How to identify and develop talents in IT: the results of the first Team Leader meetup

Andrei Plakhov : Hello everyone. Today, for the first time, we are holding a meeting of IT development managers at Yandex. We managed to collect a representative composition. I hope the meeting will be useful.

My name is Andrey Plakhov, I have worked at Yandex as a head of the search functionality department, but in general I have been in the industry for more than 15 years and 10 years.

Before turning to the main topic of the meeting, I will make an important introduction. IT development is very diverse, and people from different parts of it are in different contexts, have their own established jargon, are accustomed to different things and live in different worlds. The tradition to call it the different worlds was invented by Joyle Spolsky in 2002. Joyle himself had other worlds, but in 2018 I would single out six of the following.

It happens the development of B2B solutions, solutions for process automation. It happens B2C development of web services or programs for a mass audience. There is a development of internal tools for the needs of one organization. There are creators of video games, creators of software for some kind of hardware, where their laws are completely, and there is science and academy, where everything is theirs.

According to human nature, people from different worlds do not just do not understand each other, look down on each other. What you do yourself seems to be important and difficult, and what the others do is a trivial or pointless thought. This happens often. Somewhere a man who came to work not in a suit looks like a rebel, but somewhere like a rebel looks like a man who came to work in a suit. Yandex is more like an organization of the second type.

Somewhere it is customary to argue about technologies and programming languages, but somewhere they are sitting on the same stack, and everyone knows that nothing can be better.

I mean, I would like to ask everyone to realize and overcome potential prejudices, to understand that all the experts around you in your area have achieved a lot. This means that everyone has something to learn. If he does not look like you are used to or says something unexpected for you, it is not because it is a strange person, but because its context is completely different.

This also applies to the interpretation of the concept of team leader. In different areas, from launching rockets into space to the layout of online stores, everything is arranged differently, and the team leaders are different people. But there are common questions. We need to manage a team of qualified, usually quite young employees, who have a technical mindset, who jointly develop some kind of software with limited time and resources, with high uncertainty. We have gathered here people who are de jure or de facto involved in this - they lead development teams.

What is common for team leaders in all these worlds? Universities that would graduate them do not exist, and the existing theoretical textbooks, books, courses or lectures, usually or very general, are not dedicated to the specifics of leadership in our industry, or, conversely, are devoted to specific programming issues and practices applicable to programmers. not in the context of teamwork.

A person who was once a good developer, and for this he was appointed to the position of team leader, often finds himself completely in a new world. He has no idea what to do now, and, for example, continues to write code 100% of the time, and considers everything else as bureaucracy or chatter.

Tim leader is work. In any work there are sequences of actions that are completely formalized, which must be learned to perform regularly. You have your own good practices, specific practices, good or bad, you can do this work better or worse, there are beginners in it and there are masters.

We want to learn how to share these methods, develop terminology and industry standard.
This topic is immense, so the topic of today's meeting is working with talents, with promising employees who have their own unique personality.

The word "talent" is abstract, let's try to narrow it down. Everyone knows that in the development there are tasks routine, pipeline, and there are more complex. And there are people who solve more complex problems ten or a million times more efficiently than others. Of the competitions in solving complex problems known to me, the most representative is the International Mathematical Olympiad. Several of its winners work for us, I respect them endlessly, but by their example I clearly understand that the complexity of the task also varies. And not all the skills to solve complex problems of a certain type are poorly translated into the conditions of IT development.

It is cool to reformulate and replace the word "complex" with "incomprehensible."

In a situation where the problem seems to be understandable to everyone, but it does not lend itself to full formalization, or it is simply technically complex, and at the same time you need to look at it from some special angle. I do not want to speak for everyone, but most people in private conversations agree that people are divided into two types: those who can solve such problems and those who cannot. This does not mean that the latter are worse, there are many useful conveyor tasks everywhere, but there are tasks that certain people can solve, and I would like to narrow down the definition of talented employees to this.

This property is not determined only by qualifications, a set of technologies, education or IQ level. Starting from some IQ threshold, qualification, and so on, this skill depends on something else. This is not an allegation, it is followed by some statistics, which can be found, for example, in the 2008 book of Gladov.

