Igor Ashmanov: “The search interface is now at the DOS level”

Igor Ashmanov, managing partner of Ashmanov & Partners , in an interview with Habrahabru sets out his vision of a search engine crisis and reflects on how to deal with doorways, which will probably annoy people even more.

You often talk about the crisis of search engines. Tell us what it is manifested in? What did he become the consequence of?

I, if possible, briefly, point by point. Because I have already stated in detail at the Internet and Business Conference (the presentation is here ).
')
It manifests itself in two directions:

a) a certain stagnation in technological development. That is, in that:

* The user interface is stuck at the level of the late 90s.
Endless tape of technically-looking search results, lack of topical classification, other. As I said before, the search interface is now at the DOS level, and with management it is also a command line.

* there was a shift from the search towards "portality", that is, winding up all services around the search. Everybody moved there, including Google , which until recently tried to keep the focus on the search. Moreover, these "side" projects are often unsuccessful, no. If you look, for example, at the statistics of content projects of Yandex and Rambler , most of these services are not leaders in their respective industries, to say the least (that is, if there is free, large traffic from the portal’s head, it’s just amazing). Is that the news projects portals no doubt successful.

And the real reason is, firstly, the lack of fruitful opening ideas, as the famous character said, and secondly, you have to deal with your grandmother! Everything is already working, the contextual advertising machine brings a tumultuous stream of money, companies are rapidly swelling, hiring several people a week. Here the sky with an ovchinka will seem, not to fresh interfaces and new technologies, in my opinion. And as you know, with the rapid growth of any corporate culture, it is practically impossible to preserve, and there may be problems with the strategy. Those computer companies that are now on the heels of - Dell , Microsoft , Oracle - these are exactly those where the level of management grew faster than the market. And the rest just broke, they died - and it is in a booming market.

However, I think this fall and winter of 2007 the shock of growth will begin to pass, search engines will roll out some new items and something will move.

b) The fall in the quality of the search itself. Here the reasons are obvious, there are two of them:

* search engines do not interact with the community of webmasters. This means that along with contextual advertising, which guarantees the appearance in issuance, a place in the “organic”, free issuance also directly depends on the money invested.

But their site owners do not invest in Yandex-Rambler-Google, but in optimizing offices. Harm to the quality of the search here is not very big yet - all sites that have spent a lot of money are real, live sites with decent content and design. In principle, the search engine does not care which of the 100 relevant optimized sites to show in the top 10 places. Moreover, all these projects also actively buy contextual advertising from the same search engine.

Only the variety of issue suffers: if you ask “flowers”, you will receive mainly flower delivery in Moscow, and you will not see any other cities or other flower branches (seeds, for example).

* There is a second community - the real enemies of search engines. This is a community of search engine spammers. These promote not real sites, but carrion, synthetic pages with meaningless text, the task of which is to transfer the user to “Rupoisk” or “Notepad” , where they will be shown contextual advertising.
Here, the problem for the search engine is more complicated - it is opposed by a mass of people (in the RuNet - hundreds, in the West - thousands) without special moral or corporate restraints.
This business is not prohibited by law, the money and brains involved there are very impressive, technology is booming. For example, in runet dorveyschiki earn under 7-8 million dollars a year, and I think this is a very conservative estimate.

Here the situation is now similar to 2002-2003 , when spam in the mail seemed to explode and everyone groaned. And the presence of doorways spoils the quality of the search engine very much, impairs its competitiveness.

I think the situation with the doorways will worsen for another year or two at least. That is, the crisis will continue to develop for some time. There, in the doorway business, a lot of money, the business is growing rapidly, attracting all the new energetic people. As they write on their forums - “have you already found your button to cut the loot ?”. And ethically it is not very oppressive, not that postal spam is not a crime either. Many mail spammers can migrate to this area, like porn webmasters, however.

You said, in RuNet, doorwayers earn under $ 10 million a year. In other words, did the money go from Yandex, Rambler, Google and others?

This is how to count. I would not say that this is money that was directly stolen from search engines.

Take, for example, the analogy with the more illegal activities on the sidelines of a large business: here Microsoft usually believes that everyone who bought the pirated version of MS Office, Microsoft underpaid $ 300. Further, the list price is multiplied by the number of unlicensed copies on the market, hence the huge numbers of lost profit, loss of software producers, which are presented at all kinds of pathetic parties such as WTO, G8, etc.

But this is, to put it mildly, a stretch. These buyers of pirated “white” MS Office disks at its catalog price would never have bought them. It will not be possible to convert the pirate market into “white” by 100%. And 50% will not work.

So with doorways, it's not so obvious whose money it is. Look at how the whole chain works on the example of Google.

1) Someone buys a pack of domains, takes scripts to generate text, creates doorways (or rather, a thousand different doorways). Takes measures for its promotion: registers in Google, places links to doorways, etc.

