The main theses of ABBYY CEO Sergey Andreev from an interview with Vedomosti

The CEO of the Russian company ABBYY, Sergey Andreev, gave a lengthy interview to Vedomosti.

About the degradation of the Russian IT market, the current and future state of affairs, and competition from his position was told by a person with the deepest experience and expertise in a highly competitive market, where ABBYY is represented with its products.

"Megamind" cites the main points that were voiced in the interview.

On the crisis of the international market

The crisis that began in 2014 did not have time to harm those Russian IT companies that were able to establish international business. For the most part, they are software developers with currency earnings and ruble costs. What to do with the resulting difference - each company decides for itself, and ABBYY Sergey Andreev hopes to make a revolution in the market for searching and analyzing data using his Compreno product. As planned, ABBYY Compreno will be able to find the necessary information not only using the words of the search query, but also using the connections and context between them. It is more correct to search for information not with the help of words, but through the meaning laid down in them, I'm sure Andreev.

According to Andreev, the usual keyword search is outdated, if only because of the growing volume of unstructured information, and it should be replaced by a meaningful search, which Andreev intends to accustom to the corporate segment.

About import substitution

It is impossible to ignore. When you see that large international companies are forced to inform customers that they will not supply any products or provide technical support, the government responds: “We have to do something.” When key companies are threatened, governments try to stop these problems, and various initiatives emerge.

Now about import substitution speak in the context of security. Until 2014, this was discussed only in the context of the economy: if the local producer has a large market share, more money remains in the economy, more jobs are created. Now, more began to worry about dependence on imports. For example, there is a technological cycle of production of components for a drilling rig, there are designers, but it turns out that they work on imported software. As a result, the drilling depends on imported software. Nobody wants to live in such a risk now. In such conditions, if an analogue appears, it will be bought, even if it works worse.

- Why in all discussions about import substitution is usually not heard by consumers?
- Good question. Apparently, the organizers of the discussions have never seen consumers. Well, or maybe they didn’t think well before they proposed the initiative. It seems to me that you can think about import substitution, when the consumer says: "Yes, it is necessary."

About competition

Just to deal with a very large and very strong competitor is more difficult than with a smaller competitor. In all major segments - operating systems, databases - Russian companies began to work not the first in the world, although there were some successes. For example, the Lexicon editor dominated at one time, it did not have an office suite — a spreadsheet, power point, etc., and when competition arose with foreign vendors, they lost ground.

There is competition, it really is the engine of the market, but any system can be hacked. The free market system is hacked by monopolies. As soon as the dominant players appear, the competition does not work as the market would like. Free competition between large and small companies always works not in favor of the latter.

The US market has developed a very interesting situation with software and hardware patents. This is the only country in the world that is so treated with patents. In fact, there you can patent an idea, not an implementation. Patents of a very general form are obtained, which can later be interpreted as the invention of the Internet, for example. I remember well the case of Kaspersky Lab, which was sued because of a patent for user interaction with the program - feedback from the program.

Therefore, a new IT company is looking for segments that have not yet been mastered by giants. When searching for a niche it is important to work all over the world. If you have a niche only in the local market, it brings little money. A competitor who works all over the world will definitely come to your country, and you will not stand competition with him. This means that the United States is a very important market and you definitely need to be there. And if you license your technology to other companies, then you need to guarantee them that it is clean from the point of view of patent law and you will protect them. As a result, the patent system that protects the US market protects companies that have more patents. And who has a lot of patents? For those who have long held a market share. Therefore, we can talk about free competition in the software market only in relation to a thin layer of beginners. Here everything is really active and alive. Startups and ideas are dying and being born for a thousand a month - something interesting is born from this brew, but then, it is true, it integrates into large companies.
Therefore, the answer to the question of why we do not have large players is simple: it is very difficult to compete with global companies sitting in the local market. Someone succeeded, for example, 1C, but this is the exception rather than the rule.