It seems that, in addition to a common threshold for general consciousness and qualification, talent in this sense is determined by approaches and habits, many of which are very simple. Often, an intelligent and capable person is distinguished from a real star by a lack of understanding of a simple rule or the inability to do something completely trivial, some kind of mental block. For example, there is a rule that it’s not scary not to know, but it’s very scary to hide the fact that you don’t know and make decisions based on incomplete data. However, there is a large class of people who for some reason feel ashamed to show their ignorance, they will better pretend ... This kind of simple blocks often prevent a person from opening up and showing talents.

It is clear that talented employees who show their talent well are well motivated to do it, and not necessarily with monetary considerations. They have something else that makes them do their job much better than others.

I want to tell a typical story of two interns from my practice. Several years ago, two interns came to work for me, both graduates of the SAD, of the same age, with a similar background, both equally well and literally with the same results, had the entrance interviews. And in the first year of work, their results almost did not differ. Then they got stronger, got the qualification, I began to throw more complex, less formally assigned tasks, more cunning ones. At that moment, a clear difference began to appear between them: one of them willingly took on such tasks, connected the brain, found nontrivial solutions, and on the whole clearly grew. The second was offended by such tasks, believed that if the task was not completely clear to him, this is a problem, and the problem is of someone else, for example, me or the person from whom it proceeds. In general, the farther, the greater the difference in the results achieved by them became more and more. In the end, one began to manage a large enough unit, and the second left the company.

This is a trivial story that can be thought-provoking. For example, was it possible to find out about such a difference in the interview? And if so, what to do with it?

After some time, the second former intern in the company returned to another manager, and during his absence something happened to him that his lack of this completely disappeared. And now he has become a man who has uncovered his potential, achieved a heap of results, eventually also became a leader and generally manages a large area of ​​responsibility. And in this regard, I clearly understand that in that very first year I did something with the first intern that allowed him to open, but not with the second. And what it was, I do not know. Then I was a very bad development manager and in such terms I did not even know how to think. I really want to learn and have a certain industry standard on how to think in such terms, talk on how to help people discover talents, what processes there are for this, and what of us can do.

Andrei Plakhov : Let's discuss for the warm-up question that was in the announcement on Habré: all people are equally talented. Do you agree with this? Is it worthwhile at the process level of a company to single out a smaller part of employees, declare them talents, approach them with separate procedures? Is it worth making this status open, public, saying that these are our stars?

Alexey Shagraev : Quite by chance I am the author of this post, you know my answer, it is written there. Yes, I think that all people are talented in their own way. And our common task is to help each other open up and realize their talents. I do not think that the company should have formal segregation in this regard.

Naturally, the company itself provides different opportunities for employees with different amounts of success in a natural way: well delivered, it received a profit. It is necessary that these profits should be sufficiently open and understandable, so that a person understands that if he develops, he develops for something. Obviously, some level of separation will somehow be present: for example, someone is the leader of a group, and someone is not. Someone, we believe, will soon become the head of the group, and some will not. It is necessary that people understand what they need to do in order to go to this level.

My experience says that when a person realizes that the realization will be followed by a clear reward, he will try to do it.

Alexander Gorny : My opinion is that not all people are talented in their own way. In the definitions in which we work, we see that someone is better at solving problems, which means that someone is solving them worse.

In the usual case of an ordinary organization, which for me is somewhere between Mail.ru and RBC, there is no need to make formal categories, who is in which class. On the one hand, it demotivates, especially class B, in some sense it demotivates class C, and it is not clear what it will do for class A.

Well, we have appointed someone in green pants - so what?

In practice, this can be useful when the organization is very large, and there is a filling of new vacancies with personnel from very different departments. If there is a talented programmer in the Moscow branch, and in St. Petersburg there is not enough of a senior programmer, and they have the organizational and legal ability to pull out their man from the Moscow branch, then yes, it’s helpful to have a list of those who have a better job.
Maybe there are other practical cases when this story helps. In companies where I worked, it does not work that way. There is no need to assign any statuses to anyone.

The fact that they are demotivating is written in many books, and this is quite natural and does not need proof.

Dmitry Dolgov : The man who pushed through the mysteries of a technical interview at 1C is already talented. If we talk about the teams working for us, all people are talented there. Of course, each - in their field, on different tasks - in different ways.

I think it is not necessary to highlight this explicitly in categories. The manager should have information about which people are better at coping with what tasks, what tasks you can give them to try, and see how well they will join. The same scheme with the training of new young employees. Someone is doing these tasks well, with pleasure and it turns out, while others are not, just a talented good coder, fast and efficient.

