A few days ago, travel platform honeycomb was exposed to data fraud, more than 85% of the 21 million comments from competing websites copied. No matter how rightly the honeycomb expresses that the amount described in this article does not correspond to the facts, there is no doubt that there is data fraud on the honeycomb platform.
Industry insiders said that the problem of data fraud, including billing, brushing volume, brushing points, moving original content, etc., has become a major issue in the industry, but also many comment on the type of website hidden rules. Beijing Qingdao reporter learned that these data fraud is cheap, 2 yuan can buy 10,000 video clicks, or 3 cents to buy a micro-blog commentary. At the same time, the cost of prosecution is very high, before the video site sued the brush company, 950 million fraud in exchange for 500,000 yuan in compensation.
Hornets nest admit data fraud
Earlier, data from Horse Honeycomb showed that more than 85% of its massive reviews came from competing websites, and 18 million of its 21 million real reviews came from crawling other websites. The Horse Honeycomb has 7,454 plagiarism accounts, and has copied 5.72 million restaurant reviews and 12.12 million hotel reviews from Ctrip, Yilong, Metro, Agoda and Yelp.
Honeycomb responded by acknowledging the problem of data fraud first, and then by saying that there was not as much as was mentioned in the article. The content of the review accounts for only 2.91% of the honeycombs data volume, and the number of accounts suspected of false comments is smaller. The honeycomb has cleaned up these accounts.
But this argument is also pointed out by the other side, steal the concept. Is this data quantity the number or size of data? If the 21 million real reviews account for 2.91% of the total, then there are 700 million retrospective Honeycomb travel notes and strategies? Obviously not. If these 21 million real comments account for 2.91% of the data size, what is the significance of this comparison? A comment of 100 words, the size is only a few hundred bytes, but a picture in travel notes is just a few MB. A journal may have tens of thousands of times as much data as a review. Comments are not important because they only account for 2.91 percent of the data. The other side also listed the publicity of the honeycombs 21 million real comments on the front page to illustrate the importance of the comments to the honeycomb.
A few days later, at the honeycombs new product launch, its CEO also admitted on the spot that the data fraud problem exists, but the content of the article does not match the facts. Some people in the industry said that the problem of data fraud has become a hidden rule in the industry. Some people almost have the problem of data fraud. At present, no matter the industry or the individual, very little attention and accountability to the problem, making data fraud further become a hidden rule.
According to industry insiders, there are two kinds of fraudulent data demand, one is the merchant, the other is the platform. For merchants, the amount of bills and brushes on influential platforms can enhance the ranking and influence of merchants. More favorable comments and more advanced influence will affect consumer consumption decisions, so as to win more business for themselves. Therefore, in many e-commerce websites, and even many offline marketing companies have set up agency business to guide businesses through a variety of ways, including billing, promotion, brush evaluation and other ways to bring their own benefits.
A restaurant owner told the North Qingdao that he had opened a new shop because of the relationship between popularity and geographical location, traffic has been small. Now many consumers rely on the Internet platform to find shops, so we value this very much. Find a special marketing company to help do, they first help us to improve the stores browsing and click-through rate, and then enhance the volume of transactions and praise, very set. The fake is true. Now the customers in the shop are really much more than before. I think its worth it.
For the platform, why do many merchants brush comments or bills of behavior open one eye and close one eye, and even the platform itself will stealthily brush it? Because for many content platforms, data is life, and only if the platform as a whole maintains enough, high-quality evaluation, consumers will form a habit of using, it is true that consumers open more times, more consumption, will bring more consumption and evaluation. On the other hand, it is the need for financing. Give investors a good look at the data, is every start-ups common pursuit, short-term data upgrade is difficult? Data fraud may be a shortcut.
Previously, the platform was found to carry large-scale evaluation of other platforms, the platform recognized the existence of illegal reprinting of stores, and said: The incident occurred because the platform is new on-line trial operation of the recommendations column, in the unauthorized circumstances of the relevant content of the illegal reprinting. What about the new column of the platform without evaluation? Other platforms are reprinted. This is also a problem for many evaluation platforms.
The cost of brushes is as low as 2 yuan.
