Since the 1990s, Internet products have realized one seemingly impossible dream after another. These dreams form the foundation of the Chinese Dream. China, which has used the Internet to overtake cars on curves, is writing a perfect answer with each passing day.

On June 30, 2018, PMCon2018 China Internet product manager thousand people summit was held in Shanghai. Conference invited from hungry?, drops, netease, hkust xunfei, analysys experts, iQIYI, and many other Internet companies, to share the social, the latest technology in the field of electricity, search, etc, comprehensive introduced the artificial intelligence, big data, such as speech recognition, block chain hardware technology is how to influence people’s intelligence, transportation, delivery, such as information acquisition way of life.

During the meeting, netease cloud yi dong product specialist wang published “from netease cloud music and netease cloud reading content products to create” speech, and the vast majority of developers depth interpretation methodology, the content of a new generation of products operation and netease cloud music products, for example, with the practical experience to share for the standardization of the content of the product operation.

Next generation product content operation methodology

With the development of technology, there are more and more channels and forms for users to acquire content, and a variety of texts, videos, songs, audio, games, articles, pictures and other massive information are numerous.

And the more kinds of content, content as the core output value of the product continues to evolve. With the upgrade of multimedia, the carriers of interactive content are also constantly upgraded. From traditional paper media and radio, to the popular network video and live video, as well as the foreseeable AR and VR platforms in the future, the development of content products has found more possibilities.

Through content, users can more intuitively understand and receive products and brands until they are recognized and further consumed. For the three life stages of content products, including initial stage, development stage and rising stage, the product content operation integrating content collection, review, editing, presentation and recommendation also has corresponding scene-oriented, personalized and standardized requirements.

Contextualization: User & Content & Association

Scenarioization refers to listing the scenarios in which users use the product one by one, listing the requirements in the corresponding scenarios, and finally planning the content according to the requirements. Through this process, the enterprise connects the user with the content to better meet the needs of the user.

Scenario strategy is often the necessary strategy to get the first batch of seed users or new users, but also easier to make users reach the aha moment, and then new users into retained customers is very important. In scenarioized operations, content is made more specific and granular so that users feel that it is “valuable”. Therefore, scenarioization must be a solution to solve problems for customers by penetrating into customer usage scenarios and providing convenient and simple paths to solve current problems.

A good example of scenario-based operations is an online education company in Pakistan. In the initial search for seed customers, the company found a middle-aged housewife. For her goal of urgently needing to recharge and adapt to the workplace as soon as possible after returning to work, she systematically listed a rich curriculum catalog, which finally prompted her to choose macroeconomics. In addition, KEEP software is also a typical use case. It attracts and retains users with prominent labels such as weight loss, fitness, weight loss and strength.

Individuation: fine operation

With the popularity of mobile development, it is an inevitable trend for users to acquire content through personalized recommendation engines. Taking toutiao, Shoubai, Weibo and other apps commonly used in daily life as examples, their personalized information flows account for a high proportion of non-social information flows and are developing rapidly.

For a product, the greater the number of users that meet the needs of the product, the greater the benefits of the product itself, in terms of DAU/UV, UGC volume, brand recognition, and so on. To achieve this effect, then give the user classification according to the different latitude, such as the type of crowd, interest, activity, etc., and then pointed to different populations provide personalized push content – this is the so-called personalized operation, tracking popular content, building user portrait, again through the data analysis, modeling precision push.

Personalized operation strategy mainly involves four aspects: user portrait dimension, intelligent recommendation algorithm, big data modeling and data analysis, as shown in the figure below:

Standardization: Black, white and gray operation strategy

When the product develops to a certain scale, the importance of reviewing the product content will be highlighted. When the mass content is spread, the product owner or the operator has the responsibility and obligation to be responsible for the content itself and present the high-quality content in a more elegant way.

