YT YouTube2Text

YouTube2Text → Transcripts → 1_Продуктовые исследования_07.10.2026

1_Продуктовые исследования_07.10.2026 — Transcript

by AI Talent Hub · 19,705 words · 2,998 segments · language en · Watch on YouTube

Open in transcriber Summarize with AI Download SRT/VTT

Full transcript

0:02Radomir, you’re a product manager at
0:04VK,
0:05>> right?
0:07>> Uh-huh. What are your expectations for
0:11the course, Radomir? Do you even have
0:13any, Romir?
0:15>> Well, look, I know what product
0:17research is, so my expectation is just
0:19to finish it successfully, let’s say.
0:23>> Yeah. Okay. In short, you don’t
0:25expect to learn anything new. You just
0:27expect to cruise through it easily
0:28based on your experience.
0:31>> Well, they spoiled that it won't be
0:32easy, there will be interviews. I even
0:34already know where I’ll conduct the
0:36interviews.
0:38>> Ah, okay.
0:41>> Hey, don't scare those who aren't
0:43product people, and who are actually
0:45techies from X5 Tech. There. Hello
0:48everyone, guys. Some interviews,
0:50talking to people. Well,
0:52>> yes, you’ll have to talk to people.
0:54Ivan.
0:55>> Oh my god. That’s my favorite
0:57activity, by the way, can you imagine,
1:00actually, because you kind of get tired
1:02of programming.
1:03>> Well, exactly, who needs that
1:04programming anyway. Especially,
1:06especially, yeah, you understand
1:08yourself, a little time will pass,
1:11>> it will get even better with Claude.
1:17We’re a bit too.
1:18>> How much better can it get? We’re
1:20energetic. Yes, yes, that’s how it
1:22should be. It’s just that you have a
1:24two-hour time difference, you
1:26understand? We have. Yes. 9.
1:31>> Sofya, our job is to charge you up. In
1:34general, get charged up. Sofya. Ivan,
1:36tell us a little about yourself. What
1:39are your expectations for the course?
1:40>> That’s right. That’s right. Ivan.
1:42Ivan, I’m coming back to you. What
1:44are your expectations for the course?
1:47>> Expectations? Well, I specifically
1:49enrolled, I’m in the ICT program at
1:52Nkhab. I really want to boost all this
1:56product stuff to understand: how to
1:58sell products, how to position them,
2:00who the target audience is? Because
2:04right now I’m mostly covering
2:06internal products at X5, but I would
2:08still like to gain expertise to then,
2:10well, maybe switch to a position as
2:12some kind of product owner, not just a
2:14dev. So I’m positive and I picked
2:19everything product-related. I signed up
2:22for all the courses.
2:23>> Got it. Got it.
2:25>> So, who else? Who else? We had voices
2:27that were breaking in periodically, but
2:29I didn't see whose avatars were
2:30lighting up, so I can't call your
2:32attention. Come back to us, whoever was
2:34ready to talk.
2:36>> Can I ask a question?
2:38>> Of course, of course.
2:39>> Oh, my name is Kostya. There. And I
2:45recently had this question arise,
2:47honestly, such a, I don't know, request
2:49for, uh, where exactly is that line and
2:52the difference between a regular
2:54product manager and an AI product
2:56manager. I would really like to see or
3:05hear a clear answer to this question in
3:07one of the courses we'll be taking,
3:09whether it's mandatory or an elective.
3:15Yeah. And, well, because it seems like,
3:21um, some projects have already started,
3:24and you get onto projects, right, with
3:26teams, you sit there, talk to
3:28stakeholders, and it feels like the
3:30basic process, the one you go through,
3:32and the one we went through during the
3:34hackathon, is, roughly speaking, the
3:37same for both an AI product manager and
3:39a regular one. Yeah. The only
3:49difference is just the extensive use of
3:51some more specialized tools, for
3:53instance, like AI-powered personas.
3:59That's just an example. So far, that's
4:01the only distinction I've managed to
4:03identify. Yeah. And if it's not too
4:06difficult, or if it's already planned
4:07as part of the course, that would be
4:09great. Where is that line?
4:12>> Where does, where does,
4:14>> where does your desire to figure this
4:16out come from, Kostya? Why does this
4:18question arise?
4:22>> Because, well, since it’s been
4:23separated in this way, well, there must
4:26be some special focus on it. I don't
4:32know, this prefix really draws a lot of
4:34attention.
4:35>> I see, I understand. So, let me try to
4:37rephrase it, and you can correct me. Do
4:39I understand you correctly or not? You
4:41mean, you see this "AI product manager"
4:43title everywhere, and you expect it to
4:46be something cooler, right? Meaning
4:48that it's not just marketing, and that
4:49it actually has some extra value. And
4:52that by figuring this out, you won't
4:54just be a product manager, but
4:56something with a prefix. Well, not
5:00necessarily cooler, it's just that
5:02there are software engineers, there are
5:04different types of engineers. These are
5:08all people who specialize in their own
5:10areas. And I'd like to find out what
5:13exactly is the unique "zest" of this
5:15field of working with AI.
5:17>> Well, yeah, let me try to give you a
5:19short answer just so I don't take up
5:20too much time right now. I want to chat
5:22a bit more, kick some ideas around, and
5:24then move on to the course topic. I'll
5:25say right away that throughout the
5:27course, there isn't really a planned
5:29answer to this; that's not the goal,
5:31it’s not on the agenda. But we will
5:34have plenty of time, throughout the
5:36course we'll have quite a lot of time
5:38to explore this, talk about it, and
5:40philosophically sort it out. I would
5:44say it differs in roughly the same way
5:46an engineer differs from, say, a
5:48regular software engineer or an
5:50engineer who handles configuration,
5:52right? So it’s just a different area,
5:57a different zone of work, yes. It’s
6:01clear that a product manager’s
6:02toolkit is high-level, I mean, compared
6:05to engineers, a product manager has a
6:07fairly high-level toolkit, yes. It’s
6:09clear that an engineer working on a
6:11stack is basically working with the
6:13same math in the end, but they have the
6:15foundation of that math, and then on
6:17top of that, a whole structure of other
6:19, higher-level disciplines. In that
6:22sense, it’s less noticeable with a
6:24product, but fundamentally, yes. So you
6:26are generally right that, well, the
6:28processes are more or less the same
6:29everywhere. We are always researching
6:31something, validating something, and
6:33pulling out more or less the same tools
6:35everywhere. Fundamentally, yes. Yes,
6:37but then nuances appear in the details,
6:40nuances in solving specific tasks.
6:42Simply because product managers work on
6:44different products, there is a small
6:46domain-specific aspect. For instance, a
6:50product manager working with hardware,
6:52they release hardware products, and
6:54they use generally the same tools as a
6:57product manager releasing software. But
7:01, despite the fact that the approaches
7:03and tools are all the same, there are
7:05just a ton of nuances in the work.
7:07Ultimately, they lead these two product
7:09managers into completely different
7:11universes. Yes, my request for the
7:16course is to catch that nuance, and how
7:19will it be handled—will we be told
7:21about it, or are we supposed to
7:22synthesize it ourselves? So I need to
7:27figure out on my own where the nuance
7:28is?
7:29>> No, like I said, we can hash that out.
