Implement large data Mining adtime restore your network

With the rapid popularization of smart phones, the Chinese people's Internet behavior in the profound changes, online shopping has become the mainstream, video sites have sprung up, mobile terminals a large number of functions to begin to replace the PC. Eric Consulting 2012 China internet annual Network advertising data show that the network advertising market scale of 75.31 billion yuan, the annual increase of 46.8%, which portal decline, search smooth, electric business, video rally. This change benefits from the mobile internet era, the change in the advertising market is the response of the network behavior, in February 2013, Science and technology news website TechCrunch Guest writer Richard Zwicky wrote that Google's new version of AdWords Blur Mobile and desktop advertising boundaries of practice, Suggests that Google has acknowledged the end of the desktop era. The boundary of mobile Internet and PC Internet is more and more blurred, these changes directly affect the trend of internet marketing technology and the change of marketing mode. There are two major trends in internet advertising: precision marketing based on large data and social media marketing based on content, and two trends that complement one another.

Precision marketing based on the big data age--How to find "you" in the network

1. The current situation of precision marketing

The "narrow" concept of the emergence of the network to redefine the precise delivery of advertising, but precision is relative. First of all, through the network data to classify netizens, used to guide the launch of Internet advertising, but the behavior of netizens is changing, identity can be changed, once bought the car again buy may greatly reduce and so on, with the collection of historical data to portray and distinguish the type and attributes of netizens is difficult to do real-time, do not real-time, it is difficult to achieve precision. Second, the growing internet has also led to a growing network of data, the face of huge and messy network data, how to collect, analyze and manage the test of every internet company. Big data for Internet technology companies is an opportunity and a challenge, the future of precision marketing in the hands of these large data, can analyze large data, share large data of technology companies, with large data, internet advertising companies will continue to direct the new concept of precision marketing.

2, Adtime Precision Marketing

There's a classic word on the Internet--you never know what each other is, it still works in the age of big data, because everyone's network traces, by semantic analysis, it is difficult to portray this person's network attributes, not only the technical problem of semantic analysis, but also the network that I am another "me." But the big data age has come, the network behind the "I" is not important, as long as his network data is accurate. Adtime Precision Marketing to local conditions, the network data into user behavior data, network personal data, for different types of data taken by the precise technology is completely different.

(1) Network behavior data generally include: browsing behavior, on-demand behavior, click Behavior, search behavior and so on. This kind of information as a third-party advertising technology companies are generally rely on cookies, but the mainstream browser Microsoft (IE10), Apple (Safari) is turned on by default does not record cookies. If cookies cannot be obtained, large-line internet companies use independent IDs and personal identification information (network personal data) for accurate population positioning, and third-party advertising companies will face a more embarrassing situation, can not carry out accurate positioning ads.

Adtime with years of technology accumulation, for the network behavior data to create a CCM precision Ads intelligent Platform, its core engine is real-time decision optimization engine Rde (real time Decision Engine), aggregation bypass collection data, DNS resolution data, Cookie data and other multidimensional data, the real-time bidding, optimization effect, etc. into the engine's lowest price strategy, the optimal effect strategy, and so on, constructs the dynamic Network behavior data classification based on the time series, only the CCM platform can divide the netizen behavior thousands of attribute, and also uses the clustering algorithm, As advertisers and Internet change generate new classifications, the granularity of their classification exceeds the analytic granularity of Internet behavior based on cookies.

RDE engine relies on the technical advantage, in the large data application depth excavation, not only for advertisers and 4 a company to provide post evaluation of the effect, locking effective audiences, but also to help advertisers and 4 a company to carry out early prediction effect evaluation. Adtime cooperates with the third party organization, monitors the advertisement data and the effect, provides the advertisement evaluation for the customer, including the launch effect report, the exposure audience, the audience, the Access source report, the loss analysis as well as the time, the geographical report and so on. Adtime regression prediction model based on historical data is a means to make up fine marketing, and a phased evaluation model of optimized post effect. Regression Prediction model is also established in the historical data of intensive cultivation, effectively shorten the evaluation node time, gradually approaching real-time accurate advertising effective way to predict the effect of advertising, to find brand audiences, truly implement the Internet "fine" marketing concept, for advertisers and advertising agents to maximize the effectiveness of advertising.

With the expansion of the network scale, it is increasingly impossible to rely on a company to achieve accurate delivery. Big Data is like a book, who has mastered reading ability, can read more thin. But "reading" needs technology, industry consensus and standardization, the whole industry needs to establish a data sharing mechanism, so as to further improve the large data age network accurate positioning.

(2) network personal data including: location information, registration information, equipment information, online shopping information. Most of this information is in the database of applications and is structured and easy to analyze. But this kind of data is highly sensitive, the internet's "accurate" marketing in the large data analysis and privacy security to look for opportunities, adtime for something not, strict adherence to privacy protection, aggregation of resources, cross-screen across the network push.

Adtime's szk--finger is mobile terminal advertising platform, has been running in domestic telecommunications, Unicom's near-Chiban customization machine core, and rely on the SDK to promote the model, the finger of the installed and active volume is still increasing. However, SZK as the underlying application does not record user equipment, location and other sensitive information. With the continuous integration of mobile and desktop advertising, the finger and CCM in the technology on the integration, in the data layer through, the user device properties as a branch of the Netizen attribute classification tree. In the advertisement statistic monitoring, the finger Guest and CCM two one, truly realizes the big Data age accurate marketing. The CCM platform Fusion Szk manages nearly 20 billion internet and mobile internet traffic every month, covering 500 million online internet users.

(Responsible editor: Schpeppen)