This thesis explores the application of machine learning 8ml9 techniques in customer classification and their intergration into customer relationship management (CRM) systems within the corporate insurance industry. The research aims to address the gap in the use of AI-CRM for the corporate insurance industry. It was conducted as a case study at a Swedish insurance broker company. The study leveraged external data sources to create a data seet on customer information. The feature selection process included Variance Influence Factor (VIF) to remove collinearity and then Mutual Class Info and Random Forest, which are methods used to find which independent variables affect the dependent variable the most. Also, Recursive Feature Testing was applied to find the best feature combinations. Four different binary classification models were implemented and compared - Decision Tree, Random Forest, Support Vector Machine, and Artificial Neural Network. Note that Random Forest can be used both for feature selection and classification. The models were tested on four different feature combinations and evaluated using Accuracy, Recall, Precision, F1-score, and ROC-AUC. The study further conducted interviews at the partner company to evaluate their current CRM system. The findings show that ML-based customer classification can be leveraged to effectivize the customer acquisition process for corporate insurance. The Support Vector Machine model achieved the highest accuracy, at 80%. Depending on the avaliable data and the use of metrics, different classifiers had the best performance. The study also found that when implementing classification into AI-CRM, the specific requirements at the company need to be examined. This study found it important to conersider the data procurement process, the current customer acquisition process, the risks associated with misclassification, and present bias. The findings of this study have theoretical implications for the implementation of AI-CRM for customer acqusition. It demonstrates the practical benefits of intergrating machine learning techniques into CRM systems, emphasizing the effectiveness of AI-CRM for customer classification. Further, by comparing different classification models and evaluating their performance, the study enhances the theoretical understanding of model selection for customer classification tasks in this specific domain. Additionally, the research provides insights into effective feature selection methods, aiding researchers and practitioners in extracting relevant variables for customer classification. / Denna studie utforskar tillämpningen av maskininlärning (ML) inom kundklassificering och dess intergration i kundrelationssystem (CRM) inom företagsförsäkringsbranschen. Forskningen syftar till att fylla kunskapsluckan inom användningen av AI-CRM inom företagsförsäkringsbranschen. Studien genomfördes som en fallstudie på ett svenskt försäkringsmäklarföretag. Studien utnyttjade externa datakällor för att skapa en dataset av kundinformation. Processen för val av variabler inkluderade Variance Influence Factor (VIF) för att ta bort kollinearitet och sedan Mutual Class Info och Random Forest, som är metoder som användsför att hitta vilka oberoende variabler som påverkar den beroende variabeln mest. Dessutom användes Recursive Feature Testing för att hitta de bästa kombinationerna av funktioner. Fyra olika binära klassificeringsmodeller implementerades och jämfördes- Decision Tree, Random Forest, Support Vector Machine och Artificial Neural Network. Observera att Random Forest kan användas både för val av funktioner och klassificering. Modellerna testades med fyra olika kombinationer av variabler och utvärderades med hjälp av Accuracy Recall, Precision, F1-score och ROC-AUC. Studien genomförde även intervjuer på partnerföretaget för att utvärdera deras nuvarande CRM-system. Resultaten visar att ML-baserad kundklassificering kan användas för att effektivisera processen för kundanskaffning inom företagsförsäkring.  Support Vector Machine-modellen uppnådde högst accuracy, 80%. Beroende på tillgängliga data och användning av evalueringsmått hade olika klassificerade bäst prestanda. Studien fann också att vid implementering av klassificering i AI-CRM måste de specifika kraven på företaget undersökas. Denna studie fann det viktigt att beakta processen för dataanskaffning, den nuvarande processen för kundanskaffning, riskerna med felklassificering och nuvarande partiskhet. Resultaten av denna studie har teoretiska implikationer för implementeringen av AI-CRM för kundanskaffning. Den visar på de praktiska fördelarna med att integrera maskininlärningstekniker i CRM-system och betonar effektiviteten hos AI-CRM för kundklassificering. Dessutom förbättrar studien den teoretiska förståelsen för val av modeller för kundklassificeringsuppgifter i det specifika domänet genom att jämföra olika klassificeringsmodeller och utvärdera deras prestanda. Studien ger också insikter om effektiva metoder för val av variabler och hjälper forskare och utövare att extrahera relevanta variabler för kundklassificering.