According to the investigation by the reporter of the North Qingdao, the price of this kind of machine is very low. The cost of forgery is actually very low because of the mature related technologies such as the amount, quantity and comment of the machine.
For example, on the microblogging platform, a company that specializes in microblogging marketing offers a price of 5 yuan from the beginning, microblogging forwarding, praise, microblogging voting, the prices are as follows: the first experience price of 5 yuan 100, 100 = 10 yuan, 1000 = 80 yuan, 10 000 = 600 yuan; comments: 30 yuan = 100 (note: specify a microblog at least 100 times, less than 100 press) 100). If you go to other peoples microblog comments, the minimum is 30 yuan 100, microblog reading is 10 yuan 10,000, 80 yuan 100,000, and video playback is 20 yuan 10,000. The company also helps users increase their fans, which are grouped into junior, senior, boutique and top categories: Premier, junior are all scrap-together (no quantity, no quality), senior is head, Junior is nothing; top, boutique is a real person, with fans and blog posts. It has been revealed that many top and high-quality accounts are derived from the theft of real usersaccounts, after the capture, these previously carefully maintained accounts have become weapons in the hands of others.
And for the video website works brush, quote 2 yuan. According to a marketing studio responsible person, at present all video websites including Youku potatoes, Tencent, Aiqi Yi, Sohu, Levision, PPTV, etc. can operate the brush volume, the price varies. Among them, Aiqi is divided into two gears, one is drop, that is to say, the machine brush volume, this way is easy to be found and shielded by Aiqi technology, 10 yuan 10,000 times, within 7 days packaged; the other is no drop, the price is 80 yuan 10,000 times, but the file does not guarantee speed, need to queue up, data delay. Late update, no urgent order. Several other video sites offer from $210,000 to $60,000, some promising drop compensation and some guaranteeing fast. Some brush merchants show nearly 5000 copies of monthly sales, and the illegal gains obtained are also considerable. Prior to this, Iqiyi prosecuted a company with brush volume and was awarded a compensation of 500 thousand yuan.
And a water army business marketing company chief told the North Qingdao reporter, a film in the bean scoring input in 5000 yuan. He said that basically every movie or TV series has arranged bean paste score, micro-blog hype and other channels during the promotion period. Although the bean paste aspect has been very strict, but the bean paste for film and television products is still very important, so a large number of companies will do it. How to do it concretely? The boss said, according to the specific film, TV play specific analysis: the film is more difficult than television, film reviews are more concentrated, a flood of reviews to come in, basically on the same day on the score (show score). Unreleased new plays and new movies have not split up, the operating space is relatively large; and has been shown, it is necessary to do data collection, technical analysis. Specific prices should also be based on customer demand, a smaller crowd, better operation, about 5000-20000 yuan can be; and similar to the voice of the previous period was more questioned, later we this level, may be hundreds of thousands of dollars. He said that the general scoring can not simply brush, but also with long and short reviews, praise, discussion area to do together, and the price is usually relatively high, long and short reviews, discussion area, praise have unit price, depending on how much the project needs. Specifically, the length of the evaluation of 40 yuan / piece, discussion area 40 yuan / piece, 2 yuan / piece of praise; do these projects will be added four five-star points.
As for the catering platform, the price is relatively high, inquiry a company introduced, professional team operation review, point evaluation grid 3 to 6 VIP account is 60 yuan a, 1 to 2 star account is 40 yuan a. Large quantity can be preferential. In order to ensure credibility, the other side also indicated that only local accounts should be used.
Previous insiders also found that dragonfly FM through forcible backstage self-start and other means of monthly live data fraud, a user then in the well-known community know on the explosion of dragonfly FM backstage self-start code. By reverse compiling the Android version of Dragonfly FM, he discovered that the software contained mandatory boot codes called Prometheus and Zeus. The former can silently start a window-less transparent interface in the users mobile phone and forge the DAU (number of active users per day), while the latter can trigger the advertisers advertisement independently and send it back to the third-party data company, thus completing the operation of users click on the advertisement independently to cheat the advertising fee.