About the national operating system

Attempt to create an OS from scratch by most experts will be perceived as a utopia: it takes a lot of people and time. Approximately 5,000 people who will work for at least three years. If the work of one person is estimated at $ 100,000 per year, the annual work of the team will cost $ 500 million per year, in the end, you need $ 1.5 billion over three years. And this is just the operating system. And another question, succeed or not.
You can go the other way - to take as a basis something that exists and make your own version. Then you need about 500 people to keep track of all the changes. But this is not from scratch.

About state and IT

Speaking about the magic power of the market, we must remember that it is often needed and useful, but it can be hacked. And most often the market is hacked by cartel collusion, monopolies, and then it stops working. In terms of software, the situation is both market and non-market. There is competition between large companies, one way or another they collide with each other, and there is a market there, and the same competition exists between small companies. But on the transition of a small company in a big competition there. For the time being, this state of affairs suited everyone, and with soft measures it was possible to encourage domestic companies to develop faster than international ones.
In Russia, the situation is different due to several public failures of large foreign companies to supply software to some Russian companies. Therefore, the state should think about its function of providing business. Drilling platforms should be made. Naturally, the state at such a moment will talk about any measures that will ensure the maintenance of work. Can these measures be called market measures? Of course not. Is it possible to call market measures when you are refused delivery and break the existing agreements, because some sanctions were imposed on you? Also no. Here the market does not smell from one side or the other.

About user personal data

In other countries, the same laws apply. In the US, for example, it is forbidden to store data of Americans outside the country. When a US court wanted to get data from Europeans, Microsoft refused. Europe also prohibits storing the data of its citizens outside of Europe. They also see the risks.

Our judicial system is not perfect. But, having become acquainted with the American judicial system, I can say: what Microsoft did is phenomenal, since the company went against the judicial system. Imagine: you are the defendant. The plaintiff says: I want to get all the data that is inside the respondent company that are related to the case we are dealing with, I want to get it to prove my theory that you are guilty. And the respondent company must provide its data from the servers and the paper data to the plaintiff’s lawyers so that they can read them and select those fragments that most conveniently confirm their theses. In fact, the company must testify against itself. There is an amendment to the constitution: a person should not testify against himself, but this does not apply to companies. Failure is tantamount to a lost process. Incomplete disclosure, if they prove that it is not a mistake, but a deliberate action, also a defeat. The decision of the judge must be executed in any case. Microsoft's failure is fantastic. But the precedent with Microsoft is very bad, because here one country has demanded to disclose data that is posted on a cloud resource in another country. If before everything was moving towards globalization, now, on the contrary, the walls are reappearing. Therefore, there is nothing surprising in what is happening here in the Duma: there is a global trend.

About income

Even in IT it’s impossible to plan for five years, and when it comes to political trends, it’s impossible to predict at all. When the dollar rate rises, it’s good for a company like ours, because we are exporters: overseas markets provide about 80% of the revenue. But because of the large volume of sales abroad, they are quite large and the costs associated with these sales. The very presence of currency earnings is favorable for Russian units.

Both factors act because the Russian IT market as a whole degraded last year, and it seems to me that in absolute terms in rubles it is noticeably less than it was in 2013. It’s hard to say how it will be in 2015, because they are going unevenly: indicators in Q4 are on average always higher, usually around 40% of annual revenue. This is an element of seasonality: many large organizations close their budgets in the fourth quarter, and projects approach the implementation phase. In general, in the corporate sector, the sales cycle is quite long: if six months is already good, usually closer to eight: as a rule, what you start in January is completed by the end of the year. Large organizations save at the beginning of the year, some budgets are not unfrozen or there are other organizational problems, and by the end of the year the race begins, which is why a lot depends on the fourth quarter. This is typical for many countries, not only for Russia.
Last year, the fourth quarter was not very favorable: revenue in IT companies in Russia generally slipped, but everything turned out to be not as bad as we feared. In each quarter, we didn’t pick up about 20% of revenue in rubles compared to last year, but December turned out to be good: the increase was 9% year on year, it was even unexpected.

