Yandex is not Runet, Runet is not Yandex: FAS asked for a little equality



Does Yandex take 20% of traffic from similar services due to inequality in search results? Let's figure it out.



In short, the main problem is that Yandex, even as a private company, is abusing its dominant position in the Internet search market and promoting its own projects using methods that are inaccessible to other market participants and to the detriment of the latter. At the same time, according to antitrust law, if search is a market, then this cannot be done. If the search is the property of Yandex, you can. That is, either Yandex can do whatever it wants, since it is their search, or there is a regulated area such as the advertising market, and in this market some kind of neutrality should be observed, corresponding to the antimonopoly position of Russia.



My colleague Dmitry examined this situation from the legal side in this post . I will show what exactly is wrong, and I will show examples of what we showed the FAS when, in response to a question about the presence or absence of a market, they asked for evidence of possible abuse.



The FAS considered the problem and found that one of the main problems now in the "sorcerers", specifically - that with the help of them Yandex blocks the organic issuance of privileged services. Secondly, manipulations were noticed in the ranking, again in favor of Yandex's own services (although this is more important, it is more difficult to demonstrate).





Let's talk about the Yandex services themselves and the impact of Yandex on them



Yandex makes some services from scratch (more on them below). But Yandex buys the most successful Internet projects. And the purchase does not always benefit the project itself (remember, at least, the audience's reaction to the updated Kinopoisk - it was so bad that at first Yandex even rolled back the version to the previous one, it caused such a strong negative among users).



No less indicative are the cases when the founders of a successful Yandex acquisition leave their native project shortly after the purchase. The situation itself is not remarkable. But sometimes these same founders lead (or found) a new, similar competitive project (examples can be found here). In other words, there are services for the user, and there are services for Yandex.



Why am I starting with this question - because if the search market is ultimately not designated as a regulated entity, this can lead to the fact that any Runet project has only two ways: to join the Google ecosystem or to join the Yandex ecosystem. That is, we are not talking about projects for people, but about growing startups for resale to a major player.



The Federal Antimonopoly Service did not track such transactions for the purchase of services (they are too small, not her profile), and as a result, an Internet giant with its own ecosystem grew up, which is “a Swiss, and a reaper, and a gamer. Everything would be fine, as long as this giant is not a "mirror of the Runet". If you "sit on a tube with Internet traffic" you can very cleverly manipulate this very traffic, hiding behind good intentions (such as "search quality", etc. - despite the fact that the quality is measured by Yandex itself).



The current requirements of the FAS, if Yandex fulfills them, it seems to me, will have a positive effect on the development of competition. Ultimately, we - ordinary users of the Runet - will receive different services, and not those that are done by a single company that has decided to choose for the user what is best for him.



And what about the international arena? You may have heard of Google fines for violating antitrust laws. But Google still makes more than these fines, it's like a wealth tax. Let's hope Google will meet the requirements. I would not like the development of the situation with Yandex, when the fines on the scale of the business turn out to be small, and they refuse to fulfill the instructions of the FAS. Yandex is a big company, with good lawyers - you can sue the FAS for a long time.



Service quality



Now let's move on to how high-quality Yandex services are. Let's look at such an indicator of quality in the web industry as the bounce rate, i.e. the proportion of visitors who viewed a single page during a website visit. The lower this indicator is, the better.



: : «, - , – ».
. . NPS, , , , « », . – , . - . , SimilarWeb, « » « » — – ( , ), .



, , – , , (, , , – ).





Let's compare Yandex services and similar ones, using examples. Using the SimilarWeb service, we can find out the bounce rate for any more or less large site (data for January-February 2021).







The graph shows that 35-40 percent of visitors, getting to the Yandex.Real estate service, end their sessions, limiting themselves to viewing one page. For other sites, this indicator is significantly lower ("materiality" follows from the values ​​of the same indicator for previous periods - fluctuations occur, but in a relatively narrow range).







Likewise. And one more graph:







As you can see, only Yandex.Afisha's performance is not worse than its competitors (and one cannot say which is better), and this has manifested itself in recent months.