All people in the team need to try to develop, set new tasks, new difficulties.

Does the person successfully cope? Remarkably, it can be further developed in this direction, to give all new subordinates, more and more complex tasks. Not? We need to look for new ways of development for it so that it can develop and unfold.

Irina Fedulova : If we have found people who are better than others to cope with the tasks, then they need to be given more. Here there is an analogy with children, musically gifted, who are sent to a music school; capable of mathematics given in matshkolu. Need help unleash talent. It is necessary to set processes in a company differently, so that people who are talented in solving unclearly set tasks receive more such tasks. So will their talents be revealed.

In one of the large companies where I had the opportunity to work, this is done. There is a secret list of people who are personnel reserve; This especially applies to leadership potential, talents in the definition of Andrew.

When a person gets there, he is given opportunities for additional training, this also helps him unleash talent. Should this status be public or not?

Of course, the fact that a person is in the personnel reserve does not have to be made public, but this person got there because he is a top performer. This is a public status, and in companies it is customary to endure, reward, and publicly praise. Therefore, we can conclude about who is in the personnel reserve, simply by status.

Andrei Plakhov : Interesting: unlike previous experts, your conclusion is not at all that top performers should be publicly known in this capacity. And they said coolly that this is supported by concrete practice, when such a resource, like the opportunity to learn, is spent exactly on these people. It is better concretizing what one can do at all.

Irina Fedulova : This is a special training, like a school or a music school. It makes no sense to give all there. Naturally, it will be management training.

Nikolay Suetin : I will continue the practice of large companies. Promotion, selection is a system of grades, it is known. Meritocratic approach - everyone understands what a person got a grade for.

Starting from a certain grade, for example, from the eighth in Intel, there is an icon in Yandex, fork. The man chooses: either he continues his technical career or managerial. The talent for working with talents is a separate talent, therefore a manager who can find, build, make some incomprehensible tasks somewhat understandable is a specific job. A person who can solve well-known tasks is another.

If someone remembers the difference between a senior researcher and a researcher, a researcher is a person who can independently solve the set task. Senior Researcher can independently set the task. That is the difference.

Andrei Plakhov : Do you think these are two different talents and they need to be divided?

Nikolai Suetin : Absolutely. There are people who are brilliant managers, scientists and developers, but they don’t have enough time to do everything carefully enough. Or you are engaged in the most advanced development, then there is not enough time for management.

Sergey Berezhnoy : Each person is talented in his own way, but not always this talent is relevant for our specialization and a particular company. And, in addition to talent, a person also has a caliber, an average power. Conventionally, he may be talented in music, but the depth of his potential disclosure in this talent may be disproportionate to the average in this company. When we hire people and look at their development, sad but true, we have to dismiss someone, including because the degree of disclosure of his talent is not deep enough, and the person is not of enough caliber to remain in the team. These are fairly obvious considerations. In my practice, I try to attach each person to a masonry of natural stone, combine the edges, so that within the team a person finds his natural place. But if it turns out that the face has the desired shape, but it does not occupy a sufficient volume and does not cover a sufficient area in this laying, then it would be adequate to change this person for another.

Publicity is not very helpful. When we create a single axis: this is talented, this is not, then we have lost flexibility. It is hard for us, the axes of talents are much larger than we can imagine, cast on golden tablets and start distributing these tablets. Because situationally, such a direction may appear and an empty place in this natural clutch, where a different vector will be needed, a combination of some qualities. And it is better to leave it for a more dynamic period, make decisions situationally, on demand, formulate a request with the leaders, what kind of person we need now, and look for someone at a specific moment.

If we cast in iron, that these are talented, and these are not, then we, besides the obvious things with not very good perception of people in the company, lose flexibility.

Andrei Plakhov : Here, in fact, two diametrically opposed points of view were expressed, at least on one issue. Someone said that top performers should know everything, it should be a public status, a honor table, where it would be written who these people are. Someone said that we don’t need such statuses and people should think about themselves better. Interestingly, this is only a legacy in the respective organizations, is determined by the corporate culture or there are situations where one works better, and somewhere else another. Why do not you have what happens Seryozha?