7:31The task of answering that question
7:33isn't set for the course, nor is it the
7:34course's objective. I'm stating that
7:36right away. And I don't have the
7:39expectation that all course
7:40participants will necessarily—well, a
7:42spoiler here—we won't necessarily be
7:45developing a product within the scope
7:47of the course. Everyone will develop
7:50their own product either with a team or
7:52solo, whichever you prefer. I don’t
7:54have an expectation or requirement that
7:56the product must be email-based, for
7:58example, right? That’s not necessary
8:00at all. We are talking here
8:03specifically about the tools you need
8:05to use in order to launch a product in
8:07general. That’s point number one. My
8:10first thought. Second thought: if you
8:13specifically choose the ML domain and
8:15you definitely want—well, maybe
8:17let’s make it broader, let's call it
8:20an R&D project, or an AI project,
8:22whichever you like—you’ll really
8:24get a feel for that specific area. If
8:28you have experience in, say, standard
8:29product management—you didn't tell us
8:31anything about yourself, so I don’t
8:33know if you have such experience—you
8:35can fix that oversight.
8:38>> Well, okay, if I could take a little
8:40more time, I haven't worked as a
8:42product manager or a developer in the
8:44classical sense. I just finished my
8:47bachelor's degree. Yeah. Uh-huh.
8:50>> I spent more time on study projects
8:52there. It’s not like I’ve found a
8:55job and applied myself somewhere yet.
8:58Yeah. And within those study projects,
9:02there was someone who inspired me to be
9:04a PM more than a developer. He put it
9:09all together and did it so
9:10filigree-like. He always had thoughts
9:14about the project, the product. He
9:18could handle so much information in his
9:20head at once, and he operated with it
9:22really skillfully. He inspired me to
9:29come here and to lead some projects at
9:31the university as a product manager
9:32rather than a developer. Of course,
9:38some things worked out, some didn't,
9:40but it still seemed to me that this is
9:42probably my thing. Yeah. And that’s,
9:45in fact, why I am here. I mean, you
9:48can’t say I have a lot of experience
9:50—I have very little—but I still
9:53think, well, in short, yes, I think it
9:55will be very useful for me. Understood.
9:59Okay, let me elaborate for another 5
10:01minutes,
10:01>> and that’s all.
10:03>> Uh-huh. Alright. I propose we turn on
10:07our cameras. That will make things a
10:09bit livelier. We aren't doing too badly
10:11, but let’s turn the cameras on. Wait
10:15, I turned my camera on, but I can't
10:17see myself. What a wonderful, wonderful
10:20place for calls, of course. Oh,
10:22everything, I see myself. Wait, this is
10:24the wrong camera. Oh well, you'll be
10:27looking at me from below. Everything
10:31started moving. Oh, I even recognize
10:34some faces, I think I've seen some of
10:37you before. Right, let me introduce
10:42myself briefly, then I'll pass the mic
10:44back to you so you can continue telling
10:46me about yourselves.
10:49>> So, I'll share my screen. Sharing my
10:51screen. Sharing my screen. Screen. And
10:54you can see the screen. And the slide
10:57should have started, too. And the slide
10:59has started. Well, as for us all being
11:00here for product research, I think
11:02everyone already knows that. About me.
11:03Sorry about the pretentious photo. I
11:05don't have any normal photos. This is
11:07the latest one I could dig up. The
11:08freshest one. So, I'm Vadim, and I'll
11:11be teaching your course on, well,
11:13product research. This is the fourth
11:16time I'm teaching this course. So, you
11:18are my fourth cohort. You can
11:19congratulate me on that. You've got the
11:22most refined version of this course
11:23that exists to date. That's one thing.
11:26Two. Well, right now, for the last
11:28couple of years, I've been a product
11:30manager in MedTech, specifically at a
11:32company called Onescell, where I
11:34develop a platform for pathologists. If
11:36anyone is interested, I'll tell you
11:38more later. Some of the mentors we'll
11:40have on the course actually work with
11:42me at the company as well. And so
11:44they'll be able to give you more advice
11:46on how to develop products in MedTech
11:48specifically. So, if anyone has been
11:50wanting to try doing something in
11:52MedTech for a long time, now is the
11:54time; you probably won't find a better
11:56time than this. Before that, I was a
11:58product lead at a company called Pixel,
12:00where I launched startups. So, while
12:03I'm focused on developing an existing
12:05product here, although we are slowly
12:07launching new ones too, before that, I
12:09was launching products almost like on
12:11an assembly line. So, my main
12:13specialization is precisely all those
12:15studies necessary to launch a product,
12:18whatever that product may be; I've had
12:20experience in completely different
12:22domains. And I also had experience in a
12:24venture studio, of course. In this
12:26sense, it gives me a bit more insight,
12:28so we'll be able to analyze many more
12:30diverse cases together. And that's it,
12:33briefly about me. I don't really like
12:35it when my photo is just hanging there
12:37on the screen, so I'll go back to the
12:38course title and hand the microphone
12:40back to you. Anyone, anyone, who's
12:43there? Let me go through those who have
12:45their video turned on. Very, very sad.
12:48Alexander Esin is sitting there.
12:49Alexander Esin, you look very sad. Tell
12:52us about yourself, what are you doing
12:54here?
12:54>> Ah, good afternoon. Uh, yeah, I, well,
12:57no, I'm not sad, just a little tired.
12:59But otherwise, everything is good. Uh,
13:01I graduated from the MSU Faculty of
13:04Computational Mathematics and
13:06Cybernetics. I worked at Alfa, and now
13:10I work at the largest cloud provider as
13:12a technical expert and architect. Well,
13:16accordingly, there are a lot of
13:18projects, mainly in ML, given the
13:21specific nature of it. And, accordingly
13:25, it became interesting how to work
13:31with this not only from the perspective
13:33of infrastructure stuff, but more
13:34broadly, to look at it through the lens
13:36of business. That's why I actually
13:40enrolled here, and I'm happy about it.
13:43I'm curious how the course will be
13:45structured. I have high expectations
13:47too. Also
13:48>> voice them. Voice your expectations.
13:51I'm impatient. Well, actually, actually
13:54, I would like to understand a bit more
13:57about market sizing, about, well,
13:59specifically from the artificial
14:01intelligence standpoint—what my
14:03colleagues were asking you about—and
14:05how that translates into this paradigm?
14:11And what these products are for, so we
14:13don't just make another AI B2B product
14:15to brag about on LinkedIn, but so that
14:17there is some actual benefit from it.
14:22And so that it's truly a
14:25>> fundamental product with a
14:27well-developed problem, a pain point,
14:29and an understanding of how to further
14:31develop, maintain, and support it,
14:33because, by and large, our whole life
14:36is, one way or another, a product. Yeah
14:38. And this will be useful everyday
14:41knowledge as well. Well, I think those
14:43are my expectations. Alexander, I will
14:46try to definitely live up to your, your
14:48desires. I think, in my view, that's
14:50exactly why we are here and what we
14:52will be talking about here. Thank you.
14:55Next, I, yeah, I see Vladislav.
14:58Vladislav is against a background of
15:00something. MSU, I think, Vladislav.
15:03Vladislav Semenov.
15:05>> No, it's a generated image, actually.
15:08Where I'm heading.
15:09>> And well,
15:11>> well yeah, actually, I worked as an
15:13analyst at a bank and at Kuper. And I
15:16super love everything revolving around
15:18tests. And it's a more, you know,
15:20proven way to test a product. But it's
15:24also interesting, everything on the
15:26side, everything that we can't
15:27scientifically prove but can research.
15:31And so I went into product research
15:32simply to cover that piece of knowledge
15:34I'm missing. Uh-huh.
15:36>> What's that? Okay, got it. Well, let me
15:40tell you right away, Vladislav, I’m
15:42inviting you to a debate on the topic
15:44that A/B testing is the most proven
15:46method; we have many different
15:47scenarios here. We can run through them
15:49, just not right now. Not right now.
15:51Later.
15:52>> Later, later, maybe some other time.
15:57>> Yes. Hi everyone. I think I might have
15:59missed the question. I just need to
16:01tell a bit about myself, right? Yes, my
16:02question is, just tell us a little bit
16:04about yourself, just so we can
16:06understand. Understand the context.
