9.4 Images
- related to 8.3.3.2
Potential Variables:
Structural Complexity: complexity due to visual features (Pieters, Wedel, and Batra 2010). Measured by the Canny Edge Detection technique (Forsythe, Sheehy, and Sawey 2003), because edge information is highly correlated with observed complexity
Color Complexity: complexity due to variety of colors within an image (Reinecke et al. 2013). Measured by the color variety algorithm (Bing Zhou, Shuang Xu, and Xin-xin Yang 2015)
\(C = \sum_{i = 0}^m n_i \log(\frac{n_i}{N})\) where
\(m\) is the number of distinct colors
\(n_i\) the number of pixels of the i-th color
\(N\) is the total number of pixels
Image Permanence: how adept images are at remaining in users’ memory (after users are no longer looking at them) (Khosla et al. 2012, 2015). Measured based on AMNet memorability index (Fajtl et al. 2018), which has been verified by (Leyva and Sanchez 2021) (achieved near human consistency in predicting image memorability).
Number of faces: based on Google Vision API
Aspect ratio: height divided by width
Average HSV value: Average HSV values across all pixels within an image
Themes: LDA based on keywords (labels) from Google Vision API
Pieters, Wedel, and Zhang (2007) found that optimal feature design of ads (e.g., brand, text, pictorial, price, and promotion) can be achieved without increased costs. Moreover, the authors also propose two entropy-based measures of clutter effects, where they characterize the salience of feature ads based on Attention Engagement Theory.
Wedel and Kannan (2016) is a review on marketing analytics for data-rich environments. And competition for attention in these environments are more fierce.
Y. Li and Xie (2019) found a significant and robust positive mere presence effect of image content on user engagement in both product categories on Twitter, and high-quality and professionally shot pictures consistently lead to higher engagement on both platforms (e.g., Twitter and Instagram) for both product categories. However, the effect of colorfulness varies by product category, while the presence of human face and image–text fit can induce higher user engagement on Twitter but not on Instagram. And the fit between the image and the accompanying message matters
Jalali and Papatla (2016) study visual UGC. Where color compositions were operationalized as combinations level of hue, chroma, and brightness. Consumer engagement (i.e., click rate) is higher for photos that have higher proportions of green and lower proportions of red and cyan, as well as higher chroma of red and blue.