Irina Fedulova : I will explain. When top performer is a public status, this is when a person or even a group of people has very successfully completed something. In one quarter there can be one top performer, and this is a public fact, he was awarded, he received, say, a sign on the table. In another quarter is another person. But if the manager sees that Vasya systematically does more than others, and all the time he wants to put the top performer, but he will not do this so that there is no unnecessary tension in the team, then such a top performer falls into the personnel reserve. We pay more attention to it, give it more tasks, and it starts growing there.

Andrei Plakhov : This is a very clear idea that resonates with me. That is, the public status should not be the status of top performer, but the status of achievement, which can be indicated regardless of status - not just Vasya is good, but Vasya did well here.

Irina Fedulova : True, this is a concrete achievement.

Question : We are talking about great public or non-public achievements, about whether to praise a person or not. And let's ask the next question: so what? We declare this person a public performer, hang on the board of honor. We recently were in the same company, where, seriously, there is a honor roll. And everyone with this okay, and this is not a design institute or something miserable. Great company, there is a roll of honor, everyone likes. What for?

Andrei Plakhov : In organizations that launched people into space, there were initially honor boards too. Do not talk so dismissively.

Irina Fedulova : Why do we need to publicly tell others that someone has done something cool? To praise him, he will be pleased, he will have the motivation to continue to try. There will also be motivation from other people, they will also want to get on this board.

Question : Are there people in the hall with a psychological education? Clear. We can not control the motivation of the person. Let's stop calling it motivation and let's call it manipulation.

Dmitry Dolgov : I will say as the representative of the second side of the public reward: manipulations, any public announcements, awards, honor rolls, mailings, achievements and everything else require very formal signs of verification. Unfortunately, with regard to the development in our industry, the verification of achievements is very abstract. Here the human factor can work a lot. And by making public achievements, honor rolls and everything else, managers or a manager's board can allow a serious failure.

Andrei Plakhov : I know the specifics of your industry and began to understand what the difference is in your experience. In game development, development cycles are long, of the order of years, the team is usually indivisible within the project, until it is launched, and it is difficult to isolate a separate achievement. Most likely, for people who advocate a formal honor roll, achievements that are well formalized and accepted by all are possible. And at the same time they have a finer grain.

Dmitry Dolgov : Yes, there is this moment here. The smaller projects, sprints or tasks, the clearer the result of development. If there are formal criteria, where you can identify who to award, it works. On long-running large-scale projects with a motley team, public selection of the leaders - here the human factor will play.

Alexander Gorny : I heard correctly that now all the experts are against dividing people into grades? No one even has a public list of good people?

Andrey Plakhov : Should a grade be open?

Nikolai Suetin : Absolutely public. When a person is assigned a new grade, he is given a sign. The ability to fly another class, hotels, cars and so on. Depending on the grade of a person, he has different access to benefits.

Alexander Gorny : Grade is almost a position.

Nikolai Suetin : No, this is your qualification.

Alexander Gorny : No, the salary depends on the grade, the grade is in itself something material.

Nikolai Suetin : Optional.

Alexander Gorny : Fly in another class.

Nikolai Suetin : This is important, but for many it is important to get the next star on the shoulder straps.

Question : As I understand it, all the experts said that there are different categories of talents. There are no talent categories in any of the listed companies, although they can be taken.

Andrei Plakhov : The question of the existence of a personnel reserve or a secret manager’s list was voiced.

Question : It is not categorized, it is a general category - personnel reserve. What categories do you select within the company?

Sergey Berezhnoy : I said that formal categories are holding down. My process is as follows: a vacant position appears, I understand what an archetype of this position is, what kind of person is needed there, does he need to understand JS, CSS well, to find a common language with the product team, with the designer ... Each particular place has its own characteristics, and therefore, all you need to know about each individual person is a maximum of information. And then the person joins the real position with the help of the collective expert opinion of the leaders of this person, they make a decision whether this person is suitable for this position and why, or not.