16:07Then tell us what your expectations for
16:09the course are.
16:10>> Uh-huh. So, a little about myself. In
16:122023, I finished my bachelor's degree
16:14at HSE in organizational management.
16:17After that, I probably worked for
16:19another 3 years in consulting. I was an
16:22analyst there and dealt specifically
16:24with HR, HR-tech, and HR-analytical
16:26products. Actually, why did I become
16:30interested in the product direction in
16:32general? Because, on one hand, I
16:38realize that I don't have enough
16:40understanding of managing B2B products.
16:45I see a lot of content on how to manage
16:47B2C products. But product management,
16:53for example, for HR, seems to have a
16:55specificity that requires new tools.
17:00Yeah. And I don't really have a clear
17:01idea of how to use things like A/B
17:03tests there, for example. Yeah. And
17:05here I want to understand, because from
17:07my experience I see that B2B products
17:09and B2C products are, I would say, very
17:12different products, it seems.
17:14>> 100%.
17:15>> Yeah. And, well, I currently have,
17:18perhaps, two global goals. First, I am
17:24that B2B person, building something in
17:26the HR market, which I already
17:28understand quite well.
17:32>> Are you a solo founder? Or not?
17:36>> Well, our project actually grew out of
17:38a hackathon, so we have a team.
17:41>> Hackathon team. Okay.
17:43>> Yeah. And the second, well yes, the
17:47first point is to apply this knowledge
17:49to managing this product for HR. And
17:56the second is to potentially package my
17:58experience for what the market demands.
18:03Sorry, and in a format so that, well,
18:05hypothetically, one could also go to a
18:07company and work as a product manager,
18:11>> if our company doesn't take off. There.
18:14>> Okay. Understood, understood. Let's do
18:17this. I want to start presenting in 10
18:19minutes. We will have a dialogue with
18:21you anyway, so you can periodically
18:23throw in some more things, but let's
18:25take another 2 minutes right now. Is
18:27there anyone I won't call on who is
18:29ready and wants to supplement the
18:31previous speeches and also share their
18:33expectations for the course?
18:36>> 2 minutes, anyone.
18:37>> Can I speak on behalf of our,
18:39>> of course,
18:39>> from SUSU,
18:40>> go ahead, yes. Ah,
18:42>> I'll answer. Well, we're a bit shy. I
18:46might lag because our internet here
18:47also works poorly sometimes. There.
18:51I’m sending you greetings from Evgeny
18:53Melnikov. My expectation from the
18:55course is to learn once again,
18:59>> yes, to conduct research systematically
19:02once again, because there is logic
19:04after a bunch of, uh, technical
19:06accelerators, we also went through
19:08ITMO’s great ones and the Academy of
19:10Innovators. So I just want to try to
19:14systematize everything once more on the
19:17updated product. That’s all. Well,
19:20that’s the goal for the course. Okay.
19:24Okay. Alright, that's it, let's slowly
19:27start moving on to what you’ll be
19:29talking about. So, I ask you not to
19:30turn off your cameras, well, or you can
19:32turn them off, however you prefer. We
19:34are all free people here in general.
19:35The idea is roughly like this: for some
19:37part of the lectures, in principle, yes
19:39, I will just be explaining and talking
19:41about things. Today, everything will be
19:43super easy, a super basic topic, very
19:45simple. And further along the way, all
19:48of this will become more complicated.
19:51It will become more complicated,
19:53including the fact that you will need
19:54to apply it to your own product. As I
19:56already said, our task within the
19:58framework of this course is, in a
20:01minimal form, to go from an idea,
20:03formatted in any way, to a clearly
20:05validated, reasonably validated,
20:07confirmed or refuted, preferably not
20:09even just one hypothesis, right? So our
20:12task is the minimum. The minimum, and
20:16in a mid-level scenario, you have
20:18something appear in the form of an MVP,
20:20also confirmed and verified, with which
20:22you can reasonably try to enter the
20:24market. And at the maximum, of course,
20:27you end up with something grand.
20:29Considering that nowadays, to code this
20:32or that solution, in general, you need
20:34a couple of weeks in the worst case,
20:36right? So, in general, this is quite
20:38realistic. That's one thing. Two. Since
20:41our audience here is quite diverse—
20:43some of you are purely technical, some
20:46are already product managers, some are
20:48tech-focused but close to product roles
20:50, and others are tech experts very far
20:52from product work—people are at
20:54varying degrees of proximity to product
20:56management, so we will cover everything
20:58starting from the absolute basics. I
21:01mean, at times I will say things that
21:03seem obvious to some of you, or truly
21:05are obvious to you, but I would like to
21:08believe there will be things that are
21:10not so obvious. That is something
21:13important to note. Another part. So,
21:15every lecture, or almost every lecture,
21:18will end with some sort of, let's say,
21:20homework assignment. This homework is
21:23essentially a part of an iteration, or
21:25a step on the path toward a
21:26well-developed, verified, and confirmed
21:29hypothesis. As I said, we are here
21:31primarily to develop a product idea,
21:33that is, a product meant for the open
21:35market. So, if anyone had an idea to
21:38build something inside a company or for
21:40their own company, you can do that too,
21:43but the expectations and requirements
21:45for you in this course will be those
21:47for people developing a product idea
21:49for the open market. And in general,
21:51this is just to protect against someone
21:54saying, "Sorry, I have a secret idea, I
21:56can't share anything." That, well, that
21:58won't fly. I want to say, this is also
22:01important to note. Now let's talk a bit
22:04about how the course will be structured
22:06. We will have eight so-called modules.
22:09In reality, these are eight lectures
22:11where we will cover different topics.
22:12And 11 so-called practice sessions.
22:14What is a practice session? A practice
22:17session is essentially a consultation
22:19where you come to mentors as part of
22:20working on your product. We have about
22:23five or six mentors. I won't guess
22:25right now, I don't remember. The
22:27mentors will tell you about themselves
22:28in a few days. So, it will look exactly
22:31like this. Lectures will be held
22:33according to our schedule. Like one of
22:34them is happening right now. And seven
22:36more lectures will follow later, also
22:38according to the schedule. We are going
22:40to have a slightly more brisk start.
22:42Meaning, for the first 3 weeks lectures
22:44will be every week, and then they will
22:46be every 2 weeks so that you have time
22:48to actually complete the tasks, that is
22:50, do your homework. And also, once
22:52every two weeks after these three
22:53initial weeks, we will have
22:55consultations with you. And for the
22:57consultations, accordingly, you sign up
22:58in advance with the mentors you are
23:00interested in. They will talk about
23:01themselves in advance. You come to them
23:03and say, "So, guys, this is what we are
23:05doing.""We’ve run into these kinds of
23:06problems.""What do you think?" The guys
23:08will groom your options, right? That is
23:10, they will challenge your solution,
23:12and they will help you find solutions
23:14to the problems you are facing. You can
23:16come to the consultations not only with
23:18questions about a specific current
23:20assignment, right? You have to
23:22understand that everyone who will come
23:24to the lectures, everyone who will come
23:26to the consultations, everyone who will
23:28act as a mentor—these are all people
23:30who work or have worked in startups.
23:32They launched them, they developed them
23:33, they changed them quite significantly
23:35. Therefore, they understand this
23:38specificity of working with products at
23:40the very early stages very well, and
23:42they will help you. Yes, so the guys
23:44are solving quite specific tasks. Yes,
23:46such tasks where you have to manage to
23:49reach someone, including in B2B, and
23:51including in Gov B2B. In general, there
23:53are many different interesting cases
23:55you can explore with them. Therefore,
23:57you can come to them with anything.
23:58They will consider any of your cases.
