[1/2] Attention Factory: The Story of TikTok and China's ByteDance (TikTok. Attention Factory. Takeoff story) (Category BusinessStory)

This book 2020 Written by Matthew Brennan, the company’s growth trajectory, vision and competitive landscape of Chinese social media helped shape the global phenomenon that became TikTok. He is an expert on Chinese Internet technology and innovation. His opinions are often cited in publications such as Bloomberg, The Wall Street Journal and The Economist. Brennan lived in China 16 He is fluent in Mandarin and is the Managing Director of China Channel. I liked the presentation of the author and his thoughts on the development strategy of ByteDance.

The book can be divided into separate parts and begins with the fact that a significant part is devoted to the founder of ByteDance, Zhang Imin, and his strategic vision. It examines how his engineering education influenced the company’s algorithm-driven approach and how his leadership style shaped ByteDance’s corporate culture and business decisions. The narrative highlights how Zhang saw opportunity in the attention economy before many other tech leaders. Interestingly, by the time ByteDance was founded, he had already worked in several startups: travel, real estate, the analogue of Twitter.

Then he started a company. ByteDance I tried to run applications and stopped at Toutiao ("Titles"). It was launched in 2012 A news aggregator using artificial intelligence algorithms to personalize content. At this point, there were already a lot of content services on the market, but Imin, the founder of the company, realized that one of the formats was hardly occupied. There is a quote in the book about his approach. The efficiency of the information flow is the main theme of all my business activities. To do this, it was necessary to look at the model of content distribution on Chinese platforms. 2013 year. The model was a 2x2 matrix, where the axes were: active vs passive model, as well as who prepares it a person or a machine. It worked. 4 variant

1. The Passive Model and the Man Himself (moderation) - there were classic news portals, "moments" from Wechat. (collections without a special recommendation component)
2. Active model and it involves the person (subscription) These are Tencent’s Wechat and Weibo accounts. (It is true that the passive model also overlaps.)
3. Active model and it is prepared by the machine (search) - Baidu reigned here with his search. The user had to search for the information himself.
4. Passive model involving a machine (recommendations around content) - It was empty and the ByteDance guys decided to occupy this niche. As a result, a passive model involving a machine is the most cost-effective way for a content consumer.

To build a recommendation system, the guys took the concept of three profiles.

1. Content Profile – NLP was used to analyze the semantics of texts, as well as the date of publication to understand its freshness.
2. User Profile - here the user's behavior history was used + all the information that could be extracted from that behavior
3. Environment profile - here the authors tried to understand where exactly the author consumes the content: at work, at home, on the road between them, as well as weather conditions, communication quality, and so on. This depends on consumption patterns and recommendations.

As a result, the guys were able to spin a data flywheel, where new information about the user led to more accurate recommendations, more content consumption, and therefore an increase in the length of the use session, and therefore more data.

Continue reading about short videos in the form of Douyin and Tiktok next post.

#Management #Leadership #BusinessStory