Andrei Plakhov : Let's talk about how to identify talents in time. Even in the system with grades - I will use abstract “grades in a vacuum”, apparently, they mean different things in different companies. We have an employee of the 25th grade and we assume that someday he will become an employee of the 75th grade. So far he is very young, not everything is able, but we have found this grain in him and are going to grow it, including the favored modes, before his cherished 75th, selecting tasks for him, paying special attention and so on. How did we identify him on the 25th? And how to make him understand that he can realize his potential in our organization, that he has, where to grow here, and in order to become, conditionally, a big boss or a respected person, he should not look for a new job, go to the Valley and something such

Sergey Berezhnoy : I think it is necessary to give everyone about the same conditions, and everyone should formulate that everyone has such an opportunity. And then judge by the result, by progress, it turns out or not. Naturally, someone will do it, someone will not, and the need to identify talents in advance and engage in forecasting the future person from the cradle, like in science fiction books, that he is predetermined to become president, and we will give him the right toys in the children's class.

, - . , , , .

: . , , , , , , – , , , , ...

: , . , , , , 75 , .

, , , , .

: . ? , . - , . - , . , , . .

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

: . , , . .

: - , - , . - , , , .

Mail.ru , , , . , , . , , 2000-, , . , , , , . .

: up to eleven.

: , , , , , . , , , , , . , , , , - , — .

— , . . : , - . - , – , . - , , , .
, , - . : « , - , , ». , , . . . , , . ? ? , -, , , , . , , , , , . . , , , : «, ».

, . . , , , , : « , ?». , , : , , , .

, . , . . , - , -, .

: , .

— . , . , , , -, .

, ? , , , . , , . - , - , , , - - . , , , : , , - , . , , . : , , , , , , , corner cases, .

, , , , . – , , , , .

— , . , .

: , , . , , , . .
, , – , - , . , , , - .

: , , . . , . — .

, , cashflow, , . , , - — -, — . — , . , , , , acquisition.

: « , ?». , ? , - , ? - , , , . , , . , ? , , ? - - – .

: , - . , : . , , , , -, , , , , – . , - . ?

, ? . - ; , .

: , , , . , .

: , . , . — - . , - — .
: , .

: , . , . , - , ?

: . , , , , , , , .

, , , . , , -, , , , . . , , , . - , , , , . , , , .

: ?

: 7-10 . , - , .

: . , , , , , . , , , , . , . , , , , , .

: , , , . , .

, , , .

, , ? .

: , . , , . - , , - , - , , , .

.

: , , – . - . , .

: , .

: . , . , .

: , , , . . , , , ? . , , .

: .

: . , , .

: . – – . , - , , , , , , . ad-hoc - , . - , — , , data science. . , . , . .

: , . , . , , , , . . must, , « — , — », - . .

: , , , , , , . – , . , , , .

, - , . . , , . .

— . , , . , , .

Regarding "what, so it was possible." A person in any task has the opportunity to jump a step higher than that described in the task. The question “what could have been so?” Can be reformulated as follows: “everything that is not prohibited is allowed.” There are some limitations in any development. Roughly speaking, there is a starting point where we are located, and we want to get a project that is somewhere higher. Our ideal advancement to it forward: we got the assignment, sat down to work, two years later we were given a wonderful service or an excellent game.

In fact, it turns out wrong. This path is a complex line that will lead us either to this point, or somewhere near. It often happens that in many situations the development or actions of individuals, they do not bring us closer to the point of solution, but lead far away. To prevent this from happening, there are restrictions that must be put in order to avoid making bad bad decisions. It can be in everything: the use of methodologies, languages, middle ware, use of the workflow system, version control - whatever, in all these cases there are limitations. And then everything that is not prohibited is allowed. Any actions of employees who take us up a notch and bring us closer to the solution of the project are definitely useful, should be discussed and recorded as best practices.

Question : Here the conversation is more about the individual qualities of the developer. But they do not work separately, but together. What do you think about this situation: there are three people, they individually work poorly, and together - well, or vice versa.

Sergey Berezhnoy : Exactly what I had in mind when I spoke about artificial laying, that each situation is unique and we select people first of all with faces to each other. And the conjuncture of the fact that we have a team-lead place in a team does not mean at all that this team-leader should have the competencies of a leader and that's all. The conjuncture from team to team may differ. There are people, conditionally, more initiative and pumped in some area, there is less. Depending on this team lead, which we put in this place, must have different competencies. And it's not just a team lead. Just opened a job position in a team, and depending on how the competences of other team members are distributed, what they are stronger, what is weaker, this new hole needs to be occupied by a different person. This is what I call flexibility and efficiency. If we look at these jobs as abstract things, we will lose some percentage of the efficiency.

Question : How do you know if they work together or not?