24:00You will post your cases and questions
24:02in advance; if a mentor realizes that
24:03they don't have much to say on that
24:05question. But there is another mentor
24:06who can answer it. We will simply
24:08transfer you ourselves. We’ll handle
24:09it ourselves. There. Moreover, our task
24:12, our task as mentors, and mine as a
24:13lecturer, is to explain all this to you
24:16so that it is clear and so that you
24:18enjoy it. So, our task, mine and the
24:20other mentors 'at the consultations, is
24:22to make sure that you get your product
24:24to the result as quickly as possible,
24:26that you get your product idea to the
24:28result. We will facilitate this in
24:30every way possible. There. Uh, so, are
24:33there any questions? Maybe not. Well,
24:36in general, if anyone wants to learn,
24:37oh, right, an important nuance. An
24:39important nuance. Some of you mentioned
24:41wanting to switch to a product role
24:42within your current company, or better
24:44yet, switch to a product role in Big
24:46Tech and develop products there. Or
24:47simply develop a product within some
24:49company. So, what is worth clarifying?
24:52Some of the things we are going to
24:53discuss here are not very applicable to
24:55Big Tech. Simply because the specifics
24:57are different: launching a product from
24:59scratch into an open market is not the
25:01same as launching a product within a
25:03company that already has a huge user
25:05base where you can pour in traffic.
25:08It’s not the same at all; completely
25:10different types of actions are needed,
25:12at least for the launch. But if we are
25:14talking about you working on a feature,
25:16and really, every feature launched
25:18within a product should provide value,
25:20then the foundation—the things we
25:21will be covering fundamentally—should
25:23be especially interesting to you. Yes,
25:25that will be truly applicable. That’s
25:28it. This is, let’s say, an overview
25:30part. As I already said, some things
25:33will be obvious. I am going to ask some
25:35obvious questions now. Get ready.
25:37Anyone who wants to slap their forehead
25:39and say "that's so simple," go ahead. I
25:41urge you not to do that, or at least do
25:43it where I don't see it while I'm
25:45presenting. So, I encourage everyone
25:46else to participate in the discussion.
25:48Part one, yes, we will have three parts
25:51here. Within this lecture, there will
25:52be three of them. I might have to speed
25:54up in some places, but we will start
25:56with everything; this is still about
25:58the entire course. This is a scheme
26:01proposed for us to analyze as a
26:03traditional approach. It is clear that,
26:05in reality, everyone launches their
26:07startups and companies differently;
26:09there are various cases. But let's just
26:11look at this as something I personally
26:13call the traditional approach. This is
26:16some kind of non-product approach,
26:18without research, where people act as
26:20if they were living without us product
26:22managers or not using our product
26:23frameworks. So, what does this usually
26:26look like? And many products are
26:27actually launched this way, some even
26:29launched successfully. It's not that
26:30it's a non-working approach, it's just
26:32not repeatable. So, an idea appeared;
26:34we will have some kind of idea. So,
26:36here I am, a solofounder, I’m sitting
26:37with friends, we’re brainstorming,
26:39coming up with some ideas. And then, an
26:41idea strikes me. We sketch out a quick
26:43business plan, realizing that what
26:45we’ve scribbled in our notebooks
26:47makes us absolute billionaires. In just
26:49a couple or three months, we’ll code
26:51something up, it’ll take off, we’ll
26:52be unstoppable, we just need the chance
26:54, and we’ve finally figured out what
26:56we’ve been waiting for all this time.
26:57We think: "Okay, we need a couple of
26:59people, right?" I mean, we need to grow
27:01the team, we shouldn’t do everything
27:02ourselves, plus we’re the founders,
27:03we’re the ones who just came up with
27:04it all. We hire a team, and we all
27:06start working together. We’re doing
27:08it, writing code, building it out, it
27:09doesn't really matter how, right?
27:11It’s just a matter of time. So we
27:13start, we’ve done the work, and we
27:15begin to sell. Guess what’s on the
27:18next slide.
27:20>> Failure. No money, nobody needs it.
27:24>> They didn’t buy it.
27:25>> Just didn't buy it. Yeah.
27:26>> Yeah. They aren't buying, right. They
27:28just aren’t buying. So what gives?
27:30Well, I mean, they aren't buying.
27:31I’ll say it again, it doesn't have to
27:33be that way, right? It could turn out
27:34that they do buy. History knows famous
27:36cases, documented ones, where it works
27:39out, where guys really came up with
27:40something and it took off. Such cases
27:43exist, that’s true. But if we look at
27:47the statistics as a whole for all
27:48launched products, that urban legend
27:50circulating since the 2000s that 99.9%
27:52of products and startups fail in the
27:54early stages—it’s about those
27:56products that were launched exactly
27:58like this, right? Although, the
28:00statistics have changed since then.
28:02Actually, launching a product is not
28:04rocket science, right? Different
28:06products are launched in different ways
28:08with varying complexity, naturally
28:09working differently in different
28:11markets, but launching a product and
28:12building it so that it's useful is more
28:14than realistic, and there are
28:15approaches for that, but this just
28:17isn't the right one—so what's wrong
28:19with it?
28:22>> Well,
28:23>> because a very long time passes from
28:25the moment the idea appears to the
28:27actual validation and verification of
28:29its value. You mean, you wrote business
28:32plans, assembled a team, developed it,
28:34and only then, potentially, did you
28:36assess whether it was needed or if
28:37people were even willing to pay for it.
28:40>> Uh-huh. Let me suggest this. I see that
28:42several people are turning on their
28:43microphones at once, then one starts
28:45talking and the other gets blocked.
28:46Let's use the raise hand feature. Yes,
28:48hands are raised, so let's just go in
28:50order. I'll figure out how to see the
28:52queue now. I see the queue. So, Artemy
28:54Verbitsky is first in line. Well,
28:56second now, right? The first one has
28:57already spoken. Second. Artemy. The
29:02problem is that the hypothesis wasn't
29:04tested on the product's target audience
29:06.
29:08>> Well, can we say that, considering that
29:10someone is our hero—let it be me, let
29:12it be me, I'm from another universe.
29:15And Vadim from another universe went
29:16down this path and even tried to sell
29:18something, but it didn't work out.
29:20Moreover, look, I wrote it with a slash
29:22. Actually, maybe someone even bought
29:24something from me, they just don't use
29:25it.
29:26>> Uh-huh. Can it be said that I didn't
29:28test it on the target audience?
29:30>> Well, according to this scheme, yes.
29:32And what was in your head when you were
29:33doing it? Well, I can't. That's why
29:35I'll suggest that's understood.
29:37Accepted. Okay.
29:42>> All right, Konstantin.
29:44>> Well, I don't know about whether
29:46someone tested it or not. That's
29:48probably actually described in the
29:50business plan. It's not really shown in
29:52detail here, is it? It seems to me that
29:57, just by the look of it, this approach
29:59is like a standard waterfall, where the
30:01guys just took it, made it, and went to
30:03some kind of production. Although,
30:09perhaps the problem itself, maybe the
30:15field in which the product that was
30:17developed is applied, the situation,
30:19the time—maybe it just requires a
30:21different approach to development,
30:23where there could be several iterations
30:26between development and the business
30:28plan, for example, or, well, it just
30:30requires something else, not a typical
30:32waterfall. Maybe that's why the concept
30:37and the approach itself are just too
30:39linear. There.
30:41>> Yes. Let me, let me, summarizing all
30:44that was said previously, repeat the
30:46case one more time. So, I'm from
30:48another universe, I was sitting in a
30:50cafe with friends, came up with a
30:52product idea, wrote a business plan on
30:54a piece of paper. So, we're making, I
30:57don't know, a B2B app for pug owners.