2) As a result, at some (usually rare) request, a separate doorway pops up in the issue of Google and begins to receive, say, 1,000 visitors per month. Having set this request, the visitor sees the "organic" issue of Google and also Google contextual advertising, but for some reason does not click on contextual advertising on the issue page, but switches to the doorway, which he saw in the organic issue.

3) There, it is transferred to a page of a doorway hub like UMAX Search on the Western Internet or Rupoiska on the Runet, with contextual advertising - perhaps from the same Google or from other contextual advertising systems like Overture , Miva (and in RuNet almost exclusively from “Runner” ).

4) There, the user sees the relevant ad, clicks on it and goes to read about the product or service on the site of the ad owner. Perhaps he orders or buys something from him.

5) Then the flow of money goes in the opposite direction: the ad owner pays per click to the contextual advertising system, the system pays almost half of the doorway hub (as its advertising platform), the hub retains 15-20 percent of the commission and pays the creator of the doorway.

Who was hurt here?
User? No, he seemed to get what he wanted.
Advertiser? No, he got his client. He may have paid more expensively for the doorwayer than he would have paid Google for attracting a client, and maybe even cheaper, the grandmother said in two.
Dorveyschik and his "owner" doorway hub - also not for nothing, of course.

It seems to be clearly affected - "Google". Is his money stolen? No, not at all. In fact, he also managed to show this user an advertisement. The fact that the user has not clicked on it is the problem of the authors of the ads.
And if on the doorway there was an advertisement for Google itself? Here, in general, the circle closes, it becomes completely incomprehensible who suffered.

You know, and there are still such individual masters who generally closed this circle shorter - namely, they buy contextual advertising on Google for cheap words / phrases, and show on their page, where the user “lands”, same "Google", according to about the same words, but in a more expensive combination. And the difference is very good live. The difference in the price of ads, of course, must be more than 2-3 times for the scheme to be profitable (because search engines give the site no more than 50%).

Well, in fact, it can be said that all suffered a little: the advertiser may have paid more for the target visitor. The user also received the goods more expensive or not quite that (probably).
Perhaps the traffic was resold quite noisy, untargeted, and the advertiser simply does not know how to evaluate it and in fact pays for garbage.

But most importantly, the amount of “commercial noise” on the Internet has increased, and the issue of Google has deteriorated, being littered with doorways. This last is the main bad result for Google - the loss of quality.

In what direction do you think the search engines should move?

Here the keywords "must" and "in my opinion." It seems to me that search engines need to change the format of the issue in the direction of submitting more structured and developed information, and also create social networks in the spirit of Web 2.0, for cleaning the issue. And the one and the other search engines know, and somehow already done, as far as I know.
It is necessary to create special tools for automatic and semi-automatic detection of doorways. And this is also done, it is enough to read the news, and those who are silent, are now doing this closely.

Which Russian search engine is cluttered with doorways the most, in your opinion?

It seems to me, "Yandex". Exactly for the same reason why most viruses are written for the Windows operating system. Because Yandex is the leader, it is more profitable to spam it. The same about technical effort, and more exhaust.

According to Ru-Center , the national domain of Russia is growing by 4% per month. A significant part of this increase is the domains purchased for search engine spam, Ilya Segalovich believes . What do you think about this?

I think Ilya knows better. He does this every day. Domains are getting cheaper now, it's true. In some areas almost 3-5 dollars.
You can buy them in bulk and salt for future use. Banned the next pack of domains - immediately put into operation the next one. And the domains from this new bundle have long been registered with Yandex / Rambler, just haven't been filled with text yet. It just says something like under construction.
The time has come - I generated the texts, uploaded them to the doorways, waited for the next round of the Yandex index robot, slipped those texts to it, then turned on the Rupoisk redirect for regular browsers, and now you have your own button “hack the loot”.

If the doorway receives from the search engines 1-2 thousand visitors per month, he, apparently, already pays for the purchase of a domain. And now there are rumors that some domain registrars can make the minimum term for renting a domain name is only 3 months. Why don't the doorwayrs even get their own pocket
a registrar?

At one time you came up with "Spamtest" to fight spam, but to fight doorways, are you going to think of something?

While we have enough of our projects. There are few potential clients here - only three, even though they are monetary.
And in "Google", "Yandex" and "Rambler" already have their own departments to combat doorway.

What is the most effective medicine for fighting doorways? What tools, tools may appear in the future?

Most promising here will be social networks. That is denunciations, simply speaking.