I tried to take for comparison exactly those Yandex services that were not bought, but made from scratch, or were bought long enough for the toxic effect of "light traffic" to manifest itself. “Complex” and technological ones were also chosen. those where bounce rate as a measure of quality is applicable.



The difference we demonstrated in the numerical values ​​of the percentage of refusals turned out to be not only significant, but downright striking. We at Tutu.ru consider it a great achievement when the bounce rate decreases from 40 to 35%, or from 35% to 32% (The lower the absolute values ​​of the bounce rate, the "more expensive" each percentage - sometimes we fight for fractions of a percent - so after We managed to reduce the failure rate in one of the segments from 6% to 4.5% - 5.5% and this was a victory).



It is also known (for example, web analysts and SEO specialists) that the bounce rate is a significant ranking factor in modern Internet search engines (subject to comparisonsites from the same industry). Presumably not by chance. Below I will demonstrate another approach to measuring quality, which is based not on direct measurement of some indicator, but on indirect indicators.



Unfortunately, it is not easy to find out the bounce rate everywhere, for example, if the service is located in a subfolder. Surely there will be a Yandex service that has made all other Internet sites by this indicator, but my goal is different - to show that even such large services as Travel and Real Estate do not win at all by this indicator.



What are your specific complaints against Yandex?



Yandex is a mirror of the Runet, but not the entire Runet. Search is not the only merit of Yandex - it is the constant work of all market participants (this was the case before, and it will be so in the future). In our case, Yandex copied our models several years after their successful development. We do not mind, but only if all this is in an equal competitive environment.



It works like this: by making good sites ("sites for people" - this is what Yandex strongly recommends at specialized conferences and in manuals), webmasters attract an audience, ensure the flow of people from offline to online. Facts: Yandex was founded in 1997, but the beginning of the millennium was the start for many Russian Internet companies: Drom.ru - since 1999, Gismeteo.ru - since 2000, 2gis.ru - since 2001, Cian.ru - since 2001, Tutu. ru since 2003, Profi.ru - since 2006, Avito.ru - since 2007 and many others.



A reference to world practice and a hint that “Google has everything the same” are too short-sighted and do not pass the test. Let's compare the search results, for example, for the phrase "air tickets Moscow Osh" and compare exactly how Yandex.Travel and Google.Flights look in the search results - for convenience, in the screenshots I left only what relates to the sorcerer.



Google (screenshots from February 27, 2021):







Yandex:







Even without any comments, the difference is obvious. And yet. Unlike Yandex services:

  • Google Flights is not a separate service
  • no separate Google Flights domain
  • no seo-pages (i.e. Google Flights does not compete with other services either in context or in organic matter)
  • there is no brand of such Google-Flights (no logo, not even a branded heading)


Summary: Google Flights is an interactive response of a narrow profile, it is present exclusively in the form of an interactive response and does not appear in the SERP as a separate snippet or as a contextual advertisement (the snippet travel.yandex.ru in organic SERP - left in the screenshot for a reason - Yandex services are often "there and there").



If you still doubt that Yandex services are separate sub-brands, then please - the same Yandex.Travels have accounts on social networks:







Google Flights Interactive Response has no brand identity, i.e. is not an advertisement for a separate service. Yandex has the opposite picture - in fact, Yandex is abusing its dominant position, thus promoting its services to the detriment of other market players.



Another example is the query "repair of plastic windows", Google (missed contextual advertising, screenshots from March 1, 2021):







The same Yandex (do not forget that this is often the first content after advertising):







At first glance, there is no significant difference here. But as soon as you click on any element and the differences are already striking: if in Google you find yourself on a regular map (like searching for organizations on a map), then in Yandex a full-fledged site comes out, with filtering, adding orders, a set of performers in Yandex's own database. Services. That is, Yandex not only helps to find a contractor, but completely takes over all the functions of specialized web resources. And the audience is attracted to Yandex.Services - in an exclusive way - through sorcerers.







And a similar situation is with almost all sorcerers. None of the market participants can get into Yandex's wizards, no matter how hard they try.



