Using technology to recommend personalized reading in simple web

Intermediary transaction SEO diagnosis Taobao guest Cloud host technology Hall

With the development of "information gene" technology, Simple network found the application of the export: unmanned vertical theme microblogging, and personalized reading applications to read.

Article | Shen Lingli

Simple nets are currently developing products based on their "information gene technology", one is used on micro-blog nearly 100,000 theme microblogging, no manual operation, machine automatic screening and release. In addition, the development of personalized magazine subscriptions to the application of "read". Its third product will incorporate the user's personal recommendations to form a social organization platform based on reading interests.

Why do you think of this model?

At the end of 2006, Ding Jun, 32, resigned from the position of Director of the Microsoft Consulting Department and set up a simple network in 2007, initially positioned as a precision marketing intermediary between advertisers and users. At the end of 2009, he decided to do a good job of solving the information explosion in the quality of personalized screening, so spent a year to build a technical model, the company's internal technology called the "information gene" technology.

What is the core technology?

The core technology of simple NET is "information gene" technology, that is, the concept of semantic environment runs through the computer algorithm, and the quality score is used to establish the information acquisition model. An article in the information point of the most dense, the information relationship of the largest information cluster is its central idea, but only by "Apple" two words, can be sure that the article is a kind of fruit or company? The semantic environment "judgment is to make this confirmation, make the information of the filtering more accurate, and" quality Score "is based on the location of information source and past history, to the information to calculate an initial score to ensure that the selection of articles with relatively high quality.

What product did you do?

The present products of the net are all different manifestations of the "information gene" technology. First of all, the simple network of this technology used in the micro-blog, the establishment of tens of thousands of vertical theme microblogging, each account issued 10 per day with gold content within the information, formed a number of vertical micro-media.

Simple network industry Weibo each is very vertical, a single account is not a fan of millions, but its release of information on a daily basis can get 30多万次 forwarding, covering the total number of fans over 50 million. These tens of thousands of vertical information microblogging, both the "information gene" technology application, but also capture the user's interest and personal data entry.

Then, the simple network in the mobile terminal developed another personalized information software-refer to, A and zite, Flipboard and so on quite similar products. Users can subscribe to the latest quality content that is filtered from the whole network according to the topic of interest. Moreover, user habits and paths can also affect the recommendation engine, and the recommended content will be adjusted as personal characteristics change. Two people with similar reading interests are recommended for the first time when they use the reference, but the content they receive will be different after a period of time. User interest is more subdivided and personalized.

According to Ding Jun, there are more than 500,000 iOS users, more than 300,000 Android users. Users can choose to bind their personal microblog account using a finger-reading, one-click Sharing. "The background of the net is similar to the search, just do a ' semantic distribution '." Our information organization approach is based on individual points of interest, rather than RSS (as it is) according to the source (Organization information). ”

Read users is the basis of information discovery and recommendation, based on user interest, not limited to the subject of the accurate recommendation of information. Ding Jun Introduction, the current Simple network is developing a third product, the "people" of the views added, by the user individual to filter and recommend information. At the point of personal interest subdivision to explore the interactive point, to build a social platform based on reading interest.

Different from other similar products?

The information screening of simple nets is based on three dimensions: Semantic distribution, quality score, and "people" recommendation. Unlike search engines, which are based on keyword matching, "Information gene" technology is a bit like manual editing, not only to determine the gist of information, but also to judge the quality.

At present, most of the social recommendation tools are first to provide plug-ins, with different sites, from the site embedded code generation integrated Sharing button, and then by the user Point button to form information diffusion. The design of the net is that the Personal Computer toolbar generates a button similar to the Favorites folder. When browsing the whole network, users just want to recommend the article page click, you can share the user in the simple web of the personal homepage, and the title, summary, the form of the original link to show. The ads on the original page will also be filtered out.

How will the profitability be?

"Simple net current products, whether it is micro-media or refers to read, are guided by the category of products." We have no hope of making money in front of the products at present. In short, the future must be in or after the profit way, but in the back-end of the charges first to do the volume, the Internet business is the essence of traffic business. "Ding Jun said. In 2007 and 2011, the net won the angel investment of Qiming and Triumph Venture, a total of 8 million U.S. dollars, the team has more than 70 people.

Managing director of Dong Gobi Investment

Personalized reading is the highlight of the upsurge of social content outbreak. Simple network in the user's personal collection home page and other aspects of the product portfolio is plump, from different angles to meet the needs of users. Domestic in this direction has a lot of products, how to quickly accumulate users is this phase of the team's greatest test.