隨著競爭環境日漸嚴苛，顧客需求日趨多元，企業需要有更有效率的方式來管理與顧客間的關係，顧客關係管理CRM（Customer Relationship Management, CRM）成為企業關心的熱門議題。但過去的學者研究中，僅止於針對影響企業導入＆系統與否之決策進行研究，然而CRM對於企業經營顧客關係之重要性已被企業所認同，同時面對如此牽涉到龐大人力、資金，與時間投入的系統導入計畫，在導入過程中仍會面臨許多重要決策，因此本研究將針對響導入過程的相關因 數，並經由個案研討的實際驗證，提出具解釋力的理論架構。 經由過去文獻的整理，本研究以Rogers（1983）的創新擴散模型為基礎，將影響CRM導入過程的因素分為認知階段的環境面與組織面因素，以及說服階段認知的創新特質，另外系統供應商及顧問公司則扮演特殊的角色，同時在說服、決策及實行階段皆產生影響。其中環境面又包含競爭強度，需求不確定性及產業環境變化速度；組織面因素包含規模、結構化程度、以及高階主管態度。而認知的創新特質則包含相容性以及系統特性。另外則以資訊與資料的蒐集、累積與儲存、吸收與整理、展現與應用等為被影響的研究變數。並選取目前導入最積極的產業中之標竿企業，包含證券、人壽、行動電話系統業者、電子電腦公司，及網際網路服務提供業者等六家企業，進行深入之訪談，期能從這些正在導入的公司之經驗裡，發現可供參考與依循的準則。 經由對個案公司的深入訪談之後，本研究之研究架構有重大修正。認知階段的影響因素除了環境與組織構面仍存在外，從組織構面中抽離複雜度，並與產品特質、顧客基礎，公司依賴業務人員推廣業務及提供服務的程度共同組成新構面—業務面因素，而系統與顧問公司的選擇則成為企業採用CRM系統時重要的決策因素之一。更重要的是，本研究經由訪談對探討主題有更廣泛而完整的延伸，導入過程可以導入積極度、建置組織，以及建置方式來描述。導入積極度代表導入的速度；建置組織裡包含以專案小組的方式來推動，以及高階主管的參與程度；建置方式則包含CRM要素的建置優先順序、以及委外程度。 研究發現，環境面的因素與組織面的因素皆同時影響到CRM導入過程的積極度及建置方式裡的優先順序；而新增的業務面的因素則同時對CRM導入過程的積極度及建置方式裡的優先順序及自建或委外的選擇有較顯著影響。至於認知的創新特質中，不論是專案管理或是高階參與，其影響到的皆為CRM在建置單位上的特性。而協力單位則因為企業對於系統特性的要求有所不同，而同時成為決策結果與影響導入方式的變數。 本研究期望能藉由結合學術理論與實務應用的方案，提供給正在導入或評估規劃中的企業實用的考慮方向與實際例證，以協助企業在導入過程中自我檢測，選擇最適當的導入程序及設計組織關的配套措施，俾使導入過程順利而達到預期目標。 / To cope with the current increase in both competition and customer requirements, enterprises need more efficient methods to manage their relationships with customers. Previous researches focused mainly on the factors affecting the decision whether or not to adopt the CRM system. As the importance of CRM pertaining to management of relationships with customers has been recognized, meanwhile, with regard to implementation of the extensive software which involves investment of massive human resources, capital, and time, many critical decisions still need to be concerned. This research aims at extracting relevant factors affecting the adoption process and proposes a convincing framework verified by an empirical case study. The research in this study is based on the Innovation Diffusion Model of Rogers (1983) and divides factors affecting the CRM adoption process into knowledge and persuasion stages with environmental and organizational factors in the knowledge stage, and perceived characteristics of innovation in the persuasion stage, while system suppliers and consulting firms as joint associates. Environmental factors include competition intensity, demand uncertainty, and industry changing speed. Organizational factors consist of size, structure, and managerial attitude. Perceived characteristics of innovation are composed of compatibility and system characteristics. Data collection, data storage, data mining, and data visualization work as independent variables. Six companies in securities, life insurance, cellular telephone system, electronics and computers, and the Internet service provider industry were selected as study cases. However, the research frame was revised after investigation. In the knowledge stage, complexity is extracted and integrated with product attribute, customer base, and corporation dependence on sales representative to form one integrated factor called business. Moreover, the independent variables are amended to be more extensive, including adoption activeness, constructing section, and constructing manner. The research shows that environmental and organizational factors affect adoption activeness and priority; and that the business factor influences adoption activeness, adoption priority, as well as outsourcing decisions. Project management and managerial participation representing a CRM constructing section are affected by perceived characteristics of the innovation. Through the integration of theory and empirical data , this research hopes to provide direction for examining the CRM adoption process and organization design, so as to facilitate the fulfilling of the adoption objective.

顧客關係管理（Customer Relationship Management，縮寫為CRM）是近來十分熱門的觀念，而關係行銷是此觀念在行銷上的實際應用。關係行銷的理論認為如果企業能蒐集顧客的詳盡資訊，並據此開發出個人化的行銷計劃，就可以有效地提昇顧客的滿意度，使其成為企業的忠實客戶，為企業帶來長期性的收益。 　　在行銷活動中，售後服務是維持顧客滿意度的重要環節。本研究以售後服務中的產品常見問題集（Frequently Asked Questions，縮寫為FAQ）為主題，實作一個FAQ導覽與維護的雛型系統。使用者可以在系統的協助下，瀏覽與產品有關的FAQ，若找不到想要的資訊，則可以提出新的問題，要求公司為其解答。 　　本系統運用代理人程式的觀念，協助FAQ導覽與新問題處理工作的遂行，並利用社群討論的方式，為顧客提出的新問題尋求解答。為了讓企業重視FAQ的維護工作，本研究試圖將FAQ的維護責任與公司的客服體制作初步的結合，其作法是將新問題的處理責任加諸於公司的員工，並由公司主管監督其工作進度。藉由這些努力，本研究希望塑造出一個良好的FAQ服務環境，使得顧客與企業能互蒙其利。 / Customer relationship management (CRM) has become a very popular concept recently. Relationship marketing is one of its application in marketing area. The relationship marketing theory claims that collecting detailed information about customers and developing customized marketing plans for individual customers can improve customer satisfaction and loyalty significantly, and organizations will benefit from this strategy in the long term. 　　Post-sale service (or after-sale service) is a vital part of marketing activities for maintaining customers' satisfaction. This study focuses on product FAQs (Frequently Asked Questions) in the post-sale service category and implements a prototype system which would allow FAQ navigation and maintenance on the Internet. Users can browse product-related FAQs with assistance from the system. If they cannot find satisfying information, they could issue new requests, and the organization would then resolve them. 　　The system applies software agents to assist the process of FAQ navigation and answering of new questions. Customers' requests can also be solved by a community-discussion approach. To force organizations to pay more attention to post-sale service, this study tries to incorporate FAQ maintenance into customer services by requiring employees to answer any raised questions, and allowing the managers to monitor the performance of those employees. Through such efforts, this study establishes a FAQ service environment and creates a win-win situation for both customers and organizations.