31:01We'll help them choose the time for
31:03walking. We made up some nonsense and
31:05thought, "Damn, there are actually a
31:07ton of pugs.""That's it, this is going
31:09to take off." That's the topic. That's
31:11it, we've thought of everything,
31:12sketched out a business plan, and
31:13calculated: "Man, how much? Well, pugs
31:14cost a ton, people have money. So,
31:16anyway, let's charge 1,000 rubles for
31:18it. In short, dog walking, we
31:20calculated it, and some kind of unit
31:22economics came out. You know,
31:23calculating it literally on a piece of
31:25paper, or maybe not even on paper, but
31:27on some huge, very complex calculator
31:28or a massive, complicated spreadsheet.
31:30I calculated it all. Everything looks
31:32beautiful, I just pulled it out of my
31:34head, calculated it, put a team
31:35together, we coded it, did everything,
31:37released it, started showing it to
31:39everyone, and nobody's buying it, and
31:40those who do buy it, they bought it by
31:42accident, they don't use it, they don't
31:44need it. Let's continue the discussion.
31:47Gennady,
31:49>> let me offer a bit of criticism. I'll
31:52just express a counterpoint. What's the
31:55idea? If we assume that you want to
31:58lead us to the conclusion that, well,
32:00we still need to do at least some small
32:02product research, right,
32:04>> well, from personal experience,
32:09>> it seems that AI handles doing even
32:11remotely decent product research poorly
32:13, but it's very good and fast at
32:17>> churning out a prototype. So, isn't it
32:22true then that maybe this scheme
32:24actually didn't work well before, but
32:26now it's much more effective to make
32:27way more attempts? No, I mean not like
32:35launching the same product many times,
32:37but like spending one day to build
32:39something halfway decent instead of
32:40conducting product research. And then,
32:45it seems like this scheme actually
32:47becomes viable. But there's still
32:51something wrong with this scheme, yet
32:53your comment is accepted. I'll put it
32:55on the shelf, and I'll take it off two
32:56slides from now. I'm not saying you're
33:02wrong, there's just a nuance. I'll
33:04bring it back. Radomir,
33:05>> yes, yes, yes, yes. Even if I'm wrong,
33:08it's not a problem for me.
33:09>> No, no, I mean, it's fine, I just mean
33:11it's a good comment, I'm just setting
33:13it aside. I just need to remember to
33:15take it back out. If I forget after two
33:17slides, please remind me, two slides,
33:19including this one.
33:22>> Yes, our main problem is that we end up
33:24with our own vision of the solution and
33:26our own vision of the problem. Meaning,
33:28the problem might not exist, the
33:30solution might not solve it, yet we've
33:33poured resources, time, money, etc.,
33:35into it. And, actually, just now
33:38>> Yes, yes, keep going like that.
33:40Essentially, it's about customer
33:41interviews. It is research, after all.
33:44You quickly make a prototype, and then
33:45you research with it. That's exactly
33:47what it is.
33:49>> Put it back on the shelf. Why did you
33:51take it out? Put it back in its place.
33:52>> Alright. I'm putting it by the pie.
33:56>> Thanks.
33:58>> Well, that's valid. I agree. Sounds
34:01true enough. Daria.
34:02>> Yes. Well, the guys have already
34:04touched on this a bit. Hello everyone.
34:06Well, ideally, a business plan should
34:08simply include a proper analysis, both
34:10internal and external, of any existing
34:12constraints, including, for example,
34:14economic or legal ones. But we are
34:16assuming that we sat down with friends,
34:18thought it was a cool idea, crunched
34:20some numbers, and started doing it. Uh,
34:22it seems that if this were the right
34:24approach to a business plan, then,
34:26>> uh, maybe through analysis, some things
34:28could have already been identified. And
34:31yes, of course, product research—to
34:33check, to look at whether anyone else
34:35is already doing this, how they're
34:37doing it, how they're faring, and why.
34:40It seems that at this early stage, you
34:42need to understand and, if possible,
34:44yeah, conduct some kind of
34:45mini-research, even just using a
34:47clickable mockup to see if there's any
34:49demand at all. Mm-hmm.
34:52>> And it seems that here, somewhere at
34:54the beginning, we might also have a
34:55part,
34:57>> right? Let me just—I mean, I don't
35:01have any complaints about what you've
35:03said. I just want to throw the question
35:06back to you: what is wrong with this
35:08scheme? You say, if a proper business
35:10plan was made here, if this, if that. I
35:12already set the framework. I told you
35:14that the business plan appeared when I
35:16was sitting—as Yuri Tut says—with
35:18me and my buddies. So we sat down
35:20together and just threw together a
35:22business plan. Well, we threw together
35:23a complex business plan. I'm a product
35:25manager, I've been working at it for a
35:26long time, I can throw together complex
35:27ones. I'm like this complex, yeah,
35:29smart, very convincing person. And my
35:31product manager buddies along with me
35:32also threw it together in a complex way
35:34. Our entire analysis is just, uh, our
35:36experience. We stated all of this
35:38expertly. There was no research.
35:41>> Well, perhaps we are indeed missing a
35:43step to validate the idea on people
35:45outside of this organized group.
35:49>> Fair enough, Konstantin. Well, overall,
35:54if we aggregate everything said above,
35:56it seems to me that this step is
35:58somewhere between development and sales
36:01, because, well, in my experience, any
36:03>> this one, that which
36:05>> this what, which which step,
36:07>> it says, this step.
36:09>> Step. Well, a step. Well, this, of
36:11course, sounds a bit cliché by now.
36:14Well, something like a test run, right?
36:17>> Or or something like that,
36:19>> like somewhere before attempting sales.
36:21Yes, yes, because from the experience
36:23of any product one would want to launch
36:25, even if something was already
36:27developed, and you realize that, well,
36:29this could already be sold to a person
36:31so they could start using it. In my
36:34experience, I have never once jumped
36:37straight into sales. And even if
36:39someone actually bought it, like it
36:41would become a full-fledged product. So
36:43, somewhere somewhere between
36:45development and sales, there was
36:46something else. But what specifically,
36:48to be honest, I don't remember anymore.
36:50There. can be officially named. There
36:54>> we'll figure this out. Get to it.
36:56Thanks. Guys, why, listen, I look at
36:58the list of people, I always see three
37:00hands there. How does this happen?
37:01We're filming from above, aren't we?
37:03Where do the new ones come from, Ivan?
37:05>> Well, if we are talking specifically
37:07about the diagram, then I think we
37:09could add, since we had, yes, sales, we
37:11have an arrow going straight back to
37:13the beginning, like an idea appeared, a
37:15business plan. Well, like, what if
37:17suddenly they don't use it, because
37:19there was, well, some kind of flaw. And
37:21actually, one just needs to do an
37:23analysis there and refine something. So
37:25, I wouldn't make such a big connection
37:27, like returning, well, I mean, I would
37:29make a smaller cycle of some sort. Well
37:31, that is, for example, sales, analysis
37:34,
37:34>> sales, then, like analysis, let's say,
37:36some refinement. So there should be a
37:38cycle here. And in this case, if it
37:42doesn't work out here, then you iterate
37:44back. You didn't get any sales, forget
37:46what I wrote here about not using it
37:48yet. You just didn't get any sales. And
37:50what are you going to analyze? You made
37:52it, went to market, launched people,
37:54well, I don't know, you have a B2B
37:56product, then for you it will most
37:58likely not be ads, it will be calls.
38:00I'm star I realized, it's that I talk
38:02too much, that's why hands keep
38:04appearing. Need to be need to be more
38:06modest. Anyway, uh, you have a B2B
38:07product, yes, we were discussing that
38:08someone wants to build a B2B product,
38:10you have a B2B product, you solve
38:11everything by reaching out to the
38:12client. You have a market there that
38:14isn't very large. You go to the client,
38:16you try to sell, it doesn't sell.
