Glory to robots! Replacing men on the Internet.
Comparison of the quality audience sites.
The editorial board PC Magazine had an interesting esperiment to separate the wheat from the chaff, that is, robots of men. Omit thinking about who is in the grain and go straight to the point. The goal was to test a hypothesis about the possibility in principle to calculate the formal parameters by which one can characterize the quality of social networks and media on the web.
The value of the site is largely determined by its type. Types of sites for men in all the variety of choice of four: "content sites", "site navigation", "social system" and Web-service (for robots more types of sites, but it's understandable:)
Content site: a site where you are. You are not a robot? Then you do dolzhnobyt Interestingly, this type of sites can be assessed on the 3 criteria: "completeness of the information," "cover thematic field" and "editorial contribution."
To assess the need for these criteria to evaluate primarily the quality of information sources (especially appreciate the completeness of the information). It is clear that the sources we are renting, but still there is nothing unique, but only skillful rewriting. This was another Solomon said:
"- Would say," See, this new "- but it was already in the ages which were before us"
Well, since everything has already been many times, then turn to the original sources. They are mainly two: press releases and publishing colleagues. Accordingly, the identity can be traced to the source and our website - acceptor. Clearly, if you want this process to automate. So, was collected database of all online primary sources, which is the "Information portrait of the day."
"The editorial contribution" - the degree of processing of the original content source. Revision was based on the assumption that the redesigned content carries more weight than just skopipaschenny, as a result of processing with the new materials, the emergence of new comments of experts, the value of publishing rights for the reader (and a robot:) increases.
It is also defined and complete information, ie, the ratio of the number of keywords in the literature (more - the better).
Index "covering the thematic field" is defined as the number of significant "those days" caught in the online edition. It's simple: from the above-mentioned database with information picture of the day singled out and counted the number of topics, many of them are mentioned in the news site.
The social system built around their communities. These are social networks, blogs and forums photosites, unified video, many entertainment resources, content which creates a "collective mind" of participants. Of particular interest is the study of the effect of "reason."
The experiment editorial concluded that a significant number of visits to pages on the Internet does not generate real people, and specialized robots (or bots): Agents News Gathering, various "spiders", etc.
Web-robots typically come from outside. They can add entries to your blog, bookmarks to the social system, remarks to the forum. Such a robot can be dotatochno intelligent - he sometimes is able to vote, open links, etc. There are systems that can simulate the "discussions" in the comments or requests such as "greetings to you, catch five." Write a bot simulator "Pepsi generation", is now almost trivial (in systems where the registration is sought, such as passport, such a trick to throw harder). Not by chance, some even dating services in television commercials as one of the advantages presented with "the only real users." To estimate the ratio of robots and humans in the service edition registered the appropriate account on some social services. In the blogs posted announcements wording of Articles Site pcmag.ru, band played the role of a stable source of incoming records. In addition, we created several virtual users who post entries and references to a deliberately popular topics (the list of those formed on the basis of the rating services' Yandeks.Blogi "). The study was recorded statistics and reaction "of society." In assessing the results it became apparent that there are specific behavior patterns that distinguish the human from a robot. Summarizing, we can say that man is diverse and inconsistent, the robot is consistent and methodical. Distracted from the topic, you can continue the thought: purposeful, strong-willed and plan the next life people actually not so far from the robot, and ideally it is:)
15 to 30 percent of blogs for six months will be forgotten creators - are the conclusions of the experiment. But no matter: this percentage is more than the hordes of robots vomestyat, replacing the people: in Learn actively discussed the influx of bots script associated with the recent scandal with the abolition of basic accounts, when a sufficiently large number of users have closed their logs (system shows about 15%, but this taking into account the mass-created magazines bots, so in fact more)
In addition, the experiment allowed us to determine the level of user education, financial situation and audience, etc. In the first case, the evaluation was to establish a pool of keywords that define the cluster of interest the audience for which it is difficult to assume a high educational level. As the foundations were chosen title series and youth comedies, popular with mass audiences (such as "Happy Together", "Not Born Beautiful", etc. Data were extracted using the service "Yandeks.Blogi." To estimate the parameters of the material counted the number of mentions in diaries of purchases of expensive goods, tourist trips, on foreign missions, etc. In the column "The diversity of interests" are estimates reflecting the breadth of interests of users of the system (determined on the basis of analysis of the "tag cloud" or categories of the blog) is more interesting figure is that the editors conventionally called the "herd instinct." This figure, which reflects the willingness to discuss the proposed audience to her subjects, defined as the ratio of the average number of "topics" on the same topic (with the same tag or in the same category) to the average length of the discussion. The idea was to identify natural way folding communities, interested one way or another subject.
Another typical archetype of the sample site in 2007 - Web-service. In this case, analyze the content component - usually meaningless. Sites were chosen more for the relevance of services and technical implementation (in some cases on the basis of earlier estimates, in particular, this applies to file sharing, photo site, etc.).
In conclusion, we emphasize again: the figures - it is not evaluation, and generalized indicators that reflect some of the trends that we identified in the course of the experiment on a limited data set. They should be viewed as guidelines, artificial metrics that allow to reveal the specifics of specific resources.
Blogs and communities
| System | "The reality of the audience" | Education and intelligence | Prosperity | Diversity of interests |
| habrahabr.ru |   | | | |
| Privet.ru | | | | |
| "Blogi@Mail.ru" | | | | |
| "Learn" | | | | |
| "Rambler · Planet" | | | | |
Thematic Media
| Site | Information completeness | Covering the news of the field | "The editorial contribution" |
| lenta.ru | | |  |
| astera.ru | | | |
| utro.ru | | | |
| rbc.ru | | | |
| securitylab.ru | | | |
| klerk.ru | | | |
| regnum.ru | | | |
| 3dnews.ru | | | |
| membrana.ru | | | |
| sostav.ru | | | |
Meanwhile, the blog SEO & Money carries out the action " Advertising for advertising "