Development of a decision tool for cost justification of usability
MetadataShow full item record
This dissertation follows a three paper format. Abstract for each paper is presented below. • First Research Study A Methodology for Data Collection and Analysis to Assist Cost Justification of Usability Abstract Usability investments help companies increase benefits and reduce costs. There are many categories of increased benefits and reduced costs mentioned in the cost justification literature. It is hard for usability practitioners to justify usability through all these categories. This paper examines the significant gains from usability that practitioners should focus on based on their company. These can be listed as: (1) increased sales (2) reduced error rates, (3) reduced task times, and (4) increased traffic on website. The study involves literature review, Pareto analysis, distribution fitting and bootstrap sampling. • Second Research Study Goal Programming Model for Usability Benefit Expectations The key objective of this study is to optimize the usability benefit expectations. The fact that different types of companies expect different benefits is taken into consideration. The paper is consisted of two parts: (1) an optimization model for the assumptions of percentage increase in sales and percentage increase in website traffic, (2) an optimization model for the assumptions of percentage change in error rate and percentage change in task time. The optimization model used is goal programming. Finally, a sensitivity analysis is performed to test how sensitive the results are to small changes in the constraints and the variables. • Third Research Study Determining Constraints of Goal Programming Model for Usability Benefits The objective of this study is to analyze the interaction constraints of the goal programming models developed by Aydin and Beruvides (2014b). In Aydin and Beruvides (2014b), two goal programming models were developed for increased sales and traffic rates, and decreased error rate and task time benefits. Ordinary least square regression method was used two determine the two-way interactions. Equal slope was assumed for all quantiles of the dependent variables. The results of this study indicated that there were not any significant differences in the slopes of the regression lines depending on the quantiles. Thus, this research proved that the approach taken in Aydin and Beruvides (2014b) was accurate in terms of interaction constraints used in the models.