38:17That's it, they tell you:" We don't
38:19need it, thanks, thanks, no need. "
38:20That's it, that's what they tell you.
38:22This is what you've found out. You did
38:23this 100 times. 100 times they told you
38:25they don't need it. So you're sitting
38:27there with a finished product and a
38:29team saying:" Well, what's up? "And
38:31with 100 attempts that ended in nothing
38:32. What is there to analyze?
38:36>> Or
38:37>> should sales be placed right after the
38:39business plan? Well, if you've already
38:44watched the course, why are you even
38:46coming? Just joking, joking, joking.
38:50We'll really have people joining like
38:52that. Anna one, please. Anna. Anna, you
38:59have no microphone, but your hand is
39:01raised. You should either lower your
39:04hand or turn on your microphone.
39:06>> It's in the chat.
39:07>> Oh, damn, I have to open the chat too.
39:12So, Anna, what problem of the pugs is
39:14being solved? Who said there is a
39:16problem? What is this all for, in
39:18general? Not from the beginning of the
39:20process. Who said that pugs exist? Like
39:23. I'm giving it a like. But this isn't
39:25a product for pugs. But not for pugs
39:27either. Well, it's obvious the product
39:28isn't for pugs, it's for their owners,
39:30because I'm not a total fantasist. I'm
39:32not crazy, I'm a product manager, I'm,
39:35well,
39:35>> well, it's just unclear what problem of
39:37the pugs is being solved, where it came
39:39from, and if it even exists—I mean,
39:41it's not even here at all.
39:42>> And it's not here either. Well, in
39:44reality, there's no business plan here
39:45as such. There's only a statement that
39:47I wrote one. We don't even know, maybe
39:48the business plan there was bad. But
39:50I've already realized that this diagram
39:51needs to be removed. You are only
39:53criticizing it. Well, you need to use
39:54your imagination. Just imagine, believe
39:56me that everything here was done
39:58correctly. Well, fair enough, fair
39:59enough. And I'll put your questions on
40:01hold. We will return to this. Lower
40:04your hand, Anna, please. I need to call
40:06the next person. I suspect that you'll
40:08lower your hand, and another one, a
40:10third, will appear. Nikita,
40:12>> yes, hi, Vadim. Uh, actually, my answer
40:15has already been voiced. Well,
40:18>> no big deal.
40:19>> After the business plan, we'll put
40:21sales, make some kind of MVP, see if
40:23they buy it or not. Well, we'll try to
40:25promote a product that doesn't exist
40:27yet. Like, I don't know, launch some
40:30presales accordingly,
40:32>> see, we'll get some numbers, and we can
40:34base our next steps on that.
40:36>> But this really does look like that
40:37case we were just discussing earlier,
40:39that we can now white-code everything,
40:40generally speaking, quite quickly,
40:41right? So, in general, making a landing
40:43page and trying to sell isn't much
40:45cheaper than, like, making a product
40:46and trying to sell that, right? I mean,
40:48these days it’s not much cheaper. So,
40:50the main expenses will likely be on
40:52sales rather than on creating the
40:53product or the landing page. We’ll
40:55spend about the same amount of time on
40:56both, especially our own. Just a
40:58different number of tokens; there will
41:00be a subscription for 20 bucks, or for
41:01200. But in general,
41:06>> Konstantin,
41:07>> yeah, in continuation of the topic of
41:10sales after a business plan. Back in
41:18February at the camp, we had a
41:20masterclass on, well, like, how to
41:22create a real value proposition that
41:24can actually hook the, well, the target
41:26audience. Yeah. And it's not even about
41:37the product that doesn't exist, but
41:38rather about finding that buyer who
41:40would potentially buy, and just
41:42establishing a dialogue with them so
41:44they know about you. And, probably, at
41:49this stage, the idea might be discarded
41:51, and the business plan might be
41:53adjusted, because, well, maybe people
41:55who aren't stingy with communication
41:57will tell you, like," Well, I have a
41:59pug, but honestly, I don't even take
42:02him out of the house, he uses the
42:04bathroom at home or something like that
42:06, "right. Yeah. And in that case, the
42:10idea will change there.
42:12>> Yeah. But, yeah, it's probably true
42:15that you can really launch something
42:17very quickly with no-code and use that
42:20as a demo, right? A demo plus this
42:23value proposition equals, like, the guy
42:26is hooked, both with his eyes and his
42:28interest. Yeah. But, well, in general,
42:32yes, likely, but it's more of a first
42:34contact, just a conversation, rather
42:37than a sale.
42:37>> Well, a sale in the broad sense, yes.
42:40We sort of sold the idea, right?
42:42>> Right? So, from this stage, there won't
42:45be any users at all. They won't be able
42:48to log in. And from this stage, there
42:50probably won't be, well, any money or
42:52anything. It's just a conditional sale,
42:54really, because you could indeed show
42:56up with a product to some business or
42:58person, and they’ll just say," No
43:00thanks. "It's not even because they
43:03might not have a problem, but simply
43:05because they don't want to communicate,
43:07or they don't want you bothering them.
43:10And here, somehow, there might be some
43:12kind of insight, and from this, there
43:14might be some value for the idea and
43:16the business plan, perhaps.
43:18>> Well, there is definitely something to
43:19it, but there's still something wrong
43:21with this scheme.
43:22>> Well, yeah. So, it’s either a stage
43:251.5 between the business plan and the
43:27team, these mini-sales, or stage 3 vs 2
43:29, which is just what I was talking
43:32about, I don't remember now. Or maybe
43:34all of it together. And plus, what Ivan
43:38said about the approaches to the
43:39initial idea is also slightly different
43:42. There.
43:43>> Yes. Look, let me put everything into
43:45one category now, actually, and then I
43:47will try to explain why. Artem, I will
43:49remember. Let me say this first, and
43:51then we will get to you. I hope no new
43:53hands go up. So, three, three, actually
43:55. Raise your hands if anyone wants to.
43:57I’m joking, we can keep this going.
43:59So, we have three potential solutions
44:01right now. I am actually taking A/B
44:03tests and putting them there too, right
44:05? Well, strictly speaking, with A/B
44:06tests everything is more complicated,
44:08of course. Forget about A/B tests. So,
44:11we take" launching Windows "and put it
44:13there, right? Like, I have a business
44:15plan, I’ll launch Windows, and we put
44:17that in there too. I’ll go ahead and
44:18code the product right away. You know,
44:20that whole nightmare of hiring a team
44:22and developing it, all that crazy money
44:24being spent, you save on that, right?
44:25Everything is clear, not that much
44:27money has been spent, but there is
44:29still something wrong with this. I will
44:31tell you later, I'll guess what is
44:32wrong with it. Artem, please, go ahead.
44:38Yes, it seems to me that at stage 0.5
44:41we could add the formulation of
44:43specific product hypotheses, where we
44:45define exactly the audience, the
44:47problem, the solution, and perhaps the
44:49effect, so that when this cycle closes,
44:52we could change something in the
44:54product hypotheses and go through the
44:56iteration again—maybe not as full, a
44:59bit shorter, smaller. Artem, you are
45:03playing right along with me. That is a
45:05perfect transition to the next slide.
45:07Artem, thank you for that. An
45:09absolutely perfect transition.