Not only the course besieged the business, the quality of the market also sags, this year the trend will manifest itself more. People buy less. In general, IT projects are always associated with investments: companies invest in software products to make something better, more efficient, faster. And when the economic situation is such that you do not have enough money for turnover - to purchase materials for production, to pay salaries, then there is no time for investment. Therefore, in a crisis, the IT market is shrinking one of the first - when companies revise budgets.
This year the situation is very conservative. The financial market is unavailable, it is difficult to use credit resources for development - perhaps someone succeeds, but the financial proposal is very degraded. Investments are greatly reduced, it affects IT budgets radically.

Lack of money in the economy - the most significant factor that hinders many projects and ideas for development. The crisis is not yet so powerful as to freeze everything, many companies and organizations are focusing, choosing directions that are extremely important to keep, continue to invest in it and develop. And around these areas there are business opportunities, including in the field of IT. But the situation is unfavorable for the business as a whole.

About products and their development

Europe is growing well here, but it is not the most profitable business. If you compare the regions, Europe is the most problematic in terms of profitability. And the revenue is growing there. The reason for the imbalance between revenue and profit is in the multinationality of Europe. To sell a product there, it is necessary to have local representatives in each country, it is necessary to localize the products in different languages ​​and ensure their support. Otherwise it is difficult to sell. All these conditions make the construction of business much harder, it is necessary to increase the infrastructure and hire more people per sales unit. As a result, Europe is becoming less profitable.
The best profitable country is the USA. In general, historically, the US and Russia have always competed in terms of profitability. But the disproportion smoothes out.

Historically, in Russia, 50% of revenue was spent on financing the development. A similar figure in America. In Europe, it is closer to 35%. And this year, I suspect, the share of Russia may fall to 35% - to the European figure, but this is still not clear. Unfortunately, it is rather difficult to predict for 2015: the beginning of the year was difficult, but whether the end of the year will be bad is unclear. Last year also began difficultly, but gradually the projects began to unfreeze, people looked around, understood how the sanctions work, which projects survive, which do not. And after that, those projects that survive have gone. This year may be the same.

About borrowed funds

- Do you attract loans?
- Yes, first of all as working capital. With our dimensions, you need to have certain reserves in the accounts to ensure the timely movement of money. Although the main expenses are associated only with salary payments, in a large company such payments cause significant fluctuations in cash with uneven revenue receipts. At the beginning of the year, revenue is declining, at the end of the year it is increasing, on large contracts, revenues also go in steps. Reports on major licensing contracts quarterly. It turns out that the first two months in the quarter are rather scarce, the last one is rich. Therefore, we need working capital in the accounts, and for this we attract loans. It is possible to increase loans and use them for investment instead of investment funds, but now it seems risky, so we do not do that. Theoretically, this is a more interesting tool for the development of the company, rather than investment funds, which invest money and impose significant restrictions on the work of the organization. The foundation of the logic of work is very simple - you need to maximize return on investment: they aim at a profitability of 35% per year, and they structure the transaction according to the principle of “give and not sin”. If the company did not grow by 35% per year, you will have to sell more shares, but provide income. This is the logic of the actions of all venture funds.

- Where is credited?
- These are mainly foreign and non-state banks. Among Russian banks, there are a lot more able to give money to companies like ours, only a dozen, abroad. Sberbank tried to go into this niche, but it will take a long time to go: a large, complex organization, and the envelope is not the same. They can finance any unit for the Sayano-Shushenskaya HPP. How much do we need? A few tens of millions of dollars. Is this a scale for Sberbank? Although they tried to move towards IT.

