Researchers decode the science of scroll stop travel videos

In an age where data is currency and attention spans are fleeting, successful brands aren’t just targeting consumers. From personalized ads that seem to read your mind, to campaigns that are perfectly timed to everyday life, they use behavioral insights to shape decisions before consumers realize they are making.

By combining data, psychology and evolving digital platforms, brands are learning not only who their customers are, but what they will do next. Understanding how people think, feel and act has become the most powerful tool in marketing. And it’s completely changing the game.

At Anderson School of Management at the University of New Mexico, Hyun-Sang’s son (son) assistant professor uses machine learning technology to explore how marketers can use social media data to their advantage. His recent research, published in Tourism Management, focuses on influencer marketing in the tourism industry, using Instagram as a case study to determine what kind of videos will most promote user engagement.

This study analyzed over 4,000 Instagram videos from the top 40 travel influencers using a technique known as The Transfer Learning Approach. This is a model used by high-tech giants such as Google and Facebook. This method allows researchers to extract features from the video, such as video length, emotion, facial expressions, color saturation, and brightness.

“I can actually quantify the exact level of contribution for each feature into the number of likes, comments and comments,” my son said. “This is a combination of functional extraction to estimate effects and computer vision of traditional econometric modeling.”

He compares the transfer learning approach to the teacher-student relationship. “Teachers” are models that companies like Google train on large datasets, whereas “students” learn from that model and apply insights to smaller, more specific datasets. Without transfer learning, the accuracy of classifying content could be below 10%, says my son. But that could raise that rate to an impressive 96%.

The purpose of this study is to identify what type of video content actually attracts users. Using the automatic video feature extraction process, the team found some amazing results.

Short form videos performed better than long videos. My smile was too big and my engagement was reduced. Emotionally expressive content, including anger and even sadness, increased involvement. Product size visibility had a positive effect on user interaction. Brightness contributed to active involvement, but the contrast and rapid scene changes too much and the performance is compromised. Overuse of hashtags has resulted in lower engagement levels. Emojis and posting text also influenced user responses when used strategically.

“We wanted to see what kind of factors influence user engagement. Clicks, likes, comments,” my son said. “This data helps influencers and industries better understand how to create content.”

The findings suggest that marketers can deliver content creators more than just creative direction. For example, creators can receive recommendations on how to express their behavior, emotions, or place products on videos to increase engagement.

Son’s Research has practical applications for local and national organizations. An example is the New Mexico tourism initiative. Using similar machine learning techniques, marketers can visit Albuquerque, analyze their own social media content and identify variables such as videos, landscape shots, and the number of themes.

But my son warns against excessive generalization. “Different datasets produce different results,” he said. “Variables such as brightness, emotion, contrast, and product visibility can all affect outcomes in a unique way.”

One particularly interesting takeout is that emotions traditionally considered negative fear, sadness, and even anger can promote higher engagement than neutral content. This supports previous research showing that emotional arousal plays a major role in virality, whether positive or negative.

This research also leads to a broader theory of marketing. One is the theory of media richness, suggesting that richer media, such as videos with expressive emotions and visual details, produce more engagement than text-only content. Factors such as brightness, saturation, and contrast can define this richness and influence audience responses.

“As a marketer, I can now provide formulas for influencers. I can avoid smiles, emphasize product sizes and use bright visuals,” my son explained. “Short videos tend to work well. Are there too many hashtags? It’s not helpful.”

My son believes this study shows a turning point for how marketing strategies are shaped. Machine learning has long been in the realm of computer science, but he says that human behavior and marketing applications have great potential.

“The key role of our research is to come up with ways to apply these algorithms to better understand people,” he said. “That’s where marketing and business research can make a real difference.”