文献信息：
文献标题：New trends of Intelligent E-Marketing based on Web Mining for e-shops（基于 Web 挖掘的电子商务智能网络营销新趋势）
：
国外作者：Jon Kepa Gerrikagoitia，Iñigo Castander，Fidel Rebón，Aurkene Alzua-Sorzabal 文 献 出 处 ：《 International Conference on Strategic Innovative Marketing》,2014,175 :75-83 字数统计：英文 2554 单词，13103 字符；中文 4229 汉字
外文文献：
New trends of Intelligent E-Marketing based on Web Mining for e-shops Abstract
E-marketing is the concentration of all efforts in terms of adapting and developing marketing strategies into the virtual spaces: web, social media… In an e-commerce site, e-marketing must help consumers in their purchase. This requires precise knowledge of the customer´s preferences. For this reason, holders of e-shops must find out to whom, to what, how and when to refer to the customer, ergo, to know the “consumer decision journey” and strengthen their engagement. This analysis is obtained when the customer is visiting an e-shop because (s)he leaves a digital footprint that can be used to understand his/her needs, desires and demands as well as to improve web presence. These data can be used for data mining to understand the e-marketing and selling processes in a better way. In this paper a survey of 86 e-shops in Spain is presented. In the conclusions, some ideas for good e- marketing practices related to the buying behaviour analysis of customers are shown. Hence, new trends in e-marketing are suggested from a strategic, tactical and operational level in which different data mining techniques ease the purchase and the engagement. Key words: e-marketing; e-commerce; web mining; trends e-marketing; digital
footprint
1.Introduction In the last decade Internet has experienced a 33% high growth in the number of users worldwide. This has led to a greater acceptance of it by consumers, either as an instrument that facilitates the search for information or as a way to an end for the purchase of the products and services (“The internet a decade later”, 2013). This new technological environment has generated a significant change in the behaviour, habits and trends of the consumer and its acceptance by consumers leads to the increase of the Electronic Commerce (EC). Defining e- commerce as the commerce in products or services conducted via computer networks such as the Internet (Turban, Lee, King, Liang, & Turban, 2009). This commercial activity has forced small businesses to develop an active presence of visibility on the Internet (Chen, Chiang, & Storey, 2012) in order to get notoriety in a global market and add value to their traditional business. With the exception that the virtual and traditional customer behaviour are not identical (Eouzan, 2013). But visibility of Internet does not only give visibility to a company, also allows to: a) to have the business open to the public permanently; b) reduce costs: material stock, logistic personnel, physical space; c) to study the customer better; d) to remove geographic limitations and e) to create brand and engagement. In Spain, according to the study of the National Observatory for Telecommunications and the Information Society (ONTSI) the B2C e-commerce grew by 13% reaching € 12,383M of turnover. This growth’s main cause, in the same way as in the world market, was the increase of Internet purchasers in 2012 which matched 15% (15.2M) (ONTSI, 2013). A closer data in time (2013) and origin (region level) is available through the National Statistics Institute of Spain (INE). According to this statistics institute, in 2013 the e-commerce in Spain has increased to around 11 million people who have made any purchase through the web in the last 12 months. This represents 32% of the
total population. Moreover, the region of Spain that has used this type of trade the most is the Basque Country with 42%, followed closely by the Community of Madrid with 40% (INE, 2013). Putting the focus on the case of Basque Country, Statistic Institute of Basque Country (EUSTAT) provides some information about e-commerce by data-bank and annual reports (under the information society topic). In this report, there is an interesting section for the e-commerce. The section reflects that 39% of the users that have been recently connected to the internet have made some kind of web based purchase (EUSTAT, 2014). As it is visible on the data provided by official statistic institutions, e-commerce is a commercial activity on growth and due to its nature. It benefits the economy of a country greatly and especially to small businesses which are committed to its implementation. As it happens in real sales, e-commerce based selling must be in concordance with the customer´s