10 years have passed since the establishment of the first video website in China in 2005. 2015 was officially recognized by the video industry as the starting year for ecological marketing. Sohu, as the leader of this industry, adopted an innovative and technological approach to bring the industry to its next level. Chief Technology Officer Yanfang Guan of Sohu Video announced the beginning of a smart video era with the OTV during a forum on the age of pioneering and smart video technology held on the 18th of December.

Self-made dramas as the 3rd faction to this highly contested market

The entire video sector is reviewed from the perspective of media formats. All major video channels have focused upon self-made dramas in 2015. This fad also took China’s video market by storm. Sohu kept itself aligned with the latest trends and released many self-made dramas in 2015, including “Wu Xin the Monster Killer” and “Love Me, if You Dare” which dominated the summer and winter holiday sessions respectively. The movie “Jian Bing Man” also took the title of “most popular online IP of the year”. Statistics show that total views for self-made dramas have exceeded that of copyrighted dramas, making self-made dramas one of the top three categories in the video market in addition to copyrighted dramas and professional-generated content (PGC).

There are 2 reasons behind this phenomenon. The first being that there were 433 million online video channel members, making up to 66.7% of the total Netizen population of 649 million. Such vast membership has provided a great audience base to ensure the continuing popularity of online self-made drama groups.

The second reason was that self-made drama groups are entirely focused upon the drama itself. Materials, props, and plot contents would be down-to-earth and better received by the general user. Self-made dramas are not only welcomed by the general public, advertisers were also keen to support such dramas. Advertisers once faced the dilemma of not knowing whether a show, movie, or drama would be popular before it airs. However, product placement opportunities would be no longer available once shooting begins. Post-production brand marketing technologies on the market offer a potential solution. Advertisers could freely select the timing for product placement, helping them exert better control over the advertising process.

In addition to self-made dramas, copyrighted dramas and PGC also provide great business opportunities as well. Tens of thousands of videos and drama episodes are released every year. Such techniques would no longer be capable of accommodating the advertisers’ requirements of information dissemination. The only true solution would be Big Data analytics with the capacity for large volume analysis and dissemination of advertisements. Yanfang Guan mentioned: “Big Data for videos has been speculated for many years. Smart video discovery service would provide us with the answers needed to know what our users really want.”

Searching for great selling points for video themes

While self-made dramas achieved great success, Sohu also acquired large volumes of valuable data by analyzing the audience. Statistics from Sohu showed that viewer composition of self-made dramas are equally distributed in terms of gender. One thing to be noted is that highly educated and high income viewers make up more than 50% of the total viewer population. Such population exhibits immense purchasing potentials. Context marketing could be referenced as a potential strategy for transforming these business potentials into actual purchasing as well as the means of disseminating advertisements during appropriate scenes and environments to generate positive user response.

Context-based marketing targets the current scenario faced by the potential consumer. While describing wonderful imageries or communicating information, targeted marketing or data infiltration would be carried out to achieve consensus and persuade the potential consumer to initiate purchasing. For example, during World Cup games, many diehard soccer fans stood vigil before TV screens deep into the night to view live broadcast of the games. These audiences would cheer and celebrate when their soccer stars scored goals. Advertisements for beer or soccer gear at such moments would correspond to the celebratory atmosphere while motivating subsequent purchases. Such advertisements would not appear out of place, and are capable of retaining the audience when they appear in the scene, making them perfect for advertisers who need to motivate spur-of-the-moment purchases. From the viewers’ perspective, such advertisements would not interrupt viewing experiences and are therefore more acceptable.

Top video websites in China recognize the opportunities presented by context marketing and have started performing relevant trials. However, ideal results have not been achieved due to certain technical barriers. Yanfang Guan mentioned his expectations for smart video recognition technologies, describing it as the reason why Sohu was interested in participating in this forum.


The video industry will be undergoing 5 years of immense and rapid development. The online video market reached 24 billion Yuan at the end of 2014. The scale of video advertisement market has exceeded 15 billion Yuan. Expected annual growth is also projected at a whopping 35%. Such business environment means that major video companies such as iQIYI, Youku, and Tudou have adopted corresponding measures and established work teams to capture valuable data from videos and to conduct data matching work. However, manual processing is limited in terms of scale and scope. Identification technologies offered by video discovery servicecould be a turning point to achieve vast amounts of data capture from the video. 85% of unstructured data in the video would not remain ignored for long. Analysis by video discovery service could categorize and label all commercially valuable elements in a single instance. Such information could then be associated with relevant advertisements afterwards. These technologies would facilitate data categorization and target matching, provide additional advertisement placement opportunities in various videos, and offer clients with customized push notifications. A new breakthrough will therefore be provided for Big Data analysis of video files, allowing the surveillance for the results of better targeted advertisements.

The gateway to video advertisements is opening gradually. Advancements in smart video recognition technologies will provide better support for creating additional opportunities in linking business information with video contents.