About diversification

- How sensible is the idea to divide the company's business by region? EMEA?
- You can share in different ways. There are pros and cons to any method. To do long-term and large projects, you need a local presence. In many situations, people buy not a product, but a competence. When they see that at the other end of the wire a specialist who is very well versed in the matter is talking to you, they have confidence, and they buy it. Therefore, it is necessary to appear there and be sure to understand what is happening with the client: how he lives, what is important and unimportant for him, what can improve his life. Local presence is the key to everything. Not even knowing the language, because to some extent people are ready to go on communication in some unfamiliar languages: they just need local support. In Japan, for example, there are a small number of people who speak English. If you build a normal work with Japanese companies, you need to open an office in which engineers who speak Japanese will work. A critical factor is trust between organizations, otherwise there will be no business.

Businesses that are aimed at the mass consumer work differently: from one point through the means of communication, such as the Internet. But it became more difficult to work in the mass market with the advent of a successful business model of an Internet search engine that works like a black hole. Search engines began to include an increasing number of mass services, replacing mass products with free services. Compete with free services is impossible.

About phenomena

Search engines have not always been successful businesses. They were interesting projects for a long time, but for ten years they could not make money. And at some point, the model with the sale of keywords, which led to a significant distortion, worked - they became super-profitable. And they could not only pay off through the search, but also develop free services, which they add to the basic functionality. They earn money on what they earned, and they have not come up with anything new in the area of ​​income generation.
The black hole factor, which draws in a lot of segments, leads to the fact that an attempt to do something interesting in the mass market ultimately leads to competition with free services, which is frightening from the point of view of business logic.
If you understand how to earn money by deploying a free service, this is good, but there are few successful examples.

There are massive services not produced by search engines, but this is more like an exception than a rule. The Google Translate service appeared a long time ago, and in the iPhone - the technology of the company WorldLens, which they bought: you direct a mobile phone to a piece of text, and it gives you a translation. They bought the company and now provide its services for free. While WorldLens was an independent company, it did not look commercially successful, but hit the air and provides a good, free result.
Creating a mass service is a challenge. And so that such a service was bought - the call is doubly, this is the next level of skill, and with all due respect to Google and the rest of the operators in this market, I still do not see such masters. Their services are good, but do not pay for them.
I see so far the search engines have one way of making money - free services that catch up on traffic and ultimately sell keywords. And this black hole strongly incinerates the landscape of mass products. And so I am seeing the shift of companies from the mass sector towards b2b. Because there is not enough mass free service, customization is required there: there is another model, and search engines cannot go this way.

And about competition with search engines

When Google released a free translator, we were forced to change the direction of development of Compreno, which developed as a new generation machine translation technology. But even if you make it better than the existing, but free, then there is no business there. That is, some business is possible, but too small.
Therefore, Compreno also drifts in the direction of data analysis, on processing the stream of unstructured data, on pulling out events and facts, on smart search technologies. This is again the corporate market. And the translator had to be pushed aside.
Elements of the development of translation technologies and translation services remain, which are formed at ABBYY Language Services and are aimed at the corporate client. They have a short translation service, which can be qualified as more massive, but I think that this is still b2b. And this is definitely not a key priority.

- What could ultimately threaten search engines? Could something turn their business around?
- Theoretically it can. If the people who own the information to which the search engine provides access (the search engine does not own the information, the search engine is the pipeline, and the gas is with the people who made the content). You come to the search engine, you need content. And the mediator between you and this content starts to make crazy money. In theory, if content owners agree that they do not allow the search engine to use content for their work, index it, if it does not share with them the advertising revenue they receive from selling the words leading to this content, then a balance will arise . The search engines go super profits, and then they start thinking about profits, and not just how to make the world a better place. And at this moment the world is changing. Therefore, Google does not like the topic of copyright and in every way tries to drown it, explaining that this is inconsistent and wrong, the right to information is something obsolete.