necessities prior, during and after the sale. This requires precise knowledge of the customer´s preferences. For this reason, owners of sites must find out to whom, what, how and when to refer to the virtual visitor and develop specific marketing strategies for the Internet (Alzua-Sorzabal, Gerrikagoitia, & Rebón, 2014a). Defining e-marketing as the concentration of all efforts in the sense of adapting and developing marketing strategies into the web environment. E-marketing involves all stages of work regarding a website, such as: the conception, the projects itself, the adaption of the content, the development, the maintenance, the analytic measuring and the advertisement (Strauss, Frost, & Ansary, 2009). In short, e-marketing makes the usage of the Internet to obtain knowledge about customer preferences through interactive website and agents easier (Teo & Choo, 2001). To keep the customer’s attention on the web presence requires to build up a strong customer relationship and to offer services that attract the customer to visit the website frequently and purchase products and services. Therefore, the sales process requires a deep data analysis to know the “consumer decision journey” (Gefen &
Straub, 2000). When a customer is visiting a website (s)he leaves a trace of data, called the digital footprint, that can be used to understand the customer needs, desires and demands as well as to improve its web presence. By the use of this information applied from Web Usage Mining (Arbelaitz et al., 2013) techniques on the digital footprint left by the user during their navigation; knowledge about the behaviour of the customer"s purchase is enriched in order to increase his/her engagement. Then, this knowledge has to be converted into intelligence and, if it is possible, an entertaining presentation of the information required by the customer. In this context, the real situation of e-marketing will be seen through an on-line survey that has been supplied to retailers of Spain for their completion and four in depth interviews. After analysing the state, enriched technical solutions through the Web Usage Mining process will be proposed and linked to new discovered e-marketing trends. Finally, a few commercial and open source solutions that can help in the creation of concrete solutions for these new e-marketing trends will be displayed.
2.Methodology In accordance with the meaningful relevance that e-shops have obtained in the last years, it is crucial to conduct successful e-marketing activities in order to maintain or improve the business in these kinds of companies. At this point, it is necessary to suggest a solution from a strategic vision point of view by developing appropriate e–marketing strategies. The solution derives from a tactic vision by deciding for certain e-marketing trends and at the end an operational vision which is implemented by selected trends & practices (See Fig. 1).
Fig. 1. Business vision and e-marketing Even so, it is necessary to know the initial state of e-marketing into a sample of the Spanish enterprises dedicated to e-commerce. The methodology that has been carried out to know the initial state is based on two different manners of obtaining the real state of e-marketing paved by the e-commerce. On one hand, 86 online surveys have been made to some e- shop owners. On the other hand, out of these 86 companies four have been nominated to have an interview in depth. These interviews were conducted via telephone or face-to-face in the client"s own company and the online surveys were collected through a digital form made in Google Forms. Once the surveys were completed, we proceeded to extract the information from them in order to the draw the conclusions that drive the second stage of the process. This second step consists of: a) a reflection by experts on the detected deficiencies and b) the proposal of practical solutions that facilitate the achievement of strategic e- marketing solutions with vision of business and that incorporate techniques of Data Mining.
3.Results and discussion 3.1.Results of surveys and interviews On one hand, the analysis of the survey data shows that the majority of the e-shop owners questioned are small enterprises with less than 5 employees in the B2C
sector (See Fig. 2a) that commodities, quick sale. They are very comparable in number of products and time in the online business (1-2 years) (See Fig. 2c). They also have a maximum number of online orders that (usually) a range from 101 to 1,000 orders a year and yearly revenue of less than 50,000 € (See Fig. 2b). The main languages supported by the e-shops are Spanish and English.
