⬆️ ⬇️

“Many people choose Agile because they can't do anything else” - an interview with Dmitry Zavalishin from DZ Systems

Hello. This is again “Without Slides” , and this time the legendary Dmitry Zavalishin visited me, known to all as one of the founders of Yandex.Guru (the predecessor of Yandex.Market), founder of the DZ Systems group of companies and author of the operating system Phantom .







What we talked about:

')





All this is in a traditional video interview:







And those who prefer letters to pictures, I invite you to the cut, where you will find the transcript of this interview.



Enjoy!



Speak without slides



- Dima, how do you feel in this version, literally without slides?



- I like very much to play without slides, because, as it seems to me, I usually speak well. I have a picture of the world, which I want to express, and the slides to some extent interfere with this, because you have to follow their logic. And when you say moving in the process, new ideas arise. Sometimes you want to be distracted or swap some pieces ... And very often it turns out that I tell my entire report, showing only the first half of the slides. And the rest are just leafing through with the words: “Well, I already told you this, I already told you this, we went further.” I do not know how to make a perfect presentation - to give 5 slides, in which 5 pictures.



- Metaphors?



- Yes. Graphics or diagrams that really show what is difficult to tell. Although in reality a completely perfect picture is blackboard (blackboard and chalk) or flipchart and pencil. Because you usually need to draw a couple of squares with arrows, show some kind of entity relationship. For me, perhaps it would be the most comfortable.



About their role in the company and life



“But today we won't even give you a pencil.” Tell me better, who do you feel? Are you a businessman, IT specialist, programmer or someone else?



- You know, a businessman is a very abstract thing. After all, in English, it's just a “business man.” And in Russian, this word has some redundant and specific flair. And besides, there is little meaning.



Therefore, of course, I am still an IT specialist. It is clear that we are engaged in a business in which we must always possess this competence, wherever you are in the hierarchy. But I, of course, have ceased to be a programmer, although I still program a little.



If we talk about what I do, then when you start to go to a large number of employees, you understand that your work is mostly not in specific skills. The skills of software development have been better for my employees than for me. My task is the task of the conductor, by and large.



On the one hand, in order that no one was too dependent on anyone, nobody would create unnecessary problems for anybody, and this whole machine would work for a common result.



A completely banal example: if you roughly divide a company into parts, you will get sales and production. And you come to sales, you say: "Guys, we have a plan for the next year like this." Sales begin to sell on this plan. They come back and say: "Well, we have concluded such a big bunch of contracts." And the production here says: “Oh, and we cannot do so much. We do not have enough employees ”... And, not just abstract employees, but specialists in certain areas, project managers, analysts. And you have to build the whole team structure so that you have enough competences and working hands to smoothly go through the whole process from beginning to end. And this is a big part of life.



Another part of life is knowledge management. The fact that all employees should know and be able to. Again, I already do little of this, but I still understand what is happening there.



The third point is generally the process of managing the growth of the company. We cannot guarantee that, roughly speaking, our contracts will come with some kind of regular regularity. There are always surges, falls, market movements and changes in the economic situation in the world and in the country. It certainly affects us. And the ability to quickly respond to surges and declines is also a separate process that largely falls on my shoulders.



Any movement towards new technologies, making decisions about what we are developing and what is not. We are a production company, in contrast to many companies in Russia, we are writing software, rather than introducing western boxes.



About mobile apps vs. web applications



- Give a little more detail about what you do.



We specifically for the customer from scratch create what he needs. In this, too, there is a certain specificity, you need an understanding of what topics will evolve. Here is a simple and very crude example: some time ago it was not clear what to do on the mobile device “correctly” - web applications that were optimized for mobile resolution or native applications.



- Well, the answer is, we now already know - that everything depends simply on the use case ...



- In fact, I believe that the applications as a whole have won.



- If you use every day, then of course. Any fraction of the delay that the web gives you is all, it is impossible.



- And this, by the way, is also a very interesting, interesting situation. Look, it seems to me that in fact, in the mobile device, such an apprehension of some Zen also happened. Look, a simple example - there are two directions: banks and retail (e-commerce).



In banks, mobile apps are already alpha, omega and everything. That is, you can simply easily bury the traditional banking system. The bank has only back office and mobile application. Everything else has already disappeared. And in the future there will be only Sberbank, in which the grandmothers go, and while these grandmothers go to them ...



