Summary of Buono et al.’s “Towards the Detection of UX Smells: The Support of Visualizations”

TO: Prof. Ellis
FROM: Angela Hernandez
DATE: 3/3/2021
SUBJECT: 500-Word Summary of Article About Usability Smells

The following is a 500-word summary of a peer-reviewed article about the use of visualizations to detect usability smells.  The authors discuss the four methods of visualizations that they developed to identify usability smells in websites by conducting a study, collecting and evaluating the data.  According to the authors, “The proposed visualizations apply and customize existing visualization techniques, which are here used with the novel purpose of providing usability smells to evaluators” (Buono et al., 2020, p. 6902).  User satisfaction with software products is largely influenced by UX attributes.  Poorly designed usability can make it harder for users to navigate an interface.  Many methods can be used to evaluate systems but are rarely implemented by developers for a variety of reasons.  Research done on e-government websites has shown that developers need to be provided with the tools and methods to effectively evaluate and implement usability.  ‘Usability smells’ and ‘Code smells’ respectively indicate weaknesses in the design of an interface or code that can cause problems in the future.  The results reported by the study provide usability evaluators with the tools they need to detect usability smells.  There are two methods used for evaluating usability; User-based methods and analytical methods and several tools available to assist in the different stages of usability testing.  The tools available that provide visual representations help evaluators understand user behavior.  Graph-based structures are commonly used to visualize website navigation or general navigation paths.  The four graph-based structures that are used in this study are the following; Arc Diagram, Word tree, Sankey Diagram and Node-Link.  Each graph-based structure uses the same visual encoding, primarily nodes that are used to reveal usability smells.  In a study done in March 2019, 15 users were tasked with visiting a webpage.  In order to visit the specified webpage, users could take an optimal path to successfully complete the task but because the website has since been updated, the steps taken to accomplish the task might not be the same.  Scalability is not an issue when performing these tests because many of these tests involve a low number of participants executing simple tasks.  Participants were given a booklet composed of the four visualization techniques and tasks to be completed with each technique.  The facilitator introduces and explains the first visualization technique and the participant begins performing the tasks.  In order to check the overall research methodology, the procedure has been assessed by a pilot study.  Researchers created an excel file for each task performed in order to evaluate the support provided by the visualizations.  Two well-known questionnaires were used to evaluate satisfaction with each visualization technique and repeated measures are used to assess the significant differences in the four visualization techniques.  The data collected shows how well the visualization techniques provide support to the evaluators in identifying usability smells.  Despite there being some confusion about the paths certain visualization techniques provided, evaluators were able to detect the paths that led to task failure.  There were no differences between the four visualization techniques in terms of evaluator satisfaction.  

Reference

Buono, P., Caivano, D., Costabile, M. F., Desolda, G., & Lanzilotti, R. (2020). Towards the detection of ux smells: The support of visualizations. IEEE Access, 8, 6901-6914. https://doi.org/10.1109/ACCESS.2019.2961768