Viewing large data business passwords from network advertisements

At the end of February, at the 85th annual Oscar award ceremony in Hollywood, Ang Lee won the best director by virtue of "the fantasy drifting of the youth faction". Interestingly, this is the one of the only mistakes a--david Rothschild has predicted for all the other awards for the Oscars, David Rothschild, a Microsoft research expert, predicted by the big data analysis. Earlier, in the 2012 U.S. presidential election, he correctly predicted the election results in 50 regions of the 51 constituencies, more accurate than 98%.

This is only a microcosm, since 2012, large data has been in more and more fields and occasions have been well-known, the voice of big data, whether it is a science and technology blog, or a variety of industry, or industry forum, media topics, there will be large data related discussion. In fact, in many areas such as science, sports, advertising, public health, economic development and economic forecasts, the ability to anticipate "big data" is being developed, and data is starting to drive change in decision-making. The financial industry uses big data to identify individual credit risks, courier field through data to determine the driving route, reduce waiting time, the government through large data to find the most prone to fire and manhole cover explosion sites, shopping malls through large data to find the link between products, as well as Microsoft, such as technology companies to analyze the patient's occupancy rate, Zynga modifies game design through data analysis.

Accurate marketing on top of data

When there are a variety of data and use methods, we can not ignore the big data era of large data companies. In traditional industries, valuable data on the one hand based on the rise of the entire internet of things, from its own information sector data accumulation, which most contribute to the development of these industries themselves, but with real big data company potential is not much, the current large data companies are mainly concentrated in the IT internet industry.

The advantage of Internet enterprises is that they lead in the basic technology of data analysis and processing, and their services begin to involve the organization, management and operation of all walks of life, the data accumulates more widely, on the other hand, the behavior, demand and living of the next hundreds of millions of netizens through the Internet. The resulting large amount of unstructured data has greater commercial value.

Online advertising has become an important force for the development of large data in the Internet industry. With the rapid development of Internet and mobile Internet, more and more people will choose digital media as the channel of obtaining information. Internet users cover the mainstream consumer, their contact with the digital media and rely more than the traditional media such as television and plane, and consumer decision-making will increasingly depend on the Internet--network advertising has become one of the most cutting-edge industries to verify the value of large data.

The internet has entered a personalized, social, the era of openness and mobility, the rapid development of social media, E-commerce and intelligent terminals, daily production of a large number of semi-structured and unstructured data, such as Taobao Daily active data volume of more than 50TB, has 400 million product messages and more than 200 million registered users online activities- As Internet advertising services extend to almost all of the mature Internet platforms, they have a very substantial amount of access to these data. Using these data to support accurate advertising is the slogan that Internet advertising companies focus on in the big Data age.

Despite the fact that social platforms such as Facebook and Twitter, which are considered by the industry as the most significant data commercial value, aside from infrastructure and technology providers, Google has begun to explore more data in the field of artificial intelligence in search engines, and the rise of real-time auction advertising is based on audience data.

Chuan Yang Technology is the early domestic data as the future advertising market core value of the network advertising companies. Since its inception, Chuan Yang on the establishment of user behavior database as the focus, accumulated so far, the breadth and depth of data mining, the transmission of the database has established an industry advantage, not only has more than 900 million of the independent cookie database, Covering more than 473 million active netizens, at the same time also has a relatively complete classification model, to a user for multi-dimensional analysis of all-round.

Wang, vice president of technology and technology, said to reporters that in the advertising ecosystem, the core part is to be called by insiders to open the product "samedata" user data intelligent engine.

Samedata has intelligent network clustering, algorithm mining, integrated cleaning and other functions, it opened the 900 million cookie database and more than 1000 media resources, with the transmission of technology-leading data analysis technology and mining core, through the analysis and excavation of the kernel to continuously track, analyze, collect information, To authenticate the demographic characteristics of the audience, through the robot learning analysis model, finally, the Chinese mass of audience information data integration cleaning as can be widely used integrated database, for keyword orientation, semantic theme orientation to provide the underlying data interface and application matching display, so as to guide the entire advertising platform services business intelligence.

In Chuan Yang Science and technology advertising marketing ecological chain, samedata in which the Data management platform DMP Link, for the entire auction mode to provide real-time purchase of the audience data. In the RTB field, compared to the demand side platform DSP, supplier platform Ssp,dmp also played an important role. Because there is no DMP can not judge the target audience and individual audience prices, but also can not talk about real-time bidding. Samedata play the advantage of the huge database. Through the ABP Audience purchase platform (DSP platform) advertising, it can be in the advertisement before and after, in a very intuitive, visual form, through a variety of dimensions to advertisers and agents to show the entire process of delivery.

Specifically, Samedata in the advertisement marketing ecological chain work principle and process is this: Exchange platform or media impression generation, platform or media through the standard protocol to send requests to the DSP platform, usually carry information with media name, media classification, page URL, The contextual keywords of the page, the lowest price set by the media, and the most important information one by one user cookies. By looking up the user mapping library that has been established beforehand, the DSP converts the user to the user who can be identified by the DSP platform, and then calculates all the information received from the Exchange platform or the media, filters out the suitable advertisement and the highest bid of the advertisement, then the advertisement ID, material address, jump address, Types are passed back to the platform and media. When the bidding is successful, the platform will send the successful message to the DSP platform. The whole environmental data involved in the bidding is required to be provided by the DMP.