- Well, the web is probably also alive?



- Barely, barely. Almost completely everything went into mobile applications.



- All these endless details should be manually entered on the keyboard ...



- And now they do it differently. They use machine readable documents, barcode, QR codes ...

And today, a bank without a mobile application is simply not there. And everything else is dying. And retail, e-commerce. By the way, if we talk about automation, care in the electronics of a business process, then in e-commerce it happened a long time ago. Banks, generally speaking, later came for this. In e-commerce, this happened 15 years ago.



But in e-commerce with mobile applications, everything is very bad. Do not live.



- And why?



- Because there is no yuzkeys. A mobile application is needed when something needs to be done right now, on the run, no one knows where. You don't buy a washing machine while sitting in a taxi or on a train. Shopping is a more complex, more responsible, more serious process. They are inconvenient to do on a small screen, they require quite a lot of analysis.



Over the years of growth of the mobile world, it turned out that there are directions in which it is very lively and catastrophically business-valuable. And there it shoots with terrible force. And there are sections in which it is simply not needed. And there it dies.



It seems to me that there will not be a special mobile web, because where you really need it, there will definitely be an application with no options, and where the truth isn’t necessary, it won't be there.



- There are user cases when you want, for example, some feature to be quickly fixed on all devices. Here the web copes faster - laid out on the server and ready. And when you post your application in the Android Market or in the AppStore, then you have two or three weeks to bring it to mind. And you took it, WebView did it inside, and now you have the ability to update on the fly.



- There are a lot of other problems with the web. For example, no offline work. Although we, in general, live in a communicated world, and people are online almost 100 percent of the time, but in fact ...



- There is always another metro or dead batteries.



- Well, the battery - well, we generally lost communication. And the situation is when we have a device, but there is no connection - in this place the mobile application is still working somehow, and the web application does not work at all. And somewhere this is valuable. Some try to hide this problem. The same Facebook, for example, tells you that, they say, there is no connection, but you can write a post, and as soon as you are online, the post will be published.



About custom and box solutions



- And this is probably more or less the right tactic.



- They close the very real problem. So here the situation, it seems to me, is quite established. And, probably, there will be no other way.



But two years ago it was a serious question for us. We clearly understood that the mobile direction is absolutely perfect. It is already conquering the world. But how will it develop?

And there are still questions. For example, is a mobile application purely custom, or a mobile application platform from the designer?



The same problem exists in the "adult", "big" development. Is it a "half box" or is it a pure caste? Here, however, there is also a completely clear answer in this situation. This is a matter of cost. This is a question of what class your user is, where he is in the chain.



You, conditionally, have a wide cheap market and there is a vertical market. Of course, it is more interesting for us to work in the vertical market, in the market of unique large customers who need basically something custom ...



- Type Yulmart?



- Yes, like Yulmart, who sees himself as an absolutely unique player in the market.



- I read a great interview that their turnover for a billion exceeded a year ago.



- You see, it's not even in the back. You can find companies that have more momentum. This is exactly the position in the market. You come, you say: "I'm not one of." This is not a fixation of the result; it is a task.



Do you know why Yandex won all? And because Volozh when all this started, he said: “We must be the first. And it is not discussed. We have no other task, just no. Be the first - and that's it. We are going there.



And they came. Everything is fine. The same is with Yulmart. And in this position it is obvious that you cannot use incubator solutions simply because you have to be unique ...



“But you can't be unique in everything.” That is, probably, you can use some incubator solutions?



- Yes of course. Especially behind the scenes and in places that are not very visible. This, incidentally, is also related to the question that you have a lot of business processes. Some of them are stable, banal and replicable. They are the same throughout the business of this class. And there are business processes that are key to you, and which you sell precisely as your uniqueness. They are most likely unique. There are optimization points that are also unique. Well, and accordingly, when you work as a supplier for this market, you should already approximately understand where you want to go.



A very good example is Bitrix: a good average level is a very massive thing. But if you start to make it unique, it becomes expensive and meaningless. Because in the end, you will be able to saw through almost the entire Bitrix, you will end up with a custom solution anyway, just loaded with the restrictions of the Bitrix itself.



It is important to understand that you must roll out the box if you want to stay in the general traditional class. Moreover, you should also take business processes to which this box is used. Well, because otherwise you will have to finish heavily.