Fig. 2. (a) Number of employees; (b) Annual revenue; (c) Long online business Although, most of the e-shop owners questioned are currently using a service or tool for customer behaviour analysis in an automatic and online way, in a 40% of the cases more or less they are interested, in a service in a service that is able to analyses the behaviour of groups of customers. 80% of those interested on such a service are aiming to know the customer habits and detect new e-shopping tendencies. Being a 5% of the owners respect to total are conditionally interested. In this case, the requirement is a straightforward system (See Fig. 3a). The e-shops were also asked about: a) the competitor’s observation and the price adjustment and b) the antifraud. In the first case, they said the price is compared to the main competitors’ website. In the second case, the proportion of fraudulent cases of their total transaction volume was lower than 0.1% in 53% of the questioned e-shop owners. Out of them, 32% dealt with online payment fraud: by reviewing the incoming transactions or by transferring risk to a third party (See Fig. 3b). Fig. 3. (a) Interest of behaviour analysis; (b) Proportion of fraudulent cases
On the other hand, analysing the companies that have done the interview, it seems that all of them offer minimally processed, high-demanded products. Furthermore, 50% of the companies surveyed do not know exactly the main reasons why the customers make purchases at their e-Shop. It also is highlights, the fact that only these SMEs employ a person to ensure online sales and make little investments in their online presence (<10K €). In short, these are companies with a small volume of business and their revenues from online sales do not make a big impact on their business. Remark that, it has been less than 5 years since the e-shops started selling their products via Internet and, therefore, it is a process that must evolve towards maturity. The e-shop considers the on-line service as an added service to its business performance but not a priority. In addition to this, it can be seen that the e-shops have a certain level of infrastructure because they often have outsourced the hosting of their website and almost all use Google Analytics as a web analytics tool (Clifton, 2012; Hasan, Morris, & Probets, 2009). Out of the e-shops that use Google Analytics, very few do actually take advantage of the tool. In the same way, according to the surveys, the fraudulent case out of the total volume of transactions does not exceed 1% and they deal with it by leaving it in the hand of third parties: banks and gateways. They do not use electronic tools to observe and compare the prices of its competitors because it is the provider the one that sets them.
To sum up, with the extracted data it is concluded that many e-commerce use Web Analytics tools, specifically Google Analytics, but very few information is obtained from the purpose of the client by his/her digital footprint. This prevents to properly categorize their behaviour and facilitate effective e-marketing campaigns. 3.2.Thinking and discussion based on solutions The need for developing specific marketing strategies for e-commerce implies that some traditional principles are adapted, or even reinvented. Four activities facilitate the deployment of e-marketing strategies (Stolpmann, 2001): online
promotion, online shopping, online service and online collaboration. In addition to this, we think that it would be helpful it is necessary to choose some of the most relevant e- marketing trends and current trends & practices based on the considerations of experts who are close to the market demand and the practical trends web literature. This means that, a part of scientific base, it has obtained information from other important web origins where it might have significant information about real cases that are closer to reality. Therefore, the description of the trends: brick and mortar, Offering of increasingly complex online features, mobile everywhere... related to the e-marketing is crucial, for then, these can be crossed with trends & practices. These latest trends technological advances must be geared towards a future where retailers can get the most out of their online sales. In this regard, there are certain factors that should be taken into consideration such as the pick-up speed or omnichannel customer experience, social-networking testing (Ayada & Elmelegy, 2014), list of wishes, search engine optimization or mega markets, cross border e-sales, suggestive selling (Mussman, Adornato, Barker, Katz, & West, 2010), web banner advertising (Ozen & Engizek, 2014) and rewards. By doing so, it is obtained the necessary traceability which allows the intelligent e-marketing process be more consistent and stay closer to the e-shops reality.
4.Conclusions and future work This study aimed to test the necessity of the SMEs to become active users of web mining solutions. Overall, from the results it can be extracted that: a) the Pricing analysis is not relevant for them, because in the majority of cases is established by the products providers; b) the fraud has not any impact in the online purchases. It is inferred from the responses, that the market has not obtained the necessary maturity yet and that this issue is often outsourced to third party partners; c) Almost all the questioned companies have deployed the web analytics over their websites but they do not follow an exhaustive data analysis. Therefore, keeping in mind these key points and observing the significant importance of the digital footprint at the beginning of the web mining processes, it
must regard that in a short-term the data origins based on this system will raise and the weight that this step will have on the process will be crucial, in order to dispose a better raw data. Furthermore, as in real sales, the process of selling based on e-commerce must be in concordance with the customer´s necessities prior, during and after the sale. This requires precise knowledge of the customer´s preferences. For this reason e-commerce, must find out to whom, what, how and when to refer to the virtual visitor. In order it will take advantage of applying the data mining techniques in the phase that it corresponds to satisfy the necessity of enrich the data and add value according to better information about the client. Additionally, intelligent e-marketing processes should bear in mind the necessary relationships between e- marketing trends and Data Mining techniques in order to develop specific marketing strategies for the Internet.