台灣汽車市場從1990年起便邁入成熟期，年銷售量約40~55萬台，每公里道路的汽車數155.9台，遠高於美、日等國。由於國內經濟成長趨緩、地狹人稠，未來市場擴充性有限，12家汽車製造廠間的競爭激烈可期。而2002年起台灣加入WTO後，國產車業者受到關稅降低、進口地區限制取消等措施的影響，生存空間飽受威脅，但台灣汽車業者的因應之道卻相當有限，因為產品原始設計控制在國外技術合作對象手中，自己可以設計改型的空間及能力有限，在生產和配送上的效率也幾乎發揮到極致，最後的競爭優勢應該在行銷及服務方面。所以近幾年來，部分汽車業者便積極導入顧客關係管理，希望將過去以產品為主的思維模式轉為以顧客服務為主的思維模式，和顧客建立長期且親密的關係，以提高顧客滿意度。 　　本研究針對Mitsubishi、Toyota、Nissan等台灣前三大汽車品牌，瞭解同一品牌之製造商與經銷商如何一起推動顧客關係管理，各汽車品牌業者在推動顧客關係管理之異同點，以及造成此異同點之原因等。根據國內外相關文獻的論述，本研究認為企業所重視的顧客生命週期階段、整個服務傳送系統內所有成員的權力關係，以及組織特有的資源條件，均可能影響該企業或整個服務傳送系統之顧客關係管理的規劃內容，因此透過對國內三大汽車品牌業者的觀察與訪談，歸納出下列發現： 　　一、台灣汽車業者已處於企業穩定發展階段或是產品成熟期的市場，使得顧客關係管理偏重在維繫及強化既有的顧客關係，一般是在核心產品之外，提供可以促進交易便利的顧客服務及降低顧客成本的產品組合。 　　二、由於顧客資料記錄不完整且維護不易，難以透過統計分析找出隱藏在龐大顧客資料背後的知識，使得台灣汽車業者較少投入在維持有價值的顧客或是重新找回曾離開過的顧客，也就是說較少導入行銷方面的顧客關係管理功能。 　　三、顧客關係管理的投資報酬率並非短期立現，但為了儘速看到成果，以強化主事者的推動決心，台灣汽車業者多半從與顧客互動程度較高、較容易看到改善成效的銷售及顧客服務方面的顧客關係管理項目著手。 　　四、汽車業者擁有較長的利益持續期間及服務傳送持續期間，使得單純的價格誘因，或是財務結合差異化服務的社交方式都不易具有維繫長期顧客關係的效果，因此台灣汽車業者會傾向運用財務、社交與結構結合的方式，也就是將資訊技術與整個服務傳送過程結合起來，建立使用習慣而提高顧客忠誠度。 　　五、台灣汽車業者在推動顧客關係管理時，往往涉及整個體系的組織、作業流程、資訊系統的變革，因此由該體系內權力基礎較豐厚的製造商或代理商負責整體規劃與導入，不過該體系製造商或代理商所擁有的相對權力優勢，則會影響整體顧客關係管理計畫的順利導入與否。 　　六、各汽車業者賴以成功的策略重點（e.g.生產或行銷）、財務能力、流程整合等資源條件，也會影響顧客關係管理的整體規劃方向及計畫的順利導入與否。 / Taiwan's auto industry has come to maturity from 1990. Limited by the lower economic growth rate and the heavier vehicle density, 12 automakers are in fierce competition with others for market share. In addition, Taiwan's auto industry is confronted with competition force by global autos since Taiwan being the membership of WTO in Jan, 2002. Considering the limited capability in design and the limited improvement in production and distribution, Taiwan's automakers realize that marketing and servicing are the keys to survive and success in the long run. Nowadays, some of the Taiwan's automakers and dealers are engaged in adopting CRM system to improve the customer satisfaction. 　　This research is empirical analysis of the Top3 auto brands in Taiwan. It aims at the study of how automakers and dealers under the same brand cooperating with others in adopting CRM, and what and why the differences the three auto brands having in adopting CRM. After reviewing some literatures, this research divides factors affecting the CRM adopting into customer life cycle stages which firms emphasize, the interfirm power in channel relationships, and firms' specific resources. 　　The findings in this empirical analysis are as follows： 　　1.Facing the stable and maturity stage for the auto industry, Taiwan's automakers and dealers focus on the sustenance and enforcement of customer relationships in adopting CRM system. They prefer to deliver the product mix with convenient customer services or lower searching cost not included in the core product. 　　2.Because of some unavoidable faults in customer records, automakers and dealers can't easily find specific information from data warehouse. It makes Taiwan's automakers and dealers less focus on the maintenance of valuable customers or winning back customers who have defected or are planning to. Maybe we can say that Taiwan's automakers and dealers less focus on the marketing in adopting CRM system. 　　3.It takes long time to show the benefit of adopting CRM system, so Taiwan's automakers and dealers prefer to setup sales and services supporting functions in adopting CRM system that heavily interact with customers and can easily show the improvement on customer relationships. 　　4.Auto industry offers the products and services with long duration of benefits and long duration of service delivery, automakers and dealers would like to bundle the functions of finance, society and structure. They link IT and service deliver processes to build customer-using model, and finally increase the customer loyalty. 　　5.The channel members with power advantage will take charge of the planning in adopting CRM system. In Taiwan's auto industry, makers or agents own the power advantages. In addition, the relative power advantages that makers or agents own influence the schedule of auto brand in adopting CRM. 　　6.Success strategy for some time past, financial ability, intergrated process, etc., also influence the schedule of auto brand in adopting CRM.