Sergey Berezhnoy : Intuitive expert opinion. I will think about how they are suited to each other, based on the experience of such experiments, with whom this person used to work, and with whom those whom I share with him worked. Long observation, history.

Alexander Gorny : The universal managerial tool is working here - ask. “Vasya, did you like working with Petya last time?” - “Yes, it was awesome.” And Petya to ask - "It was awesome." So it worked out, the result was good. And if one says that Petya has piled everything up, and I did everything, and he is knocking Vasya down, it means that they did not work well. Always like that. In the forehead you ask, you get an answer. Of course, there are gradations between perfect harmony and non-working. You compare. They do not hide. This is not a secret to learn. It is necessary to ask and find out the answer.

Irina Fedulova : And in the end you can watch what is happening. Another case from practice, when there were four people, two and two worked on similar projects. We see that something is not glued, we decided to rearrange them differently two by two, and immediately went. Just watch.

Alexander Gorny : I do not quite agree. It is necessary to observe very carefully. Swearing couple can be very well done. A silent person can understand each other as a half-word, or just hate, so don’t speak.

Irina Fedulova : I watched the result.

Dmitry Dolgov : Again, these secret managerial lists, characteristics of people, information about who and how successfully they worked in previous tasks. But it all depends on the specifics of the tasks. To put on the routine task of two months ago four leads with pronounced characteristics of a penchant for learning, fundamentally different approaches to the concept of software development architecture is death like, they will rather run away after a week with mats. On the other hand, to put on this project four weak juniors - they will not be able to glue.

With such things, you need to take into account a large number of things. I had real stories when people in a team scattered banally because of different political views, right up to that. For each task, you need to collect, open your secret notebooks, look at how Peter and Vasya work, with whom they work better, pick up leads and direct performers in proportion to one lead on five juniors, mix but not shake, and so on, solve such combinatorial the task.

Alexey Shagraev : I will also tell the story, I want to add some firewood to this question. It also happens that people are one by one cool, and together they are really cool. I have a back-end back-end, a front-end back-end, designer and manager. All of them are wonderful people, I see, they are with an open mind, they like to listen to others and all that. I think it would be cool if they taught each other, communicated, so that they all the time tell what they are doing, work out common plans.

This led to the fact that the designer invented, conditionally, JSON on the frontend, he realizes the design, users interact with this ... As a result, the designer came up with the response format of the frontend backend, so that we could launch as many different visual elements as possible in production, check in experiments that fall into a variety of possible concepts. The transition allowed a large number of experiments to be run from the backend, this is very fast, the designer realized in the process, came up with a format that allowed all his ideas to be realized.

Andrei Plakhov : How did you achieve such a result? We sat in the same room, went to dinner together? What influenced?

Alexey Shagraev : People of different specialties and at the same time all successful, if left in one room, will surely fight, everyone considers himself a genius, and it’s hard to argue with him. Therefore, the beginning was such that each of them told everyone else what they were doing and what was the rationale, and in the process it was necessary for there to be an authoritative person who would tell that what these guys say is not nonsense, it is reasonable. And when they are imbued with an understanding of why something is being done in a concrete way, this interpenetration takes place among them.

Question : Regarding the list of talented employees. At what level do they work? If you are a manager, who has 70 people in submission, and you need to assemble a team of 7 people, all the lists are kept, your team leads are your own, team leaders have your own lists, and then you only cross lists above.

And in what format are these lists stored? Vasya, 2007, worked with this something. How to use them when you have 70 people in submission, and information about each grows every month, it seems that this is all out of control.

Alexander Gorny : I had a maximum of 150 subordinates. Information about who of them is talented and what, I kept just in my head in the format that when Peter gets into the task, everything is fucking fast, and when Vasya gets, then everything is fucking fast, but then twice to redo

Team leads, of course, had roughly the same information, but more accurate and high-quality. And it is important what decision it was necessary to take. Either it is necessary to distribute a new task, then, if it is ordinary, it does not reach me, and if not ordinary, I have several candidates in my head, I will plant them with team leads. If you need to raise someone - the same thing. There was no monstrous Excel.

Alexey Shagraev : I do not have such lists, I believe in all people. I also consider lists to be a rather dangerous thing, because it happens that people earn some kind of reputation and it is impossible to get rid of it. It happens that Petya did a poorly X project, it took five years, he does something completely different, and I got stuck in my head, as he had screwed up on me then. My brain should be free of it. It matters to me who he is right now. Therefore, to keep for 500 years the history of all the successes and failures of employees and present to them that they are not talented enough and three years ago did something wrong ... I do not want that.