中文译文：
基于 b Web 挖掘的电子商务智能网络营销新趋势
摘要
网络营销致力于将营销策略应用和发展到虚拟空间：网络，社交媒体……在电子商务网站上，网络营销必须促发消费者的购买欲望。这就需要精确了解客户的偏好。因此，网上商店的店主必须了解到谁、什么、怎样以及何时向客户提及才是最合适的，所以了解“消费者决策历程”并加强他们的参与，就显得很重要。这种分析可以在客户访问网店时获得，因为客户在访问网店时会留下数字足迹，通过数字足迹可以了解到客户的需求、欲望和要求，以及改善网站的呈现状态。这些数据可用于数据挖掘，以便更好地了解网络营销和销售流程。本文介绍了对西班牙 86 家网上商店的调查。在结论中，提出了一些与客户购买行为分析相关的良好网络营销实践的想法。因此，从战略、战术和运营层面提出了网络营销的新趋势，其中不同的数据挖掘技术简化了购买和参与过程。
关键词：网络营销；电子商务；web 挖掘；网络营销趋势；数字足迹
1. 引言
在过去十年中，互联网在全球用户中的增长率高达 33％。这导致了消费者对它更大程度的接受，无论是作为一种促进信息搜索的工具，还是作为一种购买产品和服务的方式（《十年后的互联网》，2013）。
这种新的技术环境已经使消费者的行为、习惯和趋势产生了重大变化，消费者对其的接受也刺激了电子商务（EC）的增长。电子商务被定义为通过诸如互联网等计算机网络进行的产品或服务中的商业（图尔班，李，金，梁和图尔班，2009）。
这种的商业活动促使小企业在互联网上积极发展知名度（陈，蒋和斯托里，2012），以便在全球市场上获得名声，并为其传统业务增添价值。只不过虚拟和传统客户的行为是不相同的（埃乌藏，2013）。
此外，互联网的可见性不仅能够让公司获得知名度，还可以：a）永久地向公众开放业务；b）降低物料库存、后勤人员、物理空间等成本；c）更好地研究客户；d）消除地域限制；e）创建品牌和参与。
在西班牙，根据国家电信和信息社会观察站（ONTSI）的研究，B2C 电子商务增长了 13％，达到 12.383 亿欧元的营业额。与世界市场一样，这一增长的主要原因是 2012 年互联网购买者增长了 15％（15.2 亿）（ONTSI，2013）。
西班牙国家统计研究所（INE）提供了更为及时的数据（2013）和原产地（地区级）。据统计研究所统计，2013 年西班牙的电子商务增加到 1100 万人左右，他们在过去的 12 个月中通过网络进行购买。这占总人口的 32%。此外，西班牙使用这种贸易最多的地区是巴斯克，为 42％，其次是马德里，达 40％（INE，2013）。
重点以巴斯克地区为例，巴斯克地区统计研究所（EUSTAT）通过数据库和年度报告（信息社会专题）提供了一些有关电子商务的信息。在这份报告中，电子商务有一个有趣的部分。这部分反映出最近连接到互联网的用户中有 39％进行了某种基于 web 的购买（EUSTAT，2014）。
从官方统计机构提供的数据中可以看出，电子商务是一种因其性质而发展起来的商业活动。它极大地促进了一个国家的经济，特别是对致力于发展小企业经济的国家。
正如在实际销售中所发生的那样，基于电子商务的销售在销售之前、期间和之后，必须与客户的需求相一致。这就需要精确了解客户的偏好。因此，网上商
店的店主必须了解到谁、什么、怎样以及何时向虚拟访客提及才是最合适的，并制定针对互联网的具体营销策略（Alzua-Sorzabal，Gerrikagoitia 和 Rebón,，2014a）。