La gestió de la relació amb el client (CRM) és el procés en què es creen relacions amb els clients i altres actors al llarg de les seves vides. Així, a la base d'aquest procés hi ha el concepte de CLV, conegut com el valor del cicle de vida del client (customer lifetime value), que permet calcular tant el valor net del client com la seva rendibilitat per a l'empresa. Aquesta tesi destaca el fet que el màrqueting teòric ha desenvolupat diferents camps d'aplicació per afrontar el procés de CRM. Un d'aquests, que cada vegada és més important, és la gestió de la relació amb el canal de distribució. Aquesta tesi és un intent per cobrir el buit que hi ha en la recerca de màrqueting sobre aquest tema i esbrinar el veritable paper del CLV en un context concret d'aplicació: les empreses de consum de masses de Lima i els seus distribuïdors. Aquest camp resta gairebé inexplorat i ha estat poc investigat, malgrat que el procés de CMR en el mercat de masses i la relació amb els distribuïdors constitueixen importants camps d'aplicació per al CLV. A partir de l'anàlisi de casos de tres diferents empreses peruanes, mitjançant entrevistes i observacions realitzades al seu personal, aquesta tesi investiga l'aplicació del CLV com a concepte clau i com a eina de mesurament en els processos de gestió de la relació (CRM) amb els seus canals de distribució o intermediaris. L'estudi s'ha dut a terme al Perú perquè aquest país té unes característiques similars als altres països llatinoamericans i la regió gairebé no disposa de cap recerca en CMR. Per això, aquest treball espera proporcionar informació significativa sobre l'aplicació del concepte dogmàtic de CMR. La tesi argüeix que, malgrat que l'evolució cap a una aproximació relacional del màrqueting s'ha estès, la pràctica encara difereix de la teoria en molts mercats, com s'ha observat a les tres empreses estudiades. Per aquest motiu, aquest treball de recerca pot constituir un punt de partida per a investigacions més profundes en la matèria. / La gestión de la relación con el cliente (CRM) es el proceso en el que se crean relaciones con los clientes y otros actores a lo largo de sus vidas. Así, en la base de ese proceso se encuentra el concepto del CLV, conocido como valor de ciclo de vida del cliente (customer lifetime value), que permite calcular tanto el valor neto del cliente y su rentabilidad para la empresa. Esta tesis pone resalta el hecho de que el marketing teórico ha desarrollado diferentes campos de aplicación para encarar el proceso de CRM. Una de estas, y cada vez más importante, es la gestión de relación con el canal de distribución. Además, puesto que este tema constituye un vacío en la investigación de marketing, esta tesis constituye un intento para reducirlo y averiguar el verdadero papel del CLV en un contexto particular de aplicación: las empresas de consumo masivo de Lima y sus distribuidores. Este campo permanece casi inexplorado y cuenta con escasa investigación a pesar de que el proceso de CMR en el mercado masivo y la relación con los distribuidores constituyen importantes campos de aplicación para el CLV. Mediante el análisis de casos de tres diferentes empresas peruanas, mediante entrevistas y observaciones realizadas al personal de las mismas, esta tesis investiga la aplicación del CLV como concepto clave y como herramienta de medición en los procesos de gestión de la relación (CRM) con sus canales de distribución o intermediarios. Este estudio se llevó a cabo en Perú porque este país comparte características similares con el resto de países latinoamericanos y esta región casi no cuenta con investigación en CMR. Por ello este trabajo espera brindar información significativa sobre la aplicación del concepto dogmatico de CMR. Esta tesis arguye que a pesar de que la evolución hacia una aproximación relacional del marketing se ha extendido, todavía la práctica difiere de la teoría en muchos mercados, como se ha observado en las tres compañías estudiadas, por tal razón el presente trabajo de investigación puede constituir un punto de partida para investigaciones más profundas al respecto. / The customer relationship management (CRM) is the process where relationships with customers and other actors are created throughout their lifecycle. Therefore, in the very foundations of this process is CLV, a key concept known as well as customer lifetime value, that allows estimating not only the customer net value over a long period of time , but also his profitability for the company. This dissertation remarks the fact that theoretical Marketing has developed different fields of application to face the complex process of CRM. One of them, in increasing importance, is the CRM with the distribution channel. Since this issue means a gap in marketing research, this dissertation constitutes an attempt to reduce it and find out the very role of CLV when CRM takes place in a particular context of application: massive consume companies from Lima and their distributors. This field is almost unexplored and has little research despite the fact that CRM process in the massive market and the relationship with distributors constitute important fields of application for the CLV. Through an analysis of three cases of different Peruvian companies, by interviewing their staff, this dissertation investigates the application of CLV as a key concept and as a measuring tool for CRM process with the distribution channel or intermediaries. This study took place in Peru because this country has similar characteristics with the rest of Latin American countries and this region has almost no research in CRM process. That is why this dissertation expects to add significant information about the application of the dogmatic concept of CRM. This dissertation argues that despite the fact that an evolutional shift to a Marketing relationship approach is extended around the world, theory and practice still differs each other in a number of markets, as it is observed in the companies selected to this research. Thus this study could be the starting point of more and deeper research about it.

The procedures of the study are mainly through the review on the relevant reference documents related to the subject of those have been published domestically and internationally, meanwhile to get a better understanding on the most recent development on Information Technology and its associated models which are being applied to the customer relationship management by the organizations and/or business units in the world. Then, through the investigation and review on steel industries to understand the real situations domestically and internationally, specifically to focus on the Sheng Yu Steel Co., Ltd. As an example, for which to plan the implementation of Extranet System for the improvement of Customer Relationship Management (CRM) by using Information Technology, what problems it will be faced and how is the best strategyfor this implementation. Then after the completion of the study, to make a constructive recommendation to the Sheng Yu Steel Co., Ltd as an useful reference when the Extranet system to be planned and implementedfor. In Chapter-1 ¡©Introduction¡ª¡Ato descibe the backgroud, purposes, procedures and the limitation related to the study¡Fin Chapter-2, the major is concentrate on the review and understand on the relevant documents of the subject related¡Fin Chapter-3, specifically to focus on the existing situations of the steel industries and the Sheng Yu Steel Co., Ltd. how it is the current situation to the subject issue¡Fin Chapter-4, to have further investigation on recent situations for the use of Information Technology in Sheng Yu Steel Co., Ltd. to make a recommendation for their reference whn they plan for the implementation of Extranet System for the improvement of Customer Relationship Management¡Fand finally, in Chapter-5, to describe the findings and issues for whicht to be expected to have further study in the future to come up a generalized model for which can be applied to all companies not only for steel industies but also to all conventional enpterprises as an important and valuable reference guideline.

