Using Voice and Biofeedback to Predict User Engagement during Product Feedback Interviews

Overview

• This paper explores the use of biometric data, including physiological and voice features, to measure user engagement during interviews about software products.
• The researchers conducted interviews with users while collecting biometric data using an Empatica E4 wristband and a laptop's audio recorder.
• The goal was to determine if biometric data could be used to complement traditional user feedback gathered through interviews, particularly in cases where in-person feedback is required.

Plain English Explanation

When developing software, it's important for companies to understand how users engage with their products. Traditionally, this feedback has been collected through techniques like product reviews and social media analysis. However, these methods may not work well for bespoke software or when companies need feedback from specific users.

In these cases, companies often resort to face-to-face interviews to get direct feedback. But the researchers behind this paper wanted to see if they could gather even more useful information by looking at users' physiological and vocal responses during these interviews.

They had users wear a special wristband that could measure things like heart rate and sweat levels, and also recorded the users' voices. The idea was that this biometric data could reveal insights about the users' emotional engagement and reactions to the software being discussed.

The researchers found that they could use machine learning algorithms to accurately predict a user's level of engagement based on the biometric data, especially the voice features. This could be really helpful for requirements gathering and emotion-aware software design, both for human analysts and potentially even voice-based chatbots.

Overall, the study suggests that incorporating biometric measurements could be a valuable addition to traditional user feedback methods, particularly for software development projects where direct user input is crucial.

Technical Explanation

The researchers conducted a study where they interviewed users about software products while simultaneously collecting biometric data. This included physiological measurements from an Empatica E4 wristband, such as heart rate, electrodermal activity, and temperature, as well as audio recordings of the users' voices.

The goal was to see if this biometric data could be used to predict the users' level of engagement with the software being discussed. The researchers trained supervised machine learning models on the biometric data and found that they could achieve an F1 score of 0.72 in predicting user engagement.

Interestingly, the voice features alone were found to be nearly as effective, with an F1 score of 0.71. This suggests that voice analysis could be a particularly useful tool for measuring user engagement, especially in remote communication settings like online meetings or when using voice-based chatbots.

The researchers believe this approach could be valuable for requirements engineering and emotion-aware software design, allowing companies to gather deeper insights about user engagement than traditional feedback methods.

Critical Analysis

The researchers acknowledge several limitations to their study. First, the sample size was relatively small, with only 24 participants. Larger-scale studies would be needed to further validate the effectiveness of the biometric approach.

Additionally, the study was conducted in a controlled lab setting, which may not fully reflect the real-world conditions and distractions users face when interacting with software. More research is needed to understand how well the biometric measurements hold up in more realistic environments.

Another potential concern is the privacy and ethical implications of collecting such intimate biometric data from users. The researchers did not delve deeply into how this data would be stored, secured, and used, which could be an important consideration for companies looking to adopt this approach.

Despite these limitations, the researchers have made an interesting contribution by demonstrating the potential of biometric data, especially voice analysis, to enhance traditional user feedback methods. As software development becomes increasingly user-centric, tools that can provide deeper insights into user engagement and emotional responses could prove invaluable.

Conclusion

This study explores the use of biometric data, including physiological measurements and voice analysis, to complement traditional user feedback gathered through interviews about software products. The researchers found that machine learning models trained on this biometric data could effectively predict users' engagement levels.

The ability to use voice features alone to measure engagement is particularly noteworthy, as it could enable emotion-aware requirements gathering and support the analysis of online meetings and interactions, whether conducted by human analysts or voice-based chatbots.

While further research is needed to address the limitations of this study, the findings suggest that incorporating biometric measurements could be a valuable addition to the software development toolkit, especially for projects that require deep user insights and engagement.