Product recommended personalization Why start with a news client

In the lack of information, the news portal because do not know the user's information needs, only through a large number of news to the user to push, so that users in the mass of news flow to find their own needs of the news to click. User is a point, news portal is a face, news portal can not only do a point, but to do face, you need "dazzling" news information.

Therefore, this is also Yahoo, AOL and other news sites can not understand the domestic news portal, one of the reasons why people need such a large amount of information? I have to say that this is a strange scenery line of the domestic news portal, the crash is all news content. The four major portals led by the news site in this regard has done a streamlining, but the new launch of a lot of garbage stations or in the "massive" flow of information mode.

As far as I can see, news clients are almost on the "personalized Recommendation" line, "personalized recommendation" has long been not the early one or two news client product patents, may not shout out on the slogan, but in the product reflects is incisively and vividly, from the homepage to the channel page are personalized recommend this route. "

Why is it a news client?

Without the exact same people, it's hard to have exactly the same user needs. Big to QQ, micro-letter and other mainstream applications, small to Jane book, painting notes and other niche applications, there may be users need to use voice, pictures, text, video and even more media, but in the product is a single, pictures, voice, text, video, etc. are fixed, the user's choice is limited, Personalization recommendations are hard to come by on these products.

News client is different, it is porting the PC side news portal mode, as long as the PC side of the news on the phone to allow users to easily browse. The problem is that the handset-side screen is smaller and the browsing habits are very different from the portal, which requires that the news client product must differ from the PC-side portal site.

First of all, the portal of various window ads, as well as the sliding screen flash ads, voice ads, video ads, etc., are not suitable for news client products appear, and this is a news portal important revenue. The news client must be a "hero", choose "broken wrist" to meet the user's reading needs on the mobile side. Such a small screen, if the rebound window, falsh ads, video ads, the user will certainly be unwanted.

News client products are very cherish their own "feathers", in the advertising progress is particularly cautious, the current mainstream has open screen advertising, information flow ads, article page tail ads three of models, they are in these models under the wind, at least a steady stream of advertisers poured in.

In addition to the most important profit model, the core part of the news client is content. Mass Information flow news portal is problematic, at least most of the news content for a single point of the user is invalid, users need to accept a lot of useless information, the news client can not go this "atavistic" mode.

To discard the traditional mass flow of information to the user to push, the user again in the mass of information flow to find the news message, you have to walk the personalized recommendation model. It is not so much news client product innovation out of personalized recommendation mode, rather than news client is forced to walk personalized recommendation mode, after all, there is no mass of information to the user pushed. Baidu News, Sohu News, NetEase News, today's headlines and other news client products have personalized recommendations of the shadow.

The main forms of personalized recommendation on news clients

Only the user-defined channel subscription function is not enough, many news client products channel subscription is very deep, users do not know how to set channel subscription. So, is the news client able to take the initiative to choose the channel that the user wants?

The main form of personalized recommendation on news clients is to push different streams of information to each user, that is, the message in the homepage of each user's flow is not the same, perhaps the majority of the social news, perhaps the majority of sports news, perhaps the majority of the science and technology news. News client products will be based on user preferences and browsing preferences, such as analysis, around the user set up to meet the needs of users to browse the home page, rather than the "thousand side" of the home page, is "Thousand Thousand Faces" homepage.

In addition to content, each user presents the content of the advertisement is not the same, the user rare enjoyment of a "personalized" in advertising.

Will personalization recommendations replace Web editing?

News client products using personalized recommendation technology, the role of the editor seems to be falling, as long as the machine can automatically get enough information, do not need to edit to create content, but this is based on the ability to have a large amount of information to give the machine access. Will personalization recommendations replace editors?

At the moment, this is a pseudo proposition. At least the content still needs someone to do, machine and technology is unable to make content, such as special production, such as high-quality content manufacturing, the latest content production, must have someone to do. Personalized recommendation optimization is the content, so that users see what they want to see, at this point we can not be confused.

Another big issue facing news clients is the censorship of information. More and more false information and sensitive information are mixed in the flow, through the news client to the user, the platform must have the responsibility and obligation to dispose of these false information.

Now the technology has not been able to do well in the anti-cheating function, only through the keyword filtering technology to filter out some spam, but the smart "text Porter" machine is unstoppable, such as Guo Jing, Guo-static, Guo Static, Ching, and so on, a slightly more complex skill, the machine takes a long time to identify the filter, and manual speed is much faster. The worry is that a lot of news client products do not have content to review this, and even have no content to review this sector, although they are not content manufacturers, but as communicators, there is the obligation and responsibility to the content of their dissemination responsibility, at this point, only the artificial will be more cautious.

After the news client, there will be more and more products on the personalized recommendation of this road, no one likes "thousand people side" things. Based on the understanding of each user's habits and exploration, will allow users to enjoy better service.