和顧客保持長期關係，對於企業經營是有利的，然而，由於資訊科技的進步與物流效率的提升，顧客越來越來越聰明，要尋找並取得替代產品與服務也越來越容易。在這樣的環境前提下，如何使顧客願意與企業維持長期關係而不轉向競爭者，是企業的一大挑戰。在文獻探討中本文發現，其關鍵是在於要能夠持續與顧客互動並了解顧客，並且將這些了解應用於產品設計、流程設計與組織設計。對於顧客的了解，本文定義其為顧客知識，就成為了顧客關係的基石。不過，由於與顧客的接觸點越多，了解顧客的管道就越多，其整合上有一定的難度，而資訊科技與傳統行銷方法都能夠產生對於顧客的了解，再加上產生並應用對於顧客的了解，以塑造顧客在消費與使用產品與服務時能得到高度的滿意，已經企業所有部門的責任，這樣跨部門的作業也不是一件簡單的事，這些議題都是企業在顧客知識管理上所需面對的。 在過往的文獻中所談的顧客知識管理，很多都混淆顧客資訊與知識，較無探討企業如何加工與顧客互動所得到的資訊，因此無法清楚了解企業在顧客知識管理上所作的貢獻為何。本文就以流程觀點，並以目前市場需求飽和且競爭激烈的人壽保險產業中的一家企業：保誠人壽為例，探討企業內的顧客知識流程，並且探討主要的影響因素，以了解企業在顧客知識管理上所作的努力。 而在本文個案的探訪後，本文還嘗試以服務業作業系統的角度，提出一種顧客知識的新分類：前台作業所需的顧客知識、後台作業所需的顧客知識，並且根據個案中保誠人壽的實際做法，修正文獻探討所提出的顧客知識管理架構，並且提出三種顧客知識管理的類型以及提出其管理重點，以作為實務界與後人研究的參考。 / It is beneficial for an enterprise to keep long-term relationship with its customers. However, owing to advances of information technology and the improvement of logistics efficiency, it is not difficult to search and acquire better products and services to substitute for what they are using now. Under these circumstances, it is a challenge for enterprises to retain their customers. How to do this? The answer in literatures is: to gain more customer insights (defined as customer knowledge in this study) through interactions with customers and apply them in product design, process design and organizational structure design. Therefore customer knowledge is the basics of long-term relationships with customers. But still, there are issues. First, an enterprise may have many contact points with its customers. Then, there are traditional ways versus information technology ways to generate customer knowledge. Finally, an enterprise has to think how to address appropriate customer knowledge to right departments and use them to enhance the total experience when interacting with customers. These are all issues an enterprise has to face with when trying to do “Customer Knowledge Management” (CKM). Most of past studies on CKM did not tell the difference between information and knowledge, thus can not clearly define what contributions an enterprise makes when implementing CKM. This study attempts to investigate practices of CKM in an enterprise in life insurance industry, in which the demand is saturated and the competition is fierce, from a process perspective to manifest how customer knowledge is generated and applied, and what main factors are to affect CKM in an enterprise. After investigation of the case, this study proposes a new framework, from the operation system in service industry, to classify customer knowledge into two categories: customer knowledge needed for front-office operations, and customer knowledge needed for back-office operations. In addition, according to practices in this case, this study will revise the CKM process framework proposed right after the literature review, and will raise keys to successful CKM in CKM processes, for the reference for following studies.

Text in English with abstracts in Zulu and Afrikaans / Despite its substantial growth and contribution to the economy of the country, the South African retail industry is faced with various challenges, including customers’ changing buying habits and expectations, the evolution of e-commerce, the entry of giant retailers into the retail market, and the growing importance of the urbanised middle class and its potential purchasing power. This creates risks that are mostly beyond retailers’ control. Therefore, there is a general drive among retail brands to build customer loyalty in order to respond to these contemporary dynamics within the industry and, ultimately, enhance sustainability. It is noted from literature that current understanding of customer behaviour in the South African retail space is largely the result of rational behavioural measurement, while the subconscious behaviour of customers is often neglected. The study on which this thesis reports attempted to address this gap in the current knowledge base. The current study makes a novel contribution to knowledge in using complementary traditional and neuromarketing instruments in a sequential mixed method research design to measure verbal, non-verbal and physiological emotions, which indicate the key influential cognitive and emotive factors shaping shopping decisions. For a comprehensive understanding of customer behaviour, the study primarily explored the emotional antecedents of customer loyalty in the formal retail industry of South Africa. In order to explore and describe both retail brands and customers’ experiences regarding the drivers of retail brand loyalty, the empirical phase of the study adopted a sequential mixed method research design. This design comprised eight individual semi-structured interviews, which focused on and explored the fundamentals of behavioural and attitudinal loyalty within the context of a customer relationship management strategy. x Based on the findings emerging from the quantitative mall intercept survey with 300 retail customers, the study identified the motivating factors of greatest likelihood among customers that best explain retail brand loyalty, namely product quality, convenience of location, customer care, product variety and availability, as well as fair pricing. Furthermore, findings emerging from the inferential analysis (structural equation