“But this box has boundaries.” In them you can design this process. Look, the whole industry works more or less with us at JIRA. But there the people in the industry themselves are poorly aware that in fact it is the designer of business processes, by and large. And on it you can pretty much everything. This is amazing at all. That is, as far as people do not understand that what they use in everyday life, this is the same garbage, just a side view.



- Yes. We have a business process in JIRA sometimes modified for a specific client.



If we talk about market positioning, then the box and the vertical are understood very well. I talked to people from Yandex who lived on JIRA, and then began to leave it for homemade solutions. Why? Because the volume is huge, the company is very large, JIRA does not pull.



Moreover, as it turned out, it has limitations that can not be removed, even if you customize it actively. They contacted Atlassian: "Guys, we want a solution for several thousand users." And Atlassian replied: “No, we will not do. This is not our market. Goodbye".



- This is generally very interesting. Because, it would seem, a huge customer with good money ...



- Yes. This is a very good understanding of the supplier of its range. And they did the right thing, because it’s inconvenient to climb horizontally. And we are also uncomfortable. We are built as a company that knows how to make unique solutions. Yes, replicated solutions - for us it is just a disaster. We do not know how to do it. We do not know how to collect market data that is needed if you make a box. Because with the box you have to get into the expectations of a large number of typed customers. Analyze how it all works. Find in this place a good common base, which suits many people. Understand that in this place you need to customize.



And we do not know how to do this. And Atlassian is able to do just that, but not the opposite. They are not able to come to this client and understand its specific tasks, and make a specific project for it. It would actually be a separate company. They refused, not their business. As a result, Yandex itself now makes decisions with its own hands. And probably this is correct. This is a very good example of understanding how you are positioning in the market and how you move in this place.



I began by saying that when we were engaged in the mobile direction, we had an e-Legion and the question was how to act. Because in the market of mobile applications there is the same dichotomy. There are companies “mobile application generators” that synthesize a standard mobile application.



- Standard Internet Bank: payments, counterparties, contacts, ATM card. Payment of mobile phone.



- Yes, all these standard things. Which, of course, are very cheap. Catastrophically cheaper. And we must either do this, or go in the direction of exactly vertical. We took the second position, we are engaged in the vertical market - we work precisely for a one-time client who needs uniquely and very well. And which these restrictions do not suit the inevitable semi-circular solutions.



- e-Legion, before entering DZ Systems holding, was engaged in horizontal development?



- Not. He, too, was already positioned in the vertical direction at that time. This is why we coincided so well and we are quite satisfied with what this company is doing. When e-Legion entered our group of companies, of course, they decided whether to change something. But, basically, they left everything as it was.



Before that, e-Legion was focused on a large unique client with high quality needs. So we coincided in this place. The only thing we changed was the closure of the division, which worked in other directions, not mobile, to a western customer.



This, by the way, is a very funny situation. When this happened, even with the old exchange rate of the ruble, our work in the West was a little expensive, and we were not particularly profitable. And so, when we analyzed the entire portfolio of orders, it turned out that Western customers have the lowest margin ...



About three customer rating parameters



- How did you evaluate them?



We have some internal scale. When we consider the order of a client, we look at three simple factors: money, fame and competence. We look at the project and think: “What will I get from it?” - and put three checkmarks. If the project is marginal sanity - put a tick. If he allows us to get a portfolio, which then the next customers will consider reference, we put one more tick. If in the process of working on a project we develop in ourselves some good new knowledge that we lacked, we put the last checkmark. And it turned out that in the projects of Western customers all three ticks are missing.



In general, in our picture of customer quality, Western companies were unexpectedly low. We discussed with our partners in Germany whether it is possible to somehow develop the direction in which we worked. It turned out that it is impossible. Then we just closed these topics, we decided that we would work out Russian customers.



Although now, for example, with today's ruble exchange rate, it is much more interesting to work in the West for money, but, on the other hand, pah-pah-pah, it seems to be enough for now.



- An interesting story about the vertical market. I correctly understood that this term means that you have a small number of large customers with whom you do more or less everything IT?

- Here the situation varies by group of companies. Because in mobile development there are usually dozens of projects a year. And in the DZ projects unit.



- Do you remember how Mikhail Samarin on Mobius said that their average mobile project is two people for two months?



