How does big Data play? This is a different idea of bat.

Last May I wrote about Baidu, Ali and Tencent, the three internet giants started digging big data. Over the past year, companies with huge amounts of data have tried to exploit data in a number of areas, with large data awareness and ability to progress rapidly, and systems and tools maturing.

Large data application practice, fruitful

Baidu in large data is impressive, there are Baidu migration such public welfare projects, applied in the people's livelihood and news and other fields. The latest trend is that Baidu Network alliance using large data based on the CTR (advertising content matching) data, AdSense average income increased 70%.

Ali said that already has 100PB data and at a happy pace of growth, Ma Yun's latest internal mail will be Ali strategy described as Cloud + large data, Ali to enter the data age.

Tencent wide-point platform no lack of bright spots, such as Beauty said the use of wide points in the mobile end of the harvest, millet mobile phone and QQ space cooperation is based on social Data marketing classic.

Baidu, Ali have to upgrade the big data for the company strategy, Li, Lei and other Internet representative figures in the two sessions have had a speech, to promote the government's large data awareness and open, large data is moving from the theory to practice, from the professional field to the stage of universal application.

The inevitability of Internet-led large data

Why is it that only a few internet giants have achieved success in domestic big data applications? Is it because they have the most users, traffic, and data? The answer is no. Because all the argument about big data is that big data is not big. The quality, nature and who owns it will determine the value and difficulty of large data being mined.

Things networking sensors, video surveillance equipment are always collecting huge amounts of data, but the value is not micro-broad, because the data can not be changed. Operators have the data from the user's communications, from voice to text messages to location, than any internet giant. Can only be wasted because operators are not allowed or able to use the data. Similarly, government departments, software companies have large data, but can only let it sleep.

The reason for bat to go in the forefront of large data applications, that is, the data with the nature of the management, with the Internet enterprise technology genes, open innovation and positive progress.

The difficulty of large data utilization lies in technology. From data collection to storage to cleaning, then to desensitization, classification, labeling, structural, and the final modeling analysis, mining and utilization, are technical jobs. It needs the server cluster, data utilization model and processing algorithm to protect, then is the result of the mining of the packaging, the cash.

The technology of Internet companies is a few blocks away, compared with other gold owners with large data. Carrier technology is outsourcing, banks are mostly technology outsourcing, and other public sectors such as government, transportation, education, energy and other industries, technology is a distant term for them.

and motives. Internet businesses are almost free of service products and must make money through other models. In the past, the three modes of advertising, gaming and value-added have come to the bottleneck of mobile ad mode, requiring a new model or strengthening the original mode. At both points, big data will play a big role.

BAT Big Data Ideas different

The data for bat three are unique.

Baidu is based on the user's search behavior demand data, Ali holds the transaction and credit data, Tencent is in charge of social relations data. Vary。 They apply a different direction to large data. Baidu and Ali are more radical. Tencent has been watching more and more, and may have said less.

The first is the motive.

Baidu income more than 95% from advertising, Taobao's main revenue model is also advertising. Baidu, Taobao and CCTV are the first three major advertising positions in China. Tencent's main revenue comes from gaming and social value-added services, which account for only about 30% of advertising revenue.

This phase of the major data to become the main way is precision advertising, which fits the demands of Baidu and Ali, two large data to upgrade to the company strategy.

Next is technology.

Search engines are technology-driven, and Baidu and its founder Robin Li have the most technical genes. Ma Yun announced that because he did not understand technology, so Ali technology is the strongest. Only Tencent does not emphasize technology, always emphasizes product capability.

The big data is the technical work, Baidu and Ali these two people think the technology is very strong to explore the first in line with common sense. Baidu and Ali have already made more layouts in large data technologies, from talent to architecture to infrastructure to technical theory.

Baidu has a deep learning institute, high prices to recruit large data fields and cooperation with universities, is building Asia's largest cloud computing room, Ali has a flying plan, has advanced across the engine Room 5k cluster, Apsara distributed computing system, as well as the data committee such a framework.

Several different attitudes on cloud computing platforms can support my point of view. The cloud platform and large data are Siamese twins. "Mobile", partner and user personal data, all require "cloud" to collect, store and process. To master large data, you must have an open cloud that hosts the data.

Aliyun was founded in 09, Baidu Cloud 12, respectively, corresponding to the IaaS model and PAAs (Amazon VS Google). Their cloud services collect Third-party Web sites, applications, hardware, and user data while providing infrastructure and cloud services to developers and users. Baidu migration can take effect is benefited from Third-party app for Baidu contribution location data.

Tencent Cloud was launched last September and started late. Although Tencent's open platform is mature, the open platform is more to share Tencent's users and resources out of the goal is not to collect data. and open platform is operation, cooperation, ecological level, cloud platform is the technical problem.

Finally, the position.

The typical Internet trading scenario is that users are caught up in the process of chatting, socializing and entertaining, focusing on "interest" or discovering new information because of interest. Then the user goes to understand, to find what they want (demand, desire, find), and finally completes the transaction on the net (electric quotient and O2O).

Here's a reference to the funnel model--a long classic marketing concept. In this model, the user's general decision path to consume is from attention (attention), interest (acquires), desire (desire) to action.

The above picture is a vulnerability that is reversed. The more the funnel to the bottom the higher the conversion rate. Ali between desire and action, Baidu between interest and desire, Tencent is at the top of the loophole.

Advertising fee model can be seen in three of the location of differences. Baidu is CPC, according to the number of clicks paid (regardless of the click of the behavior), Ali and other ads can directly correspond to the purchase behavior, the CPA (according to the actual effect) and CPS (according to the effect of commission) mostly. Tencent portal, QQ chat banner ads More pro-Lai in CPM or CPT (according to the number of shows or the length of time), a broad point is a CPA, but the highlights of the case focused on the application of the download field, rather than the transaction area.

All three do not want to be in a certain link, but the expectation of the next all. Baidu has the interest of such social products, there is video of this attention-oriented business, but also launched directly in the results page under the single "micro-purchase", up and down; Tencent reorganized the search business (combined with Sogou) and the electric business (merger with Beijing-east), and the opportunity to go down; Ali Investment Micro-blog, layout of smart TV and do shopping guide site to do micro Amoy , trying to go up.

Several internet giants of the motives, technology and location of the different ideas in the application of large data are also different: Tencent, Ali Layout first, Baidu technology first. The same is, several are trying to capture more data, data collection is the first stage, the formation of data collection capacity and mechanism is the second phase, the third phase is data mining, the current bat three are in between the second stage to the third phase, once the large data application into the third stage, accumulated deeper, Putting more Baidu in or will be expected to show an advantage.

The birth of the App Store and the iphone has brought people into smartphones and mobile internet eras, subverting the traditional software and mobile industries. In the future, the discovery of business value for large data will expand the space for Internet companies to grow, and may even lead to new business models and hardware products, like the App Store and the iphone, which can bring about disruptive changes in people's work and lifestyles.

Author Weibo @ Internet Chiu, micro-letter Supersofter

(Responsible editor: The good of the Legacy)