45:12>> Yes. There you go. So, the so-called
45:14Lean approach is essentially what you
45:16just said, Artyom. Well, technically,
45:18yes? In the sense of the classic Lean
45:20approach, right? Actually, my
45:21interpretation in the form of another
45:23diagram, which you are about to
45:24criticize, yes, I sort of took this
45:25Lean approach and mapped it onto my
45:27scheme. But in essence, in essence,
45:28what Artyom said just now is the Lean
45:30approach. This is not what we are going
45:32to do. This is the classic Lean
45:33approach, right? I mean in the context
45:35of this transition, yes? So, we have
45:37some idea, we formulate it as a
45:39hypothesis, indeed, yes? So we
45:41formulate our idea as a hypothesis,
45:43build an MVP for it, and launch it to
45:45the consumer. This is what you were
45:47talking about earlier, when you spoke
45:49about" wipe-coding, "about" made
45:50Windows "—that's it. It all still
45:52falls under this. Then we look, what is
45:54wrong? We let them touch it, and they
45:56comment on it somehow. They say," No,
45:58we don't need it, "or they don't use it
46:00. So, we move on to the next iteration.
46:02This scheme is already better. It is
46:04still not ideal. Nowadays, you can move
46:06much faster, but it is better than the
46:08previous one. My question is, again,
46:10yes, it's not written on this diagram,
46:12but I will say it, yes, just take note
46:14that we had an idea, we formulated it
46:16as a hypothesis, prepared an MVP, and
46:18went to the consumer—not to sell, but
46:21to give it to them, well, to use the
46:23MVP. To use the MVP. We aren't selling
46:25anything yet. We let them use the MVP
46:27so they, accordingly, could give us
46:29some feedback. They come back to us
46:31with feedback, and we move on to the
46:33next iteration. How is this approach
46:35better than the previous one? In your
46:36opinion, Konstantin?
46:40>> My question is more about, well, the
46:42wording. May I ask? Yes.
46:47>> Well, go ahead, attack.
46:50>> Um, honestly, I've heard this somewhere
46:52before, but maybe, maybe I've forgotten
46:55.
46:55>> Excuse me. What? I'm repeating myself.
46:58>> No, no, I'm not talking about you, I'm
47:00talking about the Lean approach.
47:01Honestly, I don't remember. So. And so
47:04I wanted to clarify. There is an
47:05iterative model, there is an
47:07incremental one. Well, they are
47:08probably somewhat similar. Maybe it's
47:10even the same thing.
47:11>> So. And how does this differ from Lean?
47:14In essence, it's the same thing. Well,
47:16you take it, build some functionality,
47:19launch it, get feedback, analyze it,
47:21refine it, and release it again. The
47:26difference is just in the, well, in the
47:28volume, where in iteration we operate
47:30with a finished product, and here we
47:33operate with
47:33>> hypotheses.
47:34>> Let me answer that. Go ahead, let me
47:36answer. Let's just compare this Lean
47:38approach with what you're saying, these
47:40increments, right? We just take it and
47:42add value incrementally, right? We
47:44increment with each cycle, adding value
47:47each time—adding features, moving
47:49iteratively, and rolling out increments
47:51. They’re just about different things
47:53. I mean, the approach is essentially
47:54the same. We move in cycles. So it's
47:56all about working in cycles, right?
47:58Well, generally speaking, it really is
47:59all about working in cycles. That’s
48:01what they have in common. But when we
48:02talk about scaling up and increasing
48:04value through increments, where the
48:06management unit is a value increment
48:08that the team carries from start to
48:09finish, that's just a different thing,
48:11right? It's a matter of focus, it's a
48:14different perspective, right? But
48:16technically, it’s the same thing,
48:18isn't it? We just go step by step. Here
48:20, the unit doesn’t necessarily have
48:22to be an increment. We could completely
48:24redo the product on the second pass.
48:26But when we talk about increasing value
48:28through increments, we’ve built an
48:30MVP—some piece of the product—saw
48:32that it has value, and then we add
48:33extra features to it incrementally.
48:35Meaning, with every cycle, it becomes
48:37better, better, better, better, and
48:39better. That is the essence of the
48:41increment-based approach. That is
48:42exactly what we are seeing here. It's
48:44just about moving in cycles: we
48:46formulated a hypothesis, built an MVP,
48:48let people test it, got feedback,
48:49realized—for instance, we realized
48:51it’s not what we needed—but we
48:53didn't just realize that; we gathered
48:55data useful for decision-making and
48:57moved on to the next step. So we went
49:00and changed it, for example, or made
49:02something else entirely. That’s
49:06called a pivot within the Lean
49:08methodology framework. So, essentially,
49:11the goal is to iterate as quickly as
49:14possible in both cases, but in one, you
49:16increase value with increments, while
49:18here, well, people pivot until they
49:20find what’s needed. Vladimir, uh,
49:25hello everyone. If I understood the
49:28question correctly about how this
49:30scheme differs from the previous one,
49:32this scheme is simply cheaper and
49:34faster. That is, we don't assemble a
49:36team, we don't waste time, we test
49:38quickly, we test cheaply and
49:40effectively, and we get results
49:42immediately. Well, look, at least two
49:44people from the previous slide would
49:46answer this for you, and they said
49:47something like this. They said:" Wait,
49:50but can't we—let me go back to the
49:51previous slide—handle the development
49:54here for, like, 200 bucks with a
49:55low-code subscription? "" I just did it
49:57, and that's it. "" Why is this cheaper
49:59than that one? "" It’s going to be
50:01the same thing. "Well, it mentions a
50:03team before development,
50:05>> that needs to be assembled. But what
50:07kind of team do you need if one person,
50:09like a low-code dev, can do it all,
50:10right, without those subscriptions?
50:12>> No, okay. But we're talking...let me
50:14clarify just in case. The lean approach
50:15is a method from the 2000s; it was
50:17invented quite a long time ago, right?
50:18So now, of course, it has evolved like
50:20this. Well, nothing works like that
50:21anymore, right? No one operates like
50:23that anymore. Conceptually, that's how
50:25it works, but in essence, there are
50:26many nuances. That’s the first—the
50:28first part. The second part. I really
50:30did add a team here just to, well,
50:32drive up the cost a bit, just to make
50:34it look a bit more elaborate. But even
50:37if you remove the team from here and
50:39leave the subscription, there’s still
50:40something very wrong compared to this
50:42scheme. So, even if our development
50:46costs 200 bucks there, something is
50:49still wrong, while this one is better.
50:52Vladimir, you can jump back in if your
50:54thoughts align with this comment
50:56somehow. Yes, nothing much to add for
51:03now. It’s just all laid out in these
51:05clear steps, a business plan, all that
51:07—it's somehow overly complicated. In
51:10reality, everything should be simpler.
51:12>> Understood. In reality, the dynamics
51:14are different, I agree. But we are
51:16looking conceptually at what the
51:17difference in approach is. And this
51:19approach might indeed be faster, but it
51:21might not be faster. It’s not even
51:22really about speed, actually. Overall,
51:24if we really cover the development in
51:26the previous approach with low-code,
51:28say, it will also be fast. It’s not
51:29just about speed. I mean, speed used to
51:32be the key advantage of the lean
51:34approach, but in our day and age,
51:36that’s no longer the case. That’s
51:38true, but there is still a difference.
51:40Ruzal.
51:42>> Ah, yes, I think that in the previous
51:44stage we were trying to sell, and to
51:47test any hypothesis, someone has to buy
51:49it. But here, we aren't really selling
51:51it, so we immediately gather an
51:53audience that we can test it on. And
51:55that is the main difference. That’s
51:58true. Okay, I’ll follow up on your
52:00thoughts a bit later. Ivan, thank you
52:03very much. Ivan,
52:05>> I wanted to say that here, unlike the
52:07previous scheme, there is an analysis
52:09of the results. Because in the previous
52:12one, if something went wrong and there
52:13were no sales, we immediately went back
52:15to the initial stage, which meant
52:17something was wrong with the idea. But
52:20here, we actually have a separate,
52:22distinct stage where we look at what
52:24went wrong. And there isn't such a
52:27black-and-white scheme here.
52:30>> Uh-huh.