modelling, two-step cluster analysis, optimal scaling and multivariate binary regression analysis) showed that brand trust, brand affect and attitudinal loyalty are key predictors of behavioural loyalty. Additionally, customer loyalty differs depending on customer predisposition towards the major retail brands tested. In this regard, Pick n Pay, Massmart and Shoprite/Checkers recorded the most favourable perceptions among customers. As Despite its substantial growth and contribution to the economy of the country, the South African retail industry is faced with various challenges, including customers’ changing buying habits and expectations, the evolution of e-commerce, the entry of giant retailers into the retail market, and the growing importance of the urbanised middle class and its potential purchasing power. This creates risks that are mostly beyond retailers’ control. Therefore, there is a general drive among retail brands to build customer loyalty in order to respond to these contemporary dynamics within the industry and, ultimately, enhance sustainability. It is noted from literature that current understanding of customer behaviour in the South African retail space is largely the result of rational behavioural measurement, while the subconscious behaviour of customers is often neglected. The study on which this thesis reports attempted to address this gap in the current knowledge base. The current study makes a novel contribution to knowledge in using complementary traditional and neuromarketing instruments in a sequential mixed method research design to measure verbal, non-verbal and physiological emotions, which indicate the key influential cognitive and emotive factors shaping shopping decisions. For a comprehensive understanding of customer behaviour, the study primarily explored the emotional antecedents of customer loyalty in the formal retail industry of South Africa. In order to explore and describe both retail brands and customers’ experiences regarding the drivers of retail brand loyalty, the empirical phase of the study adopted a sequential mixed method research design. This design comprised eight individual semi-structured interviews, which focused on and explored the fundamentals of behavioural and attitudinal loyalty within the context of a customer relationship management strategy. x Based on the findings emerging from the quantitative mall intercept survey with 300 retail customers, the study identified the motivating factors of greatest likelihood among customers that best explain retail brand loyalty, namely product quality, convenience of location, customer care, product variety and availability, as well as fair pricing. Furthermore, findings emerging from the inferential analysis (structural equation modelling, two-step cluster analysis, optimal scaling and multivariate binary regression analysis) showed that brand trust, brand affect and attitudinal loyalty are key predictors of behavioural loyalty. Additionally, customer loyalty differs depending on customer predisposition towards the major retail brands tested. In this regard, Pick n Pay, Massmart and Shoprite/Checkers recorded the most favourable perceptions among customers. As a complement to the quantitative research phase of the study, a Self-Assessment Manikin attitude measure was applied to a sample of 120 customers. This phase of the study highlighted pertinent emotional elements as key mediators in securing sustainable retail brand loyalty. In this regard, Woolworths and Pick n Pay recorded the most favourable perceptions on four stimuli, namely quality, customer rewards, customer care and brand logos. The study also applied eye tracking and galvanic skin response techniques to a sample of 10 participants to determine customers’ psychophysiological reaction to the four stimuli. Galvanic skin response findings showed that customers were especially engaged with Spar, Shoprite/Checkers and Pick n Pay, while eye tracking findings showed customers’ higher visual engagement with Shoprite/Checkers, Pick n Pay and Spar. A key overall finding of this study was the significant role of feelings and emotions in driving customer loyalty. Retail brands lose relevance when they fail to connect emotionally with customers, and without that emotional attachment, brands can easily be replaced by generic imposters. A comprehensive understanding of customer behaviour remains a quandary for marketing and business practitioners. In an attempt to address this limitation, the current study used insights from non-verbal attitudinal and psychophysiological measures to obtain a more comprehensive behavioural understanding of retail customers. / Nakuba imboni yezitolo ezihwebayo yaseNingizimu Afrika ibonise ukukhula nokudlondlobala okukhulu futhi idlala indima esemqoka emnothweni wezwe, ibhekene nezinselelo eziningana futhi ezihlukahlukene, kubandakanya indlelakuthenga yamakhasimende eguquguqukayo kanye nalokho okulindelwe ngamakhasimende nezimfuno zawo, ukungenisa ngezinkani kwemboni yokuhweba nge-intanethi nobuxhakaxhaka bobuchwepheshe besimanjemanje, ukungena kwezitolo eziyizikhondlakhondla emakethe yezitolo ezihwebayo, kanye nokukhula kokubaluleka kwabantu abahlala emadolobheni abasesigabeni sempilo esimaphakathi kanye namandla abo