网络营销被定义为致力于将营销策略应用和发展到网络环境中。网络营销涉及到一个网站的所有工作阶段，例如：网站设计，项目本身，内容的适应性，开发，维护，分析测量和广告（斯特劳斯，弗罗斯特和安萨里，2009）。简而言之，网络营销是利用互联网通过互动网站和代理来更容易获取客户偏好的信息（张和朱，2001）。
为了让客户将注意力放在网站上，需要建立强大的客户关系，并提供服务，以便吸引客户频繁访问网站和购买产品和服务。因此，销售过程中需要深入的数据分析来了解“消费者决策历程”（杰芬和斯特劳布，2000）。
当客户正在访问一个网站时，他会留下了一些称为数字足迹的数据，通过数字足迹可以了解到客户的需求、欲望和要求，以及改善网站的呈现状态。
对于用户在导航过程中留下的数字足迹，可以应用Web挖掘（Arbelaitz等人，2013）的这种信息技术；掌握客户购买行为的知识，以增加他/她的参与度。然后，这种信息必须转化为智力，如果有可能，可以对客户所要求的信息进行有趣的展示。
在这种情况下，网络营销的真实情况将通过对西班牙零售商提供的在线调查和四次深入访谈来了解。在分析了状态之后，将通过 Web 使用挖掘过程提出丰富的技术解决方案，并将其与新发现的网络营销趋势联系起来。最后，将展示一些商业和开源解决方案，这些方案有助于为这些新的网络营销趋势创造具体的解决方案。
2. 方法
根据网上商店在过去几年中取得的重要意义，对于维护或改善这些公司的业务而言，进行成功的网络营销活动以至关重要。在这一点上，有必要从战略眼光的角度提出解决方案，制定适当的网络营销策略。该解决方案源于战略愿景，通过确定某些网络营销趋势，并最终通过选定的趋势和实践来实现业务愿景（见图1）。
图 1.业务愿景和网路营销 即便如此，也有必要了解网络营销的初始状态，并将其作为一个致力于电子商务的西班牙企业的样本。
了解初始状态的方法是基于获取有电子商务组成的网络营销真实状态的两种不同的方式。一方面，对一些网上商店的店主进行了 86 次在线调查。另一方面，对这 86 家公司中 4 家进行了深入访谈。这些访谈是通过电话或面对面的方式在客户自己的公司进行的，在线调查是通过 Google Forms 中提供的数字表单收集的。
一旦调查完成，我们就从中提取信息，以便得出推动第二阶段进程的结论。第二步包括：a）专家对检测到的缺陷进行反思，以及 b）提出切实可行的解决方案，以促进实现具有业务愿景的战略网络营销解决方案，并将数据挖掘技术纳入其中。
3. 结果与讨论
3.1. 调查和访谈的结果
一方面，对调查数据的分析表明，大多数网店的店主被质疑是 B2C 行业中少于 5 名员工的小型企业（见图 2a），他们主要销售快消品。他们在线业务的数量和时间上非常相似（1-2 年）（见图 2c）。他们的在线订单数量最多（通常）为每年 101 到 1,000 个订单，年收入低于 50,000 欧元（见图 2b）。网上商店支持的主要语言是西班牙语和英语。
图 2.（a）雇员人数；（b）年度收入；（c）长期在线业务 虽然大多数网店店主被质疑目前正在使用服务或工具对客户的无意识在线行为进行分析，但在只有 40%的店主会对在服务中能够分析客户群体的行为感兴趣，对这类服务感兴趣的人中有 80％旨在了解客户的习惯并发现新的网络购物趋势。5%的店主对所有的都感兴趣。在这种情况下，需求是一个简单的系统（见图 3a）。
同时网上商店也被问及：a）竞争对手的情报和价格调整，以及 b）反欺诈。第一种情况，他们表示将价格会与主要竞争对手的网站进行比较。第二种情况，53％的受质疑的店主的欺诈案件占总交易量的比例低于 0.1％。其中，32％涉及到网上支付欺诈行为：通过查看收到的交易或将风险转移给第三方（见图 3b）。
