Visual Attributes of Thumbnails in Predicting Top YouTube Brand Channels: A Machine Learning Approach

With video marketing platforms growing rapidly, brands create YouTube channels to distribute their content and communicate with customers. This study proposes a prediction model analyzing 16,278 image data sets collected based on 153 brand channels generated before September 26, 2022. We analyze the factors affecting the number of image views of brand channels using the data set, and constructs a view prediction model. The study found that the characteristics of the thumbnail image, offline top brand characteristics, and the size of the channel (number of subscribers, number of channel videos) affect YouTube’s top online channel views. The results of the study make it possible to predict branded content before uploading it, and it can help the brand for planning content production and marketing activities through YouTube.