We are in 2001-2005 . built a linguistic laboratory "Spamtest" to combat mail spam, it is effective, it distributes updates in real time, it protects only 25–30 million mailboxes in RuNet.
This experience shows that a small team of specialists, armed with smart and powerful software, is fully capable of protecting tens of millions of users in real time. That is, if the program does 99% of the analysis, and the person only polishes, making decisions in doubtful cases - everything works fine. Especially if there is a stream of complaints from Mail.Ru- scale services.
So, if you add a social network for the production of complaints from webmasters, for whom the fight against doorway for a place in the top ten is a very natural thing, then it’s quite realistic to resist doorway technologies and the community of doorway writers.

What do you think about the project “Answers@Mail.Ru” ? Grishin, the CEO, is ruffled: Mail.Ru opens a new chapter with him in the history of search engines in Russia.

I think the project is cool. As another Mail.Ru content project, it will probably be successful, but it will not be able to compete with the search yet. There will be acute problems of completeness of the list of questions and motivation of experts. It is unclear how they are solved.

How to cover all user needs through voluntary responses? It is not clear. Frequent requests to search engines show that there is a “heavy tail” of low-frequency requests, which practically do not repeat. You will not get enough answers for them. And inquiries in search engines Runet - tens of millions per day.
And besides, if the “experts” are self-named, not verified by anyone, then the quality will not be very high. Although as an entertaining and educational service and a place for a party - not bad.

Does another content project Mail.Ru do with you? Tell about him?

No, I probably said it badly.
I meant that the “Answers” ​​project as another Mail.Ru content project is likely to be successful. People, shows, advertising and so on. But as a search engine to replace the "Yandex" and "Rambler" - hardly.

All last year, everyone was frightened by the arrival of Google in Russia, and now he came. How do you think something has changed?

Well, at least the fact that in "Ozone" is now another director ...
It has not changed much (although Google now is able to incline Russian words, which is also quite a lot), but it will definitely change in the fall. I think Dolgov put aside a number of events there. In September, everything will start, prepare quills and paper to comment.

Has Russian “Google” itself learned to incline Russian words, or have they bought technology on the side?

As I understand it, they took something in the open source and screwed themselves to the search.
Perhaps they bought something on the side, from a private developer. In any case, morphological dictionaries now do not make any sense themselves.

When will Ashmanov & Partners launch its own - large - search engine? Yes, you have thematic projects, but is there any desire to get into the garden of “Yandex”, “Rambler”, “Google”?

I have a desire, thoughts, work, I don’t want to tell yet. Who knows how it goes.

How many people work for your company now?

And no one considers them, everyone changes their last name to Ashmanov and sends them to the trenches.
In general, I will not say for sure, many work remotely, scattered around cities and villages, but I think 50 people.

What is your main source of income? What do you earn the most?

On the sale of technology and search engine optimization.

How many companies buy context technology on a page? “Runner” in this sense is your biggest customer, but there is information that RLE and Vebalta can also become your customers.

Few of them are real wild. To buy a contextual recognizer for contextual advertising, you need to have an infrastructure - advertisers, ways of making ads and money. Or start building such an infrastructure in parallel with the licensing and commissioning of the recognizer. So these negotiations are usually long, painful, the customer doubts, he cannot build the necessary infrastructure, his IT-service wants to program everything herself, the customer is testing competitors' engines, but then relations and money are stable.
In general, the usual painful business for the supply of OEM solutions.

There are no more than a dozen or one and a half buyers in RuNet on our engine, but it’s not the number of firms that matter, but the number of millions of impressions per day for a particular buyer. Because we are selling a technology for recognizing a topic (called the “Semantic Mirror”) on terms of deductions for 1000 requests to the resolver's server, or for% of revenues when using the alienated library. In RuNet there are monsters with tens and hundreds of millions of hits per day.

About the specific negotiations can not tell, of course.

What business do you think should be paid attention to in the near future? What seems promising to you? What niche are you planning to take?
I find promising all sorts of paid marketing services on the Internet, as well as all sorts of social search engines. In general, I would like to occupy a niche somewhere on the beach and sleep.

How much sleep per day? 4 hours, like Kutuzov?

I, alas, not Kutuzov.
And as I have to sleep, I have a small child, daughter, the day before yesterday was a year old.
I will be able to sleep 12 hours - I will sleep 12 hours.

What are you reading now?

I read a lot of emails, five hundred or six hundred a day, a lot of contracts, acts and invoices - this is a basic reading, for the soul.
In addition, I read the Tor 100 counter, lists of new personal searches, Novotek, read Habrahabr and Vebplaneta .
At home, I read BusinessWeek weekly, an excellent book by The Tipping Point - about social epidemics, Thank You For Smoking - about bad PR of tobacco companies. I read in English, using “MultiLex” , of course.
He began to read the classic Crossing the Chasm book, but Tutubalin took it away. Fiction practically do not read. « -». – - , 200-300 .

?

. , , .

, , , -, , ?

, , «», , , .…
(«», «») , .
, , «» «» «», , . .
.
— .

, .