1.1 Research Background and Motivation Nowadays in the global market, many companies face high levels of competition. When a company introduces a new product to market, its competitors will always introduce some similar products which may have stronger and better functions. How a company can capture a better opportunity to promote its products depends on whether it can effectively and efficiently advertise its new products. Whenever a company wants to advertise its new products; one of the important factors that have to be considered is selecting the advertising agency. Once company chooses the wrong agency it can cause a serious problem. It is very difficult to recover from agency selection mistakes. If a wrong agency is chosen, the company will probably have to live with the consequences for decades. These are complex decision-making processes that have a capability that places the right agency in the right jobs at the right time. Many research studies have defined the advertising agency into multidimensional concept, most studies focused on advertising agency-client relationship, not the evaluation and selection of the agent. Choosing the best advertising agency is not that easy, so many criteria need to be evaluated. It is not just to talk about cost or the good reputation, but there are some of qualitative criteria, such as creativity, chemistry, techniques, and communication skills etc, which have to be considered and cannot be properly quantified. Besides, the criteria that should be considered may have different important weights in the perspective of different decision maker. As a result, this study proposes fuzzy multiple levels multiple criteria decision making (MLMCDM) model that can be a very suitable approach for evaluation and selection of advertising agency. This research will develop an n-level hierarchy structure for criteria which extend the work done by Chu (2007) to obtain the final ranking value of each alternative. 1.2 Problem Identification Advertising agency selection is a complex decision making process. It is basically a problem in identification, weighting, and evaluation. The company must first identify all criteria which are perceived to be significant in evaluating potential candidates. These will vary somewhat from one hiring entity to the next. Once the criteria have been selected, they must be weighted. Someone must decide whether experience is more important than education, and whether creativity is more or less important than a proven track record. This problem is complicated by the fact that some of the criteria are completely quantifiable and others completely subjective. Besides, criteria may be quantifiable but in different units, so that they cannot be conveniently compared. After that, each potential candidate also must be rated for each identified criteria. This evaluation is subject to some of the same difficulties encountered in assigning weights (Taylor, Ketcham-, and Hoffman, 1998). The advertising agency evaluation and selection are generally and theoretically introduced; but in many advertising agency research studies, most studies focused on advertising agency-client relationship, not the evaluation and selection the agency. One of the researches that focused on evaluating and selecting the agency is the research of Yuksel et al. (2007). They mentioned 14 criteria to select and evaluate advertising agency, but they treat all of the criteria as quantitative criteria. Besides, the common approaches for this issue, which is statistic, are not really effective to deal with qualitative criteria and multiple level multiple criteria decision making problems. Traditional approaches such as statistics considering only exact crisp values, this method cannot well-support decision-making procedures for such evaluation problems (Ou et al., 2009). Therefore, fuzzy multiple level multiple criteria decision making (MLMCDM) can be a highly suitable approach for this issue. 1.3 Research Objectives The objectives of this study are the following: To analyze criteria for evaluating and selecting advertising agency. To provide Chi Square test in order to analyze the independent/dependent criteria. To establish a fuzzy multi-level multiple criteria decision making (MLMCDM) approach for evaluating and selecting advertising agency. To conduct a numerical example to show the computational procedure and the feasibility of the proposed model. 1.4 Research Framework The framework of this research is shown in Figure 1-1. This study is divided into six chapters including introduction, literature review, criteria selection and development of multi-level hierarchy, fuzzy set theory and model establishment, numerical example, and conclusion. The introduction presents the research background and motivation which leads to problem identification and the establishment of research objectives and research framework. Literature review introduces the existing models of advertising agencies evaluation, decision criteria, nature of criteria, and hierarchy structure of criteria. Chapter three deal with the fuzzy set theory. Chapter four introduces the independent test of criteria and model establishment. Chapter five shows the feasibility of the model with a numerical example. Finally, chapter six concludes the research with managerial implications of the result of the research. Figure 1.1. Research Frameworks