Therefore, the samedata-driven DMP will no longer only focus on real-time bidding ads click Optimization, and more to provide enterprise audience segmentation as the basis, and further to display ads precision orientation, site optimization, search positioning and multiple repeat positioning, Audience research and data mining dmp the audience (consumers), and for these audiences subdivided advertising opportunities for procurement or sale. The data management platform will provide more reasonable and scientific algorithm and analysis basis to help customers and media executing agencies to determine what audience is high quality, what kind of advertising platform has more such high-quality advertising audience groups, so that the results of marketing promotion more credible and better.

To be aware, the most direct meaning of large data for precision advertising is the ability to orient netizens ' interests through data analysis, and then price individual target audiences. Based on the user's behavior interest exploration, Chuan Yang application of the latest algorithm research results, the Netizen's interest is divided into 33 categories of interest, 168 interest in class, 857 interest in small categories, all-round users interest tracking technology, close to the target audience.

2012, Chuan Yang technology based on the Netizen data intelligent engine Samedata, has issued a car netizen interest map, the mother and child industry network audience user behavior data and many other industry research reports, for different age, sex, social stratum of users of different interests in commodities, browsing habits, focus, There are detailed statistics and descriptions in each aspect of correlation.

The evolution of the advertising industry

The arrival of the large data age, for accurate advertising to provide opportunities for development, but only the precision as it brought to the network advertising, is far from enough.

Alibaba research Big Data engineer has such a discovery: Only through the customer's purchase experience to screen its similar products to recommend to customers, this is in vain, similar products rather than make customers purchase desire, or even have negative effects, weaken the desire to buy.

You can't say that this kind of advertising is not accurate, because theoretically it is true, the problem is that RTB and all the traditional ads, is still the producer of the Big Horn, is a producer-centric, push technology to push the consumer ads. However, the past is playing a loose bomb, a dozen pieces, now more than a shot paramilitary. The consumer accepts the advertisement the way, still with "lie to get the gun" almost, is not to come to the door, but actually cannot hide.

Qiping, Secretary general of the China Academy of Social Sciences Information Research Center, said that the previous world Wide Web was gradually upgraded to a semantic web, which would radically change the living environment of online advertising. The semantic web, which is connected with the Internet, mobile, positioning and so on, conforms to the standard of large data, is the aggregation of heterogeneous data, and secondly, it represents the front line of the HTML5 from the web's Peer-to-peer to the mobile web, and it is the self organized and coordinated social network calculation. The semantic web uses the way of adding tags to the content, which is equivalent to breaking the boundaries between the advertising screen and the advertising screen, and the semantic web will make the user each network activity become an intelligent advertising behavior.

Samedata Intelligent Data Engine, is also based on the coverage of 1000 of the media for the long-term content of the Web page of the semantic model analysis and research, for the audience browsing the content of the Web site, the content of the Web page through the continuous application of mature training, real-time update, composed of intelligent web search engine, Provide the underlying data interface and application matching display for keyword orientation and semantic theme orientation.

Simply put, in the semantic web, online advertising will be from the current accurate delivery to the future meaningful push. Only the meaning of consumers to get through, accurate only targeted, to avoid the negative effects of the above. The next generation of advertising precision is no longer just like RTB, lbs positioning to people, but along the meaning of the meridian, direct positioning to the heart, for the hearts of the dorsally. Future advertising, but also real time, but more emphasis on this point, here and the accurate collision, and quickly gathered quickly scattered, speed.

This is also consistent with the view in the recent article in Wired magazine that "the Internet will shift from space mode to TIME mode": All information on the Internet will soon become a time based structure, moving from static to dynamic. We no longer need to rely on Amazon to inform us whether a new book of interest has arrived, and that there will be a huge change in e-commerce that we don't have to look for in any other product.

Of course, from precision to recommendation, this means that a product's advertising is different for everyone's content, meaning a transfer from a central computing model to a local calculation. So, the future of the specific form of advertising companies is hard to say. However, Nomura has proposed a future-related trend is "output reversal", meaning that in the future dispersion model, consumers become the starting point, the producer has become the end point. For large data, for advertising, production reversal will lead to the emergence of new advertising head-foot inverted.

The exploration of the technology in the field of Internet advertising is prescient, its relying on the cloud database to obtain a large number of interactive structure and unstructured data, and from the bottom of the data analysis platform to support the upstream of the application services, through the PC and mobile Internet data channels, and gradually create a vertical industrial chain form. Chuan Yang to build the Samedata, the data content composed of high intelligent search engine and Czech filter matching content, for the accurate implementation of directional data interface and application matching, become a large data import, collection and integration process of the company.

In a sense, the future of the mature network advertising companies must also be large data companies. Transmission can be at the bottom of the industry chain deep layout data collection and analysis platform, with this cross terminal, Cross-platform, cross-industry to provide personalized, accurate, intelligent mechanism, and on this basis to extend the data processing industry chain, the data processing industry chain will serve the media, agents, advertisers, users, third-party consulting and research institutions, Electric business enterprises and even the traditional enterprises within the production chain, so that consumption and production to achieve a highly efficient match and matching, explore new business model.

In the view of American it Queen Dyson, a viable business model is that, with sufficient data, future advertisers can build a data collection and analysis platform that interfaces with customers ' local data and dock with a multitude of package platforms with many lifestyle designers For consumers to provide one-on-one life-consuming advice or other personalized services. In this model, advertisers will not earn the media money, but the business money.

(Responsible editor: The good of the Legacy)