- This is not enough. We usually have five people for 3-6 months. And then there are accounts on which we work for years. Bank of Moscow, for example. We have been developing and constantly developing this application for him for many years. Probably, someday he will want to change the contractor and everything will change.



But overall, this is exactly the kind of account situation in which we close a certain direction for our customer with long-term work. There are, of course, one-time projects. For example, Afimoll, a shopping center in Moscow. We made for him an internal application that gives navigation to the shops inside the shopping center. This, by the way, is a very interesting experience, especially if you return to the question about e-commerce as a whole and about the failure of mobile applications in this area. We have done several projects on navigation in large retail chains, and this is a popular mobile business case.



- This is for the zone, where is the buyer? Or for a warehouse?



- Yes, it is the buyer who puts himself mobile apps.



Here Afimall, it is huge, it has a lot of things. And the navigation tool that helps you, falls very well under the mobile use case. Suddenly right here, right now, an urgent need. The mobile application answers this task perfectly. Well, again, returning to the casualness of the order - it was such a casual order. The customer came, we made him a very specific application addressed to a very specific task. But this is not an account, because the application is done - everything works - goodbye, thank you. And there is no long history with this client. And for a bank it is a long story, because the bank will develop its mobile application all its life. There is a real brawl on the market of mobile applications for who will do the best. And a good application is a great value for the bank.







About the problem of finding good frames



- You mentioned the production department. You said that there are situations in which the sales department comes to you and says: "We have a million-dollar order, we must do it." “Good,” you say. But where to get people?



- This is actually a big problem for Russia in general.



- It is Russia? That is, in the world there is no such problem?



- Unemployment in Russia is less than in the world, and significantly less so. We have, if I'm not mistaken, there is something in the region of five percent. And in Europe, about 20 percent. That is, we have a situation preduratsky. Our country is really loaded with work above the head. There are not enough working hands. The economy needs to develop, but it has no place to develop. Because there is no one to actually do it. Well, really no one.



- Why did it happen, and what to do about it?



- Here, here in Sochi, the RIF took place, of all the regional RIFs, the worst. I participated in the HR section there, colleagues from Headhunter and Superjob were sitting nearby. And Vladimirskaya. And in particular, I rolled out a claim from the customer. I say: “Guys, you see, I’m really very dissatisfied with the quality of headhunting in Russia, because you bring it anyhow. In fact".



I sit, they talk. “You understand that when you say“ programmer ”as a profession to each other, it shakes me because there is no such profession.” Inside the programmer there are still one and a half thousand cuts, which are not compatible at all with each other. And to appeal to you with the task of choosing a real person who really needs me in this position is unreal. At the very least, you can still search for keywords.



And then very different people come. We are now actively recruiting, for example, analysts. I talk to them, and it pisses me off. Not that bad people come across. No, they are good. Positive. Workable. Very motivated for the result. It is evident that they have done a lot in their lives, they are not afraid to include them in projects. They will plow. But at the same time, these are people with a very fragmentary knowledge of what they do. And this knowledge is basically just gathered from work experience in some similar areas.



- That is, they have no base just?



- Yes Yes Yes. It is evident that they have no serious professional base at all. I'll give you an example. We have several case studies for analyst testing. Well, a simple question: “We are going to release a new Skype. Tell me, please, any tasks that we should put into the basis? What do people do to him? ” Do you know what the analyst answers? "Files transfer".



- Well, this is serious. For the analyst.



- Do you understand what the problem is, yes? The problem is that there is no “ride” for the car. There is no such person who says "Let's drive a car."



- Yuzkeys - go somewhere.



- Go to the country. Go to work. Yes, even in the store. You see, yes?



The task is, it is a little bit behind this. There is no “send file” use case. And pay attention: it so happened that I was just on an interview with some of them on Skype. You see, and the analyst is sitting, talking to me on Skype, and he cannot take out my jazz case "to have an interview on Skype."

You see, yes? And this is a good analyst. He is, in general, quite good. This is a man who worked, wrote documents. I looked at these documents - they are imputed.



On CodeFreeze, I told you an enchanting case about an intercom that didn’t have a “no unlock” function. That is, it was designed by a person who simply did not know that there are alternative scenarios.



The same thing - I am interviewing an analyst, I say; “Well, look, let's discuss the setup of e-commerce with you. Let's say mail order delivery. ” He describes well. Perfectly describes. Directly the truth passes some even unobvious things. I say, “Well, fine. And if not delivered? ". He: "They sent and did not deliver?". I say: "Why sent? And if not sent? ".