In addition, the CEO of Dragonfly FM announced that the number of users exceeded 200 million two months after the announcement of 150 million, that is, 50 million users in two months, many industry insiders have said that the data must be problematic, and the purpose of data fraud should be to increase advertising revenue and promote financing. Although FM responded to the AB control tests and statistics of relevant user indicators when the new features came on line to facilitate the technical framework for product decision-making, it was still impossible to explain why it increased the number of ad hits and daily life.
The platform itself will crawl users information or comments with the help of Crawler and other technical means. According to an engineer to the North Qingdao reporter, some websites can use the web crawler to capture, crawler is normal people can browse the content, with open query interface read out, and then summarized into documents. Because the content of evaluation is open, it is easy to be crawled by crawlers. Some websites not only crawl evaluation content, but also copy the content of registered users.
Baidu had experienced similar problems before. Baidu maps and Baidu knows that unlicensed Baidu maps and Baidu knows that a large number of information from the public review network, in Baidu maps and Baidu knows that a product search for a merchant, the page will display the users evaluation of the merchant information, most of which are from the public review network. The latter will appeal to the court, the court found that Baidu Map notarized by the merchant review information, which involved the catering industry 1055 merchants used a total of 86 286 comments from the public review network, an average of 81 per merchant. More than 75% of the comments used by 784 merchants came from Popular Reviews, and all the comments were displayed in full text and were mainly at the front of the page. Therefore, at last, Baidu was sentenced to compensate for the latters loss of 3 million yuan.
The law stipulates that the amount of brushes should be held responsible.
In fact, brush volume behavior will not only reduce the credit of websites or merchants, but also lead to unfair competition. Many of the platforms or companies in the above cases were accused by the court and were eventually sentenced to compensation by the court. However, why do we still have a large number of brushes and brushes? According to a marketing company, on the one hand, because of the amount of brushing merchants will also increase the activity of the platform, so many platforms open one eye and close one eye, will not report; on the other hand, it is difficult to prove that many marketing companies will use manual, technical and other means of brushing, disguised as true evaluation, not easy to detect, not to mention Easy to obtain evidence; third, the cost of safeguarding rights is high, it takes time and economic costs to go through judicial channels, and the ultimate amount of compensation is not high, such as the amount of Aiqi art brushed 950 million times, and ultimately was only compensated 500,000 yuan. However, Wang Limin, President of the court of intellectual property in Xuhui, Shanghai, believes that rights should be protected. He said that in the wave of information technology in modern society, information and data become an important force to promote social progress, large data and corresponding application technology is particularly important, has become an important factor for market operators to grasp the competitive advantage. The protection and regulation of the great commercial value contained in the big data information by the legal system should adhere to the basic idea of safeguarding transaction security, promoting technological development, respecting honesty and credibility and recognizing commercial ethics. The act of brushing volume will damage the commercial interests of video websites and the legitimate rights and interests of consumers. Tort liability. Wen / reporter Wen Jing this article source: Beiqing Net - Beijing Youth Daily editor in charge: Wang Fengzhi _NT2541
In fact, brush volume behavior will not only reduce the credit of websites or merchants, but also lead to unfair competition. Many of the platforms or companies in the above cases were accused by the court and were eventually sentenced to compensation by the court. However, why do we still have a large number of brushes and brushes? According to a marketing company, on the one hand, because of the amount of brushing merchants will also increase the activity of the platform, so many platforms open one eye and close one eye, will not report; on the other hand, it is difficult to prove that many marketing companies will use manual, technical and other means of brushing, disguised as true evaluation, not easy to detect, not to mention Easy to obtain evidence; third, the cost of safeguarding rights is high, it takes time and economic costs to go through judicial channels, and the ultimate amount of compensation is not high, such as the amount of Aiqi art brushed 950 million times, and ultimately was only compensated 500,000 yuan.
However, Wang Limin, President of the court of intellectual property in Xuhui, Shanghai, believes that rights should be protected. He said that in the wave of information technology in modern society, information and data become an important force to promote social progress, large data and corresponding application technology is particularly important, has become an important factor for market operators to grasp the competitive advantage. The protection and regulation of the great commercial value contained in the big data information by the legal system should adhere to the basic idea of safeguarding transaction security, promoting technological development, respecting honesty and credibility and recognizing commercial ethics. The act of brushing volume will damage the commercial interests of video websites and the legitimate rights and interests of consumers. Tort liability.