okuthenga. Lokhu kudala ubungozi obuvamise ukuba ngaphezu kwamandla nolawulo lwezitolo ezihwebayo. Ngakho-ke, kunomkhankaso oqhutshwa yizitolo zonkana ezihlose ngawo ukwakha ubudlelwano namakhasimende kanye nokuqinisekisa ukwethembeka kwamakhasimende ukuze zikwazi ukumelana nezimo ezikhona esikhathini samanje embonini yezitolo ezihwebayo, futhi ekugcineni, ukuze zikwazi ukusimamisa ibhizinisi lazo liqhubeke njalo futhi likhule lidlondlobale. Kuphawulekile emibhalweni efundiwe ukuthi ukuqonda kanye nolwazi olujulile olukhona njengamanje mayelana nokuziphatha kwamakhasimende eNingizimu Afrika emkhakheni wezitolo ezihwebayo ikakhulukazi kuwumphumela wokukalwa kwezinga lokuthatha izinqumo ngobuhlakani nokwenza izinto emva kokucabangisisa, kepha-ke ngakolunye uhlangothi kuvamise ukushaywa indiva ukwenza izinto kwamakhasimende ngaphandle kokuqaphela nokucophelela futhi ngokuthathela izinto phezulu noma ngaphandle kokucabangisisa ngalokho akwenzayo. Lolu cwaningo okubikwa ngalo kulo mbiko wocwaningo luzamile ukubhekana naleli gebe elikhona njengamanje eliphathelene nokuntuleka kolwazi. Lolu cwaningo lwamanje lufaka isandla esinohlonze kanye nemibono ephusile olwazini olukhona, ngokusebenzisa amathuluzi asekelayo ajwayelekile kanye nalawo asebenzisa ubuchwepheshe bokukala ukusebenza kwengqondo, ngaphansi kwedizayini yocwaningo egxile kwindlela-kusebenza exubile yokwenza izinto ngokulandelana kwazo, ngenhloso yokukala imizwa ezwakaliswa ngomlomo, ngezenzo noma ngeminyakazo yomzimba, okuyinto ebonisa izimo nezici-bunjalo ezisemqoka futhi ezinomthelela omkhulu eziphathelene nokusebenza kwengqondo kanye nemizwa, okuyizona ezilawula izinqumo zamakhasimende eziphathelene nokuthenga. Ngenhloso yokuqonda ngokuphelele ukuziphatha kwamakhasimende, lolu cwaningo lwacubungula futhi lwahlaziya ikakhulukazi izici-bunjalo zemizwa elawula ukwethembeka kwamakhasimende embonini yezitolo ehlelekile yaseNingizimu Afrika. Ukuze kucutshungulwe futhi kuchazwe kokubili izinhlobo zezitolo kanye nezimo ahlangabezane nazo amakhasimende mayelana nezici-bunjalo ezenza amakhasimende athembeke esitolo esithile futhi athenge kusona ngaso sonke isikhathi, isigaba socwaningo esiphathelene nolwazi oluphathekayo nolubonakalayo sasebenzisa idizayini yocwaningo egxile kwindlela-kusebenza exubile yokwenza izinto ngokulandelana kwazo. Le dizayini beyibandakanya ama-semi-structured interviews enziwa nababambiqhaza bocwaningo abayisishiyagalombili, okuyizingxoxo ezabe zigxile ekucubunguleni ukwethembeka kwamakhasimende okuphathelene nokuziphatha kwawo kanye nendlela abuka ngayo izinto ngaphansi kwesimo seqhingasu lokuphathwa kobudlelwano namakhasimende. Ngokususela kwimiphumela eyatholakala kwinhlolovo ebandakanya abathengi abangama-300 okwaxoxiswana nabo ezikhungweni ezinenxanxathela yezitolo, ucwaningo lwahlonza izimo nezi-bunjalo ezikhuthaza futhi zihehe amakhasimende okuyizona ezingasetshenziswa ukuchaza kangcono ukwethembeka kwamakhasimende ezitolo ezithile, okuyikhwalithi yomkhiqizo, ukufinyeleleka kalula kwendawo lapho kutholakala khona isitolo, ukunakekelwa kwamakhasimende, imikhiqizo ehlukahlukene kanye nokutholakala kwayo, ngokunjalo namanani-ntengo angambi eqolo. Ngaphezu kwalokho, imiphumela eyatholakala ohlaziyweni lwezinqumo ezithathwe ngokubheka ubufakazi kanye nokuqonda okujulile (structural equation modelling, two-step cluster analysis, optimal scaling kanye ne-multivariate binary regression analysis) yabonisa ukuthi ukwethenjwa kohlobo lomkhiqizo, ukukhanga nokuthandeka komkhiqizo kanye nokwethembeka okuphathelene nendlela-kubuka kuyizibikezeli ezisemqoka zokuziphatha kwamakhasimende okuhambisana nokwethembeka emikhiqizweni ethile. Futhi ngaphezu kwalokho, ukwethembeka kwamakhasimende kuyehluka, futhi lokhu kuncike ekuhehelekeleni kwamakhasimende ezitolo enkulu ezihloliwe. Mayelana nalokhu-ke, abakwa-Pick n Pay, Massmart kanye nabakwa-Shoprite/Checkers yizona zitolo ezibonise ukuthandwa kakhulu ngamakhasimende. Njengendlela yokweseka nokunezezela phezu kwesigaba socwaningo olukhwantithethivu olwenziwe, kwasetshenziswa isikali sendlela-kubuka esibizwa nge-Self-Assessment Manikin kwisampula yamakhasimende ayi-120. Lesi sigaba socwaningo saqhakambisa izimo nezici-bunjalo eziphathelene nemizwa ezifanelekile njengezinto ezisemqoka kakhulu ekuqinisekiseni ukwethembeka kwamakhasimende esitolo esithile ngokuqhubekayo. Mayelana nalokhu, abakwa-Woolworths kanye nabakwa-Pick n Pay babonisa ukuthandwa kakhulu ngamakhasimende mayelana nezici-bunjalo ezihehayo futhi ezikhuthazayo ezine, okuyikhwalithi, imivuzo yamakhasimende, ukunakekelwa kwamakhasimende kanye nemibala egqamile, igama elidumile nezimpawu zesitolo. Ucwaningo lwasebenzisa futhi nezindlela zokucwaninga eziphathelene nokulandela umkhondo wamehlo amakhasimende kanye nezimpawu zemizwa yamakhasimende ezibonakala esikhumbeni semizimba yawo, okuyizindlela zocwaningo ezasetshenziswa kubabambiqhaza bocwaningo abayi-10 ngenhloso yokuhlonza indlela-kusabela kwamakhasimende ephathelene nengqondo nomzimba kulezi zicibunjalo ezikhuthazayo futhi ezihehayo ezine ezishiwo ngenhla. Imiphumela yendlela yocwaningo ebheka izimpawu zemizwa yamakhasimende ezibonakala esikhumbeni semizimba yawo yabonisa ukuthi amakhasimende ahehelekela ikakhulukazi kwabakwa-Spar, Shoprite/Checkers kanye nabakwa-Pick n