,营销专家普遍认为以下变量对储流量和/或购买决定有影响(小,1998年)(图4)气象数据(温度,雨量,)对服装销售非常有影响。然而这些数据无法控制,非常波动。天气数据的预测期在服装行业的有效利用也很短。同样,商店地理位置也是影响销售的一个因素。竞争在全球化的服装环境变得越来越重要,因此它也是销售变化的重要因素之一。公司一般可能都知道,他们竞争对手的销售量,营业额,市场占有率,但有些因素如品牌影响力(GUPTA等人。,1996)或竞争对手的广告活动却难以融入一个预报系统。日历数据,如节假日,往往引起销售额大幅波动。一般人都知道提前将这些变量纳入预测之中。营销战略包
46、括促销和有或没有价格下降的广告拍卖。这些行动是分销商所采取的,逻辑上,应该被考虑在预测范围内。政治性零售对应经销模式(超市,专门店,邮购),网店的数量,专卖店的位置,营销技术。预测应考虑这些因素从一年到另一年的发展。项目的特征如颜色和大小的数量,样式,以及它们与时尚潮流匹配数,也对销售有影响。但是,量化其影响可能比较难。宏观经济和趋势的数据(购买力,失业率)可能对全球的销售数量有影响。值得注意的是,我们可以根据这些变量对销售的影响加以区分。事实上,他们中一些对总销售量没有显著影响的变量总会产生准时波动,比如节假日销售或结束季节销售。其他销售影响更全球化,如宏观经济数据或零售战略(图5)。根据前面的市场,然后提前建立一个合适的预测系统,从业人员必须牢记说明变量是必不可少的服装销售模式,如果可能,最相关的变量,必须综合在预测的计算值内。这些变量多种多样,我们不可能建立一个详尽的清单。每个变量的影响是特别难估计的,而且随着时间变化每个变量都不是常数。这些变量可能与相互关联。这会使变量对销售影响的理解和建模严重复杂化。有些变量是不可用(即竞争者数据)的或可预测(即气象数据)的,因此不能在预测系统中集成。