Google sorcerers are available to all players and promote not only its services - and this is exactly the approach that is global practice. Moreover, their cost for a third-party player is exactly the same as for an affiliate. And Yandex does not even pay anyone for this (really, why shift from the left pocket to the right? Let's save on VAT!)



And what about the decrease in competitors in the search results?



The comparison methodology is based on the fact that if the site is of high quality, then the search engines have differences in algorithms, but they will not be able to NOT show a good site, because this will worsen the quality of the search itself. Imagine that Google stopped showing Wikipedia or something similar in the SERP. Of course, to get a full ban, you still have to try. But if you just downgrade an undesirable site (competing with its own services) in the search results a little, then it is difficult to notice this, because there is always room for arguments like “just your developers make a site that ranks better in Google, and worse in Yandex”, “the core of our search engine - trained by MatrixNet, we ourselves do not know how it works, but it works well "and" Yandex audience likes Yandex services more "and the like.



But most of this argumentation breaks down if you look not at a specific site, but take a representative sample, and conduct a comparative analysis of industry leaders (the analysis will be for 6 industries) and the corresponding Yandex services.



Description of the technique: we used SimilarWeb as an available source of data for all sites. The compared parameter is the share of search (organic) traffic from the Yandex search engine and the Google search engine. Only desktop traffic was considered (since there is no available data on the share of search engines in mobile traffic). Data were taken for 2 years by months - this allows you to see not only the current state, but the dynamics of changes in the investigated parameter over time. In particular, none of the players will be able to influence these data - they are historically “fixed”.



First, the results for each of the 6 industries considered (Auto, Real Estate, Cinema, Work, Travel, Concert Tickets, etc.).



The solid lines in the graphs are the share of traffic from Yandex, the dotted line is the share of traffic from Google (roughly equal to “100% minus the share from Yandex”).



The red lines in the graphs are data for Yandex services (different colors have been selected for other sites). That is, the red solid line is the share of traffic from Yandex specifically for the Yandex service. They drove.





Solid red (traffic from Yandex to auto.ru - belongs to Yandex) - higher than others.





Likewise, the





imbalances here are not so striking (apart from the beginning of 2019).





This is perhaps the only case when Yandex's service confidently receives a high share of traffic from Google. It seems that this is the case when Google admits that Kinopoisk is a "quality site" that is not inferior to other players. The question is whether this was due to the fact that Yandex redistributed traffic to it earlier, which is generally possible.









Here, as in the previous graph, the share of traffic from Yandex to the corresponding Yandex service is downright indecently high (sometimes more than 90%).



The most interesting thing becomes clear if you build all these graphs at once on the same axes.





A clear trend is already visible - the share of traffic in Yandex for all non-Yandex sites is not just below the corresponding services of Yandex itself (I wrote about Kinopoisk above - together with ivi.ru these are the two lowest lines on the graph), but there is also a clear trend towards a decrease in this shares for the last 2 years.



For ease of perception, I will allow myself a little liberty (let the literalists-mathematicians forgive me) - I will draw something like a trend line - I will average the percentages (here it is moderately correct, because the accuracy of the lines can already be traced with the naked eye).







findings



  1. I believe that Runet is created by the efforts of a large number of companies, tens of thousands of specialists every working day. This is a common space, and it became so thanks to openness, fair competition and equal rights for participants in the Internet industry.
  2. ( -), . .
  3. , , ( ) , , , .


FAS noticed this and took the first step. Here's another analysis of the lawyer .



At the same time, I believe that Yandex is a good company that has made an excellent Russian business and is confidently rebuffing Google (which does not happen in every national market). But I am worried about the fate of Runet, and the prospects of the company, in which I have been working since its foundation, and, ultimately, my personal income, due to the fact that Yandex slightly violated antitrust rules. And I also understand where this is leading: the effect that Yandex has on the Runet is disastrous for the Internet industry. And this effect has long been clearly visible to specialists, and now, I hope a little for you.



It is interesting that Yandex and Google (and other search engines), when the search market is recognized as a full-fledged market, will have to show their criteria for the quality of services. That is, if I rely on the bounce rate, and Yandex relies on its own metric, this “own metric” should be transparent and open.



All Articles