Pay, kanti ngakolunye uhlangothi imiphumela yendlela yocwaningo ebheka futhi ilandele umkhondo wamehlo amakhasimende yabonisa ukuheheka kwamehlo okuphakeme ezitolo zabakwa-Shoprite/Checkers, Pick n Pay kanye nabakwa-Spar. Umphumela omkhulu futhi osemqoka otholakale kulolu cwaningo yindima ebaluleke kakhulu edlalwa yimizwa ekukhuthazeni ukwethembeka kwamakhasimende. Izitolo zilahlekelwa yiwozawoza yazo kumakhasimende uma zihluleka ukuxhumana namakhasimende ngokwemizwa, futhi ngaphandle kwalokho kuxhumana ngokwemizwa nokuzwana, izitolo zingazithola kalula zisesimweni lapho indawo yazo isithathwe yizitolo-mbumbulu ezingosandanezwe. Ukuqondakala okuphelele kwendlela aziphatha ngayo amakhasimende futhi enza ngayo izinto kusalokhu kuyinkinga nesithiyo esikhulu kwizazi nezisebenzi ezenza umsebenzi wokukhangisa nokuqhuba ibhizinisi. Njengomzamo wokubhekana nalesi sithiyo, lolu cwaningo lusebenzise ulwazi nokuqonda okujulile okutholakale kwizikali zendlela-kubuka yamakhasimende kanye nezikali zesimo sengqondo nesomzimba samakhasimende ukuze kutholakale ukuqonda okuphelele futhi okujulile okuphathelene nokuziphatha kwamakhasimende athenga ezitolo kanye nendlela enza ngayo izinto. / Ondanks sy stewige groei en aansienlike bydrae tot die landsekonomie, het die Suid-Afrikaanse kleinhandelsbedryf met verskeie uitdagings te kampe. Veranderinge in kliënte se koopgewoontes en verwagtings, groeiende e-handel, reusekleinhandelaars wat die kleinhandelsmark betree, die toenemende invloed van die verstedelikte middelklas en sy potensiële koopkrag is enkele hiervan. Dit gee aanleiding tot risiko’s wat grotendeels buite die beheer van kleinhandelaars is. As gevolg van die hedendaagse dinamiek in die bedryf, poog kleinhandelsmerke om kliëntlojaliteit te verstewig en hulle eie volhoubaarheid te verseker. Dit blyk uit die literatuur dat die kleinhandelsektor se siening van kliëntgedrag in Suid-Afrika op die meting van rasionele gedrag berus, en dat die onbewuste gedrag van kliënte verontagsaam word. Hierdie tesis doen verslag van ʼn studie om hierdie leemte in die huidige kennis te probeer aanvul. Dit lewer ʼn bydrae deur aanvullende tradisionele en neurobemarkingsinstrumente in ʼn opeenvolgende gemengde metode as navorsingsbenadering te gebruik. Verbale, nieverbale en fisiologiese emosies is gemeet aangesien hulle die kognitiewe en emotiewe faktore is wat inkopiebesluite bepaal. Hierdie studie het hoofsaaklik die emosionele antesedente van kliëntlojaliteit in die Suid-Afrikaanse formele kleinhandelsbedryf verken om ʼn grondige begrip van kliëntgedrag te kry. Ten einde kleinhandelsmerke en kliënte se ervarings van die drywers van kleinhandelsmerklojaliteit te beskryf en te verken, is ʼn opeenvolgende gemengdemetode-navorsingsontwerp gevolg. Dit het uit agt halfgestruktureerde, private onderhoude bestaan. Hierin is gekonsentreer op die grondslae van gedragsen houdingslojaliteit teen die agtergrond van ʼn kliënteverhoudingsbestuurstrategie. Op grond van die bevindings van ʼn kwantitatiewe winkelsentrumopname onder 300 kleinhandelskliënte, is die motiveringsfaktore met die grootste waarskynlikheid aangetoon. Hulle is produkgehalte, ʼn gerieflike ligging, kliëntesorg, produkverskeidenheid, beskikbaarheid en billike pryse, en bied die aanneemlikste verklaring vir kleinhandelsmerklojaliteit. Volgens die bevindings van die inferensiële ontleding (struktuurvergelykingmodellering, tweestaptrosontleding, optimale skalering en meerveranderlike binêre regressieanalise) is merkvertroue, merkaffek en houdingslojaliteit betroubare aanwysers van bedragslojaliteit. Afgesien hiervan verskil kliëntlojaliteit na gelang van kliëntpredisposisie jeens kleinhandelsmerke wat getoets is. Kliënte was vir vier stimuli (naamlik gehalte, kliëntbelonings, kliëntesorg en merklogo’s) die ontvanklikste vir Pick n Pay, Massmart en Shoprite/Checkers. Aanvullend tot die kwalitatiewe navorsingsfase van hierdie studie is ʼn selfasseringsmodelhoudingmeting op ʼn steekproef van 120 kliënte toegepas. Hierdie fase het die aandag gevestig op pertinente emosionele elemente as sleuteltussengangers in die opwekking van volhoubare kleinhandelsmerklojaliteit. Vir die vier stimuli (naamlik gehalte, kliëntbelonings, kliëntesorg en merklogo’s) was kliënte se persepsies van Woolworths en Pick n Pay die gunstigste. Hierbenewens is oogvolg- en galvaniese velreaksietegnieke op ʼn steekproef van 10 deelnemers toegepas om hulle psigofisiologiese reaksie op die vier stimuli te bepaal. Die galvaniese velreaksietoetse het aangedui dat kliënte besonder ontvanklik was vir Spar, Shoprite/Checkers en Pick n Pay, en volgens die oogvolgtoetse visueel meer aangetrokke tot Spar, Shoprite/Checkers en Pick n Pay was. Hierdie studie het bevind dat emosies in die algemeen ʼn deurslaggewende rol in kliëntlojaliteit speel. Kleinhandelsmerke raak irrelevant as hulle nie ʼn emosionele band met kliënte smee nie. Sonder hierdie emosionele verbintenis kan generiese bedrieërs maklik die plek van kleinhandelsmerke inneem. Bemarkings- en sakepraktisyns het nie ʼn deeglike begrip van kliëntgedrag nie. Om dit reg te stel, het hierdie studie nieverbale houdings- en psigofisiologiese toetse toegepas om die gedrag van kleinhandelskliënte beter te begryp. / Business Management / D. Com. (Business Management)