Question : Since I, due to my occupation, have to conduct an interview, we are obliged to talk with the candidates about their failures. Good weather doesn't make good sailor. It is impossible to learn something without making a mistake. Each admin will necessarily crash the production base once. We need someone who has slammed.

Sergey Berezhnoy : You raise by hierarchy. The number of subordinates is seven plus or minus two and plus or minus each manager must know about his subordinates in such terms in great detail without any Excel. You may know a little less detail, as Sasha said, and when the time for making a decision comes, you are going to the room, voice the task and make a decision. And if my recommendation is interesting to someone, save your time on making such lists, they will not help, they will not allow you to make an effective decision, in my opinion.

Dmitry Dolgov : Since I made the phrase "secret lists", I will explain that this does not mean at all the entries in a notebook, in Excel under a password or something. This is information that can be somehow classified, starting with the grade, hobbies and basic skills of a person, it can be represented in the head. There are 70 people, it’s physically impossible to keep detailed information on them. There is a recommendation of seven plus or minus two, but usually it turns out who you can work with more or less effectively, this is a maximum of 15-20. Next comes a strong drop in quality.

If some kind of team resorts are needed, if they are not people of direct subordination, but for some kind of hierarchy, then this is either the task of the head of this group, or my joint task with them to figure out how to reformat a particular unit.

Irina Fedulova : 70 people are already working, they are busy with something. We can't just pull out seven people.

Question : Super important project.

Irina Fedulova : And let's say it is temporary. Need to strain yourself and in addition to your responsibilities to do something?

Question : Find seven people, find an important project and then dissolve them.

Irina Fedulova : We had a list for this, which was formed from the visitors and participants of this community, people who are interested in doing something extra to their duties. Opposite the surname was not the color of the pants, but simply the skills, what they know how to do, and what they would like to do, and we played these skills for the task.

Andrei Plakhov : Did you just ask people about it from time to time?

Irina Fedulova : Yes, in the process of internal mitaps.

Dmitry Dolgov : I have experience limited to the gaming industry, but we have a situation that there is a super-urgent project, let's take seven talented people on it, they will correct it in a month and return to their direct duties. It works extremely disgusting. The situation when a person on a project is pulled off to solve other problems is in itself a great deal of stress. And if there is still a need to somehow figure it out, dive into the project that is already under way, started by someone, some team that does not cope - this is rather a strong demotivator than development assistance.

If such a project arises, there it is necessary either to transfer people to a permanent place, or to recruit new ones, or to take one sensible one so that he collects the team. To avoid such a game of dominoes. I have personally and several times experienced the transfer from project to project, and this is a very bad managerial decision.

Andrei Plakhov : How to grow a person with tasks? What to do if right now in the organization there are no suitable tasks for growth for a talented employee, and he is clearly ready to take on more? Or, on the contrary, a new task has appeared, it is clearly difficult, and it is not clear who among them will cope with it, and nevertheless all who have grown to this task are busy and cannot be pulled out. Do you have good practices on this topic? What do you think?

Nikolay Suyetin : It happens that people are jerked off, including top managers, and they sit there, they code, everyone checks, there have been such cases. Therefore, if there is a problem for which you do not have an answer, you need to hire new people that meet the level.

If you have strong guys who, as you understand, have nowhere to grow, they need to honestly say about it that while there is no work for you. And for them it is a kind of signal. You said not to leave. Why? If we play hockey in the yard, it does not mean that you will play well in the NHL. Now, if you play well in the NHL, this is already a criterion. So here. If a person can prove himself in the Valley - this is great. Why do you want to limit it? You can not put your personal interests as a manager above the interests of people who have the potential for development. I think I must be honest.

Alexander Gorny : There are three different situations. There is a department that makes admin panel: every day a new page, a new request is nice to bring out, the work of a monkey - boards in the fence. And next to it is a department that builds a spacecraft, where everything is very cool, difficult and interesting. The person in the first section has no interesting problems. In all my practice and in theory, it follows that it is economically expedient for everyone to drag him into the second section, if he wants and can, he needs it there. And we will find the boards in the fence. And there is no problem, you just need to drag.

