Measurement of User's Satisfaction of Digital Products through Emotion Recognition

• Ricardo Cordeiro Galvão Sant'Ana van Erven UnB
• Edna Dias Canedo UnB

Resumo

Context: Measuring user satisfaction is an essential tool to create value in digital transformation. Measurement allows you to identify future purchase intentions, user loyalty, and retention. The traditional way of measuring satisfaction using self-assessment has problems, such as subjectivity. Therefore, a more objective approach is needed that allows automatic measurement of satisfaction. Objective: This research aims to make a systematic literature mapping (SLM) to identify the state-of-the-art use of Artificial Intelligence tools and techniques related to emotion recognition visually and then use the results to propose a model that can lead to an automated user satisfaction measurement. We present the model with a work process, an exploratory experiment, and an application project to measure the user experience from visual emotion recognition, automatically calculate the satisfaction assessment of a user when using a digital product, and present the positive, neutral, and negative points of using this product. Methods: A systematic literature mapping was used to identify the primary studies related to techniques, benefits, and challenges that associate the measurement of user satisfaction through the recognition of emotions with the use of Artificial Intelligence (AI). In the proposed model, an experiment was planned in two phases to validate research hypotheses related to the objective of the work. The experiment, which will apply the projected application, will be carried out by videoconferencing with the approach of exploratory usability testing on a web product. Results: 10 primary studies were identified in different areas of knowledge: restaurants, television systems, retail stores, artistic shows, and usability testing in a call center system. Two systematic reviews of the literature were also identified. The technique most commonly used in primary studies is the convolutional neural network (CNN). The use of cloud services for emotion recognition was also verified. Benefits related to user feedback, such as user profile mapping, were reported, and challenges for emotion recognition were found, such as user privacy and capture environment inadequacies. Besides, a model to automatically measure user satisfaction in a scalable and replicable way was proposed. Conclusions: The possibilities of applying emotion recognition are countless in terms of contexts, techniques, forms, and components. Despite this, it was possible to identify good practices that could guide the creation of a tool to measure the satisfaction of using a digital product through emotion recognition and Artificial Intelligence. The main contribution of this work is the proposal of a model that guides the creation of an automatic satisfaction measurement tool to identify positive, neutral, and negative points in the use of a digital product. This research also defines a measurement experimentation process that any organization or researcher can reproduce.

Palavras-chave: User Satisfaction Measurement, Emotion Recognition, Artificial Intelligence