About planning and investment

It is advisable to build prototypes of programs: so that they appear every six months or a year. It is necessary to evaluate how interesting such a product is - if there is demand, you need to finish it quickly, in about six months, and release it. Cycles are shorter, planning can take only a few weeks. Remade, checked, if something is wrong, brought to mind and offered to customers. The key is interaction with customers. So iterations closer and closer to happiness.
The project cycle of 5–7 years is an exception, it makes sense to make it when there is a steady trend. For this you can get involved in a fight, but more often people prefer short terms.

With fantastic speed, the amount of information that people store in computers and storages is growing. Previously, there was a task to simply put information into a computer, and now they have placed so much in it that using such an array of data becomes a serious challenge that requires new approaches to working with information. ompreno is designed to deal with today's triumph of unstructured information, which is very difficult to handle. For example, you know exactly what you need in this piece of iron, but modern tools do not allow this information to be pulled out. You try to find this way and that way, you use different types of queries, combinations, it still does not work. We brought the computer to work not with words, but with the meaning behind the words, to a combination of meanings. We think in categories, images, and then convert them into words. There are systems based on rules and statistics on the market today. Our technology is based on artificial intelligence, semantic-syntactic analysis. This gives a different quality result. Such technologies are needed by the market.

About semantic search

Compreno is a basic technology, on the basis of which you can create many completely different solutions. Some areas do not exist yet, and it is necessary to form demand and change business processes. Some areas for us are clear and defined by the market. We will talk about these development directions in the near future. By the way, exactly one year ago we presented the first solutions based on technology - solutions in the field of intelligent corporate search.

Not everyone needs magic. Sometimes you can solve a problem in simple ways - and the attempt to remake on the basis of magic does not meet with understanding. From the good of good not looking. But there are many situations where the simple way to solve does not work. Suppose you remember that you met half a year ago with guys who talked about environmental products. You are trying to remember when it was and where, in order to find a correspondence, scoring key words - “meeting”, “ecology”, “Petersburg” - it is impossible to find. But it does not work, because in other words it was written. Not ecology, but green technologies, not Petersburg, but Vyborg. Search for the word "technology" is generally wrong - it is in every fourth sentence. Key words need information is difficult to find. A person remembers everything approximately - and often with the wrong words. Often the question is not in a specific synonym, but in a combination of words, then the search generally collapses. The main trick - we see not only the facts, but the connection between them. When we parse the text, we make it a complete analysis - we parse the sentence syntactically, we understand how concepts are connected and why the word is used in one sense or another. This removes the ambiguity. Except us, no one pulled this connection. That is why we feel confident.

Our intelligent search will improve corporate search engines. Compreno is able to search by meaning, that is, to find information that can be expressed in other words.
We focus on b2b solutions. People were taught to search for information by keywords. And they learned to do it. And all business processes are based on this model.

We are standing on two pillars that ensure our bright future: computer vision and computer linguistics, all these are elements of artificial intelligence. You can move in these directions, you can add something, but for now we plan to focus on these two elements.
The main business - recognition - is quite profitable, but all that we earn, we spend on computational linguistics. There is income, but so far small. When we manage to bring this segment to the trajectory of beautiful growth and we will make good money on it, then we will start looking further. While we go into the depths.

About geographical and product niches

- There are geographical niches, where you have not gone and do not plan? African countries, for example.
- The third world countries - Asia, South America, Africa - have a good potential: they are growing faster and faster than in Europe. You can improve work with languages: we have Chinese, Korean, but you can improve, you can work with Arabic. There are financial resources in these countries, and now they are invested in computerization. China is an interesting place with good potential, but it is difficult to understand how everything works there, and so far, unfortunately, it is not possible to find explosive growth paths.

- And by sector?
- Historically, we loved to make universal products that can be customized. , , , , , , – . .
– . , , PassportReader: - , . – , .
, , – , , , , . .
, .
– , . .

– , 90- .?
– 90- . . . , . , . , . , , , . . – - .

- Chinese software can cause problems for you?
- Any competition creates a problem for suppliers and benefits to consumers. Here is the question of where the border of using market and non-market methods will go. Competitors from China will come, we will deal with them.