Another situation, when no one has more interesting tasks, everyone does something uninteresting. Then you have to quit and, most likely, for you - if your subordinates do not have interesting tasks, then what are you doing here? The industry still allows, you can find something where interesting.

In my practice, the most standard problem is when there is a task, but there is no one to solve it, more often. And we must give it to someone who is not ready, but not at least ready, to help him in battle, if he is interested in sitting until the night, he will be ready. Nothing else has been invented. I have never seen any deep training do levelap. Work till night is easy.

Andrei Plakhov : Yes, a hungry faint.

Alexander Gorny : Yes, but training is not.

Dmitry Dolgov : Regular time pressure in terms of human resources. There are always tasks, and interesting tasks. If for some reason it turns out that for a person there are no interesting tasks for growth, then at least one can find R & D tasks in geymdev, which may not be needed right now, but a person can spend some time on research tasks and grow independently, show results, success, in parallel with the main routine. Routine is always a dangerous thing, it can ruin any person. Talented - even faster than the ordinary performer, who sits quietly from morning to evening and closes the tickets.

Routine should try to dilute the interesting tasks that lie in the sphere of human interests and allow him to grow, to walk from step to step.

Sergey Berezhnoy : The ideal solution to a problem when a person is locked in routine tasks is to try to formulate the task of getting rid of the routine. If he constantly makes up small landings, ask him to make the designer of these landings so that they do not need to be typesetted, so that they are either made by managers or selected by machine learning. When a person constantly does regular work, puzzle him with the invention of the pipeline - this is the best next level task.

Question : I would like to return to the question when there is a difficult task, but there is no person. What to do if you give a person a difficult task, and suddenly you have an external specialist who can cope with it well. What then to do with whom you originally gave her? There is a conflict situation. How to be?

Alexander Gorny : I had an old story: we launched the first version of Mail.ru Agent and the first programmer did not cope with the task very much, we were behind. We hired the second from the market, doubled the development. Things got even worse, they started working together. Everything

I talked to the first programmer, we decided that it was not him, and suggested that he deal with the server. With the server, it began to turn out well, he is now a top manager of Mail.ru, the story ended well - Mail.ru Agent was launched almost on time.

External person must be connected. If the original performer is humanly normal, he will normally accept what he began to help. If the second one shows that he is really better and cooler, the first one will normally perceive that he is better and continues, and I am going somewhere. And if he is not normal humanly, and he still does not cope with the task, well, well, dismiss.

Question : In my experience, the problem is not even that there is no person who will perform the task - there are people, but people are also busy with important and incomprehensible tasks. : , , . , . What to do?

: .

: ?

: . , , , , , . , -, , , . , , — .

. , , , , , - code review , - -, . , .

, , . , , ? . , . , . - , - , - , - , . .

: , , 100% , , . , , , , - , - - , - – , , , , , . - , .

, , , . , , , , , , , -. , , , , .

, , - , 4-5 . , – , , , , , . , , 5-10 .

. , – , ? - , , , . . , .

: . : , , , , , , , – , . - , . , .

: , , . . , , , . , , pet projects, , - , , – , - , . , , , JS, 3D-, .

: ? , , , top performer, . , , . , , , , , - .

: , , , . , , . , — .

, , . , , —, , , , , .

, , : ? , , , . , , , , . , , .

, , — , – .

, . , , , , bad practice , , , . , , 218 , 4, 1.

, , bad practice, , , , , - .

. — , .

: , . . , , . . ? ? , , , . , , , . .

: . , — , . – , – «» - . . , – , - .

: , - . , , . , , , . , — . , , , . , , : , , , , .

, , , , , , , . . , , . , , . .

: . , , - . , , , , , 10 . . — - , .

: ?

: , , . , . , , , -. – , , .

: , — . . , . , - , - 30-50%, — . , , - , — - . .

, 50 . , . , , , , . .

, . : — .

: , , . , , . , - , - , . , , , , , , , . , , , . , , . , , , , below expectations.

: . , . , , , .

: , .

: , , . , . , , , , . , , , - – , « » , , . , .

, .

: — ? . , – - ? 100%.

: . , . . , . , . .

: , , , . — , , . , . ?

: — , , . , , , , , . , , , , , . , , .

: , , .

-, : , - , , . ( ), . , — , , . — , ?

, , , . , , .

: . , . .