图 3.（a）行为分析的兴趣；（b）欺诈案件的比例 另一方面，对接受采访的公司进行分析，似乎所有的这些公司都提供微加工高需求的产品。此外，有 50％的受访公司并不清楚客户在网络上购物的主要原因。这也凸显了这样一个事实，即这些中小企业只雇用一个人来确保网上销售，而很少对他们的网络形象进行投资（小于 1 万欧元）。简而言之，这些企业的业务量很小，网络销售的收入并不会对他们的业务造成很大的影响。
注意到，网上商店开始通过网络销售产品不到 5 年的时间，因此，这是一个必然走向成熟的过程。网上商店将在线服务视为业务绩效的附加服务，但不是优先事项。
除此之外，可以看出，网上商店需要一定的基础设施，因为他们经常外包其
网站托管，几乎全部使用谷歌分析来作为网络分析工具（克利夫顿，2012；哈桑，莫里斯和珀倍特，2009）。在使用谷歌分析的网上商店中，很少有人真正利用了这个工具。
同样地，根据调查，交易总额中的欺诈案例不超过 1％，并将其交由第三方银行和网关处理。他们不使用电子工具来观察和比较其竞争对手的价格，因为他们的提供商就是设置者。
总之，利用提取的数据，可以得出结论，许多电子商务使用 Web 分析工具，特别是谷歌分析，但是只有很少的信息是通过客户的数字足迹获得的。这样就无法正确分类其行为并促进有效的网络营销活动。
3.2. 基于解决方案的思考和讨论
制定电子商务的具体营销策略的必要性意味着一些传统原则被改变，甚至被重新制定。促进网络营销策略的部署的有四个活动（施托尔普曼，2001）：在线推广、在线购物、在线服务和在线协作。
除此之外，我们认为，有必要根据贴近市场需求的专家和网络文献的实际趋势的考虑，选择一些最相关的网络营销趋势和当前趋势与实践。这意味着，作为科学基础的一部分，它已经从其他重要的 Web 来源获取信息，其中可能有更接近真实情况的重要信息。因此，趋势可以描述为：实体，将提供越来越复杂的在线功能，移动无处不在……相关的网络营销至关重要，因此，这些可以与趋势和实践交叉。这些最新趋势的技术进步必须着眼于未来，零售商可以从他们的网上销售中获益最多。在这方面，有一些因素应该被考虑到，例如快速或全方位的客户体验，社交网络测试（阿雅达和埃尔麦奇，2014），愿望清单，搜索引擎优化或大型市场，跨境电子销售，建议性销售（穆斯曼，亚多拿图，巴克，卡茨和韦斯特，2010），网页横幅广告（欧森和艾吉奥戴克，2014）和奖励。
通过这样做，获得了必要的可追溯性，使得智能网络营销过程更加符合和贴近网上商店的现实。
4. 结论和未来工作
本研究旨在测试中小企业成为 web 挖掘解决方案活跃用户的必要性。总的来说，从研究结果可以得出以下结论：a）定价分析与他们无关，因为大多数情况
是由产品供应商确定的；b）欺诈对网上购物没有任何影响。从反馈中可以推断，市场尚未达到成熟，而且这一问题往往被外包给第三方合作伙伴；c）几乎所有被质疑的公司都已经在他们的网站上部署了 web 分析，但并没有进行详尽的数据分析。
因此，在 web 挖掘过程开始时，要牢记这些关键点和观察数字足迹的重要性，必须考虑到，在短期内基于该系统的数据来源将会提高，而这一步对过程的权重将是至关重要的，以便处理更好的原始数据。此外，与实际销售一样，电子商务的销售在销售之前、期间和之后，必须与客户的需求相一致。这就需要精确了解客户的偏好。因此，网上商店的店主必须了解到谁、什么、怎样以及何时向虚拟访客提及才是最合适的。将数据挖掘技术应用到相应的阶段，以满足丰富数据的需求，并根据客户端的更好的信息来增加数据的价值。
此外，智能网络营销过程应该牢记网络营销趋势和数据挖掘技术之间的必要关系，以制定互联网的具体营销策略。
相关热词搜索： 外文 文献 网络营销