The core of content operation standardization includes white box, gray box and black box strategies:

White box strategy: let high-quality content emerge spontaneously, causing rational interaction among users; Grey box strategy: let poor quality content (not conducive to product development, but does not violate laws and regulations) into the rules; Black box strategy: ARTIFICIAL intelligence machine intercepts inferior content that violates national laws and regulations.Copy the code

On the basis of controlling the basic tonality of products, content operators should pay full attention to the operation of content safety, so that high-quality content is presented gracefully, and the illegal content is reviewed and deleted in a timely manner.

Netease is the actual operation of product content — take netease Cloud music as an example

Five years ago, the industry generally believed that music player has been settled in the Red Sea market, many people have questioned netease’s motivation to enter this market? In fact, at that time, there was no music product that could really meet the needs of users. However, by now, netease Cloud Music has become the largest UGC platform, the largest music social platform and the best music media in China — all thanks in large part to its accurate personalized recommendation system.

Cloud music is a music platform, but it is also a social and media platform, but in essence, it is a data platform. It is the data behind the content platform that provides a solid data foundation for personalized recommendation system.

To provide users with precisely customized content

Millions of songs and 400 million song lists. By analyzing 300 million users’ listening behaviors, collecting behaviors and social behaviors, netease Cloud Music can obtain nearly 100 billion behavioral data every day and make precise personalized recommendations based on it.

Recommendation algorithm matrix

At present, netease Cloud Music has formed a mature and systematic recommendation algorithm matrix, including song modeling algorithm, sorting algorithm, basic recommendation algorithm, user modeling algorithm, feedback algorithm and so on.

Make use of UGC (user-generated content) to realize the integration and transformation of high-quality content

Original content is one of the important behaviors of users.

Therefore, netease Cloud Music can push songs through algorithms, screen users’ active comments and carry out a set of procedures for operation and planning activities based on users’ high-quality original content, so that each planning can directly hit users’ hearts.

Sublimation intelligent recommendation, take care of users’ habits “in all aspects”

These upgraded intelligent recommendation schemes can take into account users’ usage habits more comprehensively and meet users’ usage needs, such as the use of points, notification of new songs on shelves and one-stop service of listening and reading songs.

The combination of artificial intelligence effectively reduces the cost of content audit

For products in each register, login, interactive use in the scene, need to be registered, log in protection, identity authentication, filtering and advertising, smart as yellow, violence could identify, politics detection, such as security, and the use of artificial intelligence technology, greatly reduces the netease cloud music in the business and the content in all aspects of the safety cost. It is worth mentioning that the security system supporting netease cloud music operation has been exported through the cloud service of netease Cloud ESHIELD.

Finally, share small dry goods – Yun Village police station’s content audit rules

Reviews have always been one of the most distinctive features of Cloud music. In order to provide users with a high-quality comment area atmosphere and better user experience, The Village Committee of Cloud Village has specially formulated rules for netease Cloud Music Comment area. Specifically,

Prohibited comments include:

Contents that violate national laws, regulations and relevant regulations and contain sensitive statements such as extreme politics, crime and religion; View the comment section as an advertising area to publish content containing advertising, promotion of other platforms and other relevant information; Attack, abuse singers, publish content with insulting language, and other comments that affect the atmosphere of the community; Comments containing vulgar, pornographic and other illegal content; Attack users, abuse each other, personal attack curse, etc.Copy the code

For such comments, yuncun police station will delete them, warn them, ban them and even ban them.

Unwelcome comments include:

Unoriginal jokes completely unrelated to the song; Brush B under A; Lyrics Song error type comments (expired comments); Evaluation of netease Cloud music category (partial jokes and dirty words); Group or machine brush praise on hot comments, rub heat; Other comments not related to songs or artists.Copy the code

For this kind of comments, Yuncun police station will carry out repeated comments and cleaning up the contents of irrigation and brushing screen.

Welcome comments include:

Song background introduction; How you really feel when you listen to music; Personal experiences born of music; Original funny jokes related to the song.Copy the code

Cloud music quality reviews are numerous, as long as it is related to music “rational, healthy, friendly, sincere” review content, are very welcome.

Free trial netease cloud music with the same security service.