This study designs an electroencephalogram (EEG) experiment to record neural signals during consumers' exposure to short video advertisements (SVAs), with the aim of exploratorily digging the frequency bands and brain regions (cerebral cortex and non-internal structures) that are linked to engagement preferences and using machine learning and deep learning methods to predict engagement behaviors.
This study selects different types of SVAs, designs an EEG experiment and chooses participants. It records the EEG data of participants while they watch SVAs as well as their participation willingness data after watching. Based on logistic regression, it explores the optimal viewing time for influencing consumer participation, analyzes the channels and frequency bands that affect consumer decision-making and predicts consumer engagement behaviors by comparing convolutional neural network and machine learning algorithms.
The experimental results reveal that consumers' SVA viewing time in the first eight seconds has the most significant impact on their engagement behaviors, while the theta band activity in central region and gamma band activity in temporal region exert the most profound influence on engagement behaviors. Notably, the prediction accuracy of consumer engagement behaviors has achieved 82.63% by using all the channels and frequency bands data.
This study demonstrates the potential of EEG for the first time in assessing consumer engagement behaviors in SVAs. Additionally, it provides a data-analytic paradigm for SVA research and offers actionable insights for advertisers crafting impactful SVA content. The neuromarketing framework established herein bridges cognitive neuroscience principles with marketing practice, enabling more evidence-based optimization of consumer engagement behaviors in social media marketing.
