Using Neural Network Method to Construct 3H Content Model

Hui-Hsin Huang

Abstract

This paper proposes a 3H (hero, hub, and hygiene) content model by neural network method to predict the conversion rate of audiences between each of the three layers. The empirical data is collected from a brand of beauty groceries store. There are134,569 audiences who subscribe to this brand’s Youtube channel over one year are involved to make parameters estimation, weight value of each attribute in each content, and calculate the probability of conversion rate. The results show that in the hub content, the weight value of community and engagement are higher than episode. It means that contributing audience or customer engaging and forming communities are important goals of hub content. In hygiene (help) content, the results show that the weight value of cultural factor is the highest value than the second highest relevant and evergreen and real time is the lowest weight value. It demonstrates that in creating hygiene (help) content, the company shows more attention on the cultural element. For the probability of conversion rate, there is a higher rate of hub content from hygiene (help) content than from hero. Finally, the conclusions are made for future application.

Abstract

Keywords