- In general, this is surprising, because the basic book of Coburn about the use cases explains well that there must be scenarios, in each place - one variant of the refusal is worked out. Otherwise it won't do at all.



- In fact, this is also a separate funny song. The fact is that exceptions are a problem from the point of view of the business itself. Because for business, the mainstream business process is always licked. And when your exceptions start to pour suddenly, it breaks down the work very easily.



This is also some kind of human thought. Because from the point of view of psychology, this is a negative case. He is such that he does not want to think about him.



And then there is another problem that is also in the area of ​​the task of analysts. And about which, too, very few people understand, until it works with large complex systems. I had a smart example.



That was a long time ago. An order came for a description of business processes in some large company, an Internet provider. They implemented a new ERP system. And we performed purely analytical work for them. Our analysts have written huge documents. Huge. The final document was pasted on the wall, and he occupied half the wall. Business processes are also written with text. We really picked up a huge job on which the system was implemented. Not by us. And when this system started, and worked for a month. And at the end of the month, a certain “Aunt Lena from Accounting” said: “Where is my Excel? We had an Excel in which everyone entered, what bills were issued to customers, and I then went in and corrected for Vasya and for Lena - we have oral agreements with them about what we have there, for them we make such a discount . I corrected in this. Where is he?".



And nowhere. No more of it. Everything. Goodbye. It happens in the system, and there is no place in it where Aunt Lena can climb and fix something. What is the problem? The fact that it was not possible to reveal the presence of this business process was possible. This is not a mistake of analysts, it is simply an immanent property of our Universe. You can never guarantee that you have completely covered the whole map of knowledge about a certain area. Just because there is no certain verification tool that you have covered.



This aunt Lena, until the day of invoicing came, simply did not remember that she was doing it. She does it on the full machine. When they ask her what she does, she simply did not remember. And this business does not close. Books are being written on this subject, and a lot of effort is being made, but this I have shown a very unique case. For all my life - and I, there, for more than twenty-five years, I have been doing this - it was one such case, when a process came to light that could not be identified. But in general, the taxonomy of work on the requirements, it is about that. About the fact that if you will ad hoc collect them, somehow on your knee and situationally, then you will miss a lot of things. , , .



Agile —



— ?



— . Agile — . Agile - , Agile, .



- . : «, , Agile, - , ». , in-house, . , , . , Agile,





, Agile, . , , , Agile , , . , , , . Agile .



, Agile , , . – : « Agile». : « RUP?». , , RUP , , RUP , Agile ( ), .



90 , agile, RUP . Agile , , . .



,



—C, . . — PMBoK, ISO9000, CMMI…



— CMMI , , . , CMMI , . . , , , , CMMI . CMMI.



. , , . , , . .



, . , project- . , . , , . «». , , , . , . . , .



, . . - , .



— ?



— . PM-a ?



— , , : , , . .



— . , . , , , .



— , .



— , ? . . , . . .



, . . . .



— ?



— , , , , . , . . , – . «», .



— , -?



— , , , , . , , , project plan.



Why is it important? -, PM, . , . -, , . . , « , PM ». , PM , .



. ? - trade-off'. , , . , . .



, , . — ? , . , . , - , , . : - , . .



. , . . , , , , – . .



, ? . , .



— ?



— .



— ?



— . . , 10 « », . . , , , . . , . .

, , , - .



— .



— , , , « » . . . .



,



, , , ?



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



— . , .



— . . . , , - : «. – ».



— , « » — ?



— . .



, . , , PR-. , , . .



. : « – ». Okay. . , , , , , . , , , – , . – .



: , . , . .



— ?



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



: e-commerce . , . , , . : « ; , , , ». : «. ?! . , ».



: « . , . . ». . . : « , , . «», . . . . «» . – «». «» — , «» — . ».



. . , : « ?! ?! , - ». : «, , - . ?».



— ?



— , . , , . : «, . , ».



, . , – . , …



, : «, , . , , . , – . , . , 200 /, … , . , ».



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



, 90% , 5% . . . , , . – , , , , . . , . , , . , .



Everything. --. , . ? Can.



«» – , -. . -! . , . – , , , , . , , , , , . , , . , - .












:



Source: https://habr.com/ru/post/269803/



All Articles