52:31>> There. Yes. Let’s have the two people
52:34who raised their hands speak now. Don't
52:35raise your hands anymore for now,
52:37because it's time for me to finish the
52:38thought I’m developing. Daria.
52:40>> Oh, yes. Well, in general, it seems
52:44that in this scheme we have a faster
52:46feedback loop, and we plan less, do
52:48more, test immediately, and return to
52:50re-validating the idea. That is, we are
52:56more focused on trying to do something
52:58before planning it in detail, launching
53:00it, and selling it; we test more at the
53:02initial stages and see if it’s
53:03actually needed in practice.
53:07>> Uh-huh. Yes, thank you. Understood. I
53:10like" feedback loop. "We’ll keep that
53:12in mind. I’m using it now. Konstantin
53:14, let’s have you be the final one.
53:17>> I don’t like this scheme because it
53:19lacks some kind of final outcome. Right
53:22. Well, like, it seems to me, and as it
53:27always happened, all the projects of
53:29mine that died were developed using
53:31this exact scheme.
53:34>> Right. Because from the first stage,
53:36from the moment the idea appeared to
53:38the launch for the consumer. In the
53:41first cycle, that consumer is most
53:42likely not the one who will pay you
53:44money and actually use your product.
53:47Well, because they like it. Most likely
53:50, it’s either some fake consumer you
53:52just found—like a friend or an
53:53acquaintance—whom you just tried to
53:55push your idea onto so they could test
53:57it for you, and you could look at their
54:00results. Or, most likely, it will just
54:03die at that first stage because
54:04there’s nothing to analyze, since
54:06nobody ended up using it. Nobody bought
54:09it, nobody did anything there, yeah,
54:12maybe you attracted someone, but, well,
54:14it's unclear here, like, if they didn't
54:17buy it, then what happens, right? If
54:19there are no results, then what do you
54:21do? It just dies again, that's it.
54:24>> There. And
54:25>> thank you. That's, that's, yeah.
54:27That’s it, let’s end it here. I’m
54:30interrupting you quite rudely because
54:31this is the thought I want to leave at
54:33the end. And the difference is truly
54:35fundamental in that the result here is
54:37binary. We either didn't sell or we
54:39sold. If we sold and then they start
54:41giving us feedback, super, we can work
54:43with that. But if we didn't sell, we
54:45learned nothing. I mean, and what do we
54:47do about that? We just didn't sell. We
54:49have no data, we don't know how to act
54:51next. So, we just didn't succeed. We
54:53didn't succeed and that's it. We can
54:55try 50 more times. We can keep
54:56wireframing until we're old. Meaning,
54:58you can test these ideas as much as you
55:00want, right? So, it will go on
55:01indefinitely. Even if it costs us, like
55:04, 5 bucks. For each new test, you can
55:05run them for as long as you like. Every
55:07time we'll just be shooting in the dark
55:09, hoping it works out. That is the
55:11fundamental difference from this scheme
55:12here. But as was just perfectly noted,
55:15this scheme can also go wrong if, well,
55:17hygiene rules aren't followed. We'll
55:20talk about these hygiene rules next,
55:21but the scheme will actually change its
55:23look now, it will be like, well, okay,
55:25I’ll talk about that now. Hygiene
55:27rules must be followed. And basically,
55:30the essence of the lean approach is
55:32that we shouldn't just go to any random
55:34person and pretend they are our user.
55:36It must be an actual representative of
55:38our target audience. We shouldn't just
55:41give it to them or ask them to touch it
55:43sometime; it should be integrated into
55:45their processes one way or another. I
55:48mean, in the lean approach, right, and
55:50we must collect feedback from them
55:52based on their usage experience. They
55:54might say," For example, I don't need
55:56this at all. "We can then find out what
55:58they do need. I mean, we were building
55:59something for you; you had some reasons
56:01to do it. There. So, in this sense,
56:04with this approach, we are constantly
56:06bumping into our potential client,
56:08gathering more and more data with each
56:10iteration and pivoting. In this sense,
56:13this approach, even though it’s
56:15actually an outdated one, is quite
56:17realistic now, precisely because
56:19development costs are quite low. But we
56:21will be using a different approach.
56:22However, I’ll start by dwelling on
56:24this one, to talk about it in a bit
56:25more detail. So, technically speaking,
56:28yes, this is just its classic scheme.
56:30It’s divided for us, in a way, into
56:32three stages, yes, this very lean
56:34approach is divided into three stages,
56:36but that’s a very rough split.
56:37Firstly, we are constantly conducting
56:40discovery; we make a MVP and run it
56:42until customer validation happens as a
56:44way out of the cycles. We are simply
56:46finding the so-called Product-Market
56:48Fit. Many of you have surely heard this
56:50term; it’s not really necessary to
56:52focus on it right now, but it’s just
56:54about the user or the market confirming
56:56the need for the product you’re
56:57bringing. So, in one of the iterations,
56:59you finally finish it to the point
57:01where it’s actually useful. And here
57:03they tell you:" Oh, yes, let’s do it,
57:05you found product-market fit. "And at
57:07that moment, you only start to scale.
57:09This is where sales start to happen.
57:10That is the essence of the lean
57:12approach. Now I am going to speed up a
57:13little because we’ve dragged things
57:14out. Well, I’ve dragged things out
57:16with you. But the essence is roughly
57:17like this. The idea is that we first
57:19find out what is needed. This is
57:20fundamentally what we will take from it
57:22. We first find out what is needed, we
57:24spend time on this, we find out what is
57:27needed, right? After that, we build
57:29what is needed, and then we will simply
57:31refine and increment it in cycles,
57:33which is what we discussed earlier, and
57:35then it can be scaled. But the whole
57:37idea is that until we collect data,
57:39until we understand what the user
57:41actually requires, what they will pay
57:43for. Firstly, yes, there are actually
57:45several questions, yes, firstly, who
57:47will pay? We just touched on this issue
57:49; yes, we might reach the wrong people,
57:51so until we find out who will pay us
57:52and, like, what they will pay for, we
57:54don’t start investing resources into
57:56it; we just try to iterate and collect
57:58as much data as possible. Our task is
58:00to accumulate as much data and as much
58:02knowledge as possible. Hands went up
58:04and down. That suits me. Well, the
58:07essence of this lean approach, or the
58:09lean startup approach, is a development
58:11methodology where we constantly iterate
58:14, cycle after cycle, to determine
58:16viability before we start scaling. Well
58:19, I think that is basically clear. The
58:21main businesses, this is important, we
58:23will partially take them for ourselves.
58:26So, folks, instead of a complex, deeply
58:28developed business model with
58:30preliminary analysis, they throw
58:32together a business model canvas, right
58:34? It has changed now. We won't be
58:36working with that one; we'll be working
58:38with a Lean Canvas. But the idea is, in
58:40principle, the same. It is a short
58:43canvas with specific questions, the
58:45answers to which we actually collect
58:47with every iteration. We draft it first
58:51, go into an iteration, enrich our
58:53understanding of how the market and the
58:55user are structured, fill out this
58:57canvas, and with each iteration, we
58:59become more informed. And customer
59:02development, item number one. Anyone
59:05who calls an interview" custdev "will
59:07go to a separate boiler in hell.
59:10Custdev and interviews are not synonyms
59:12. Custdev is a methodology. An
59:15interview is an interview; there are
59:17problem interviews, expert interviews,
59:18different kinds of interviews. These
59:20are different things, they are not the
59:21same. For some reason, it's common here
59:23to say we conducted a" custdev, "we"
59:24custdev'd "a little bit. That, in short
59:26, is not it. No, we are not"
59:27custdev-ing, "we are conducting an
59:29interview. When we conduct interviews,
59:30we might actually be doing customer
59:32development. But it is more than just
59:34an interview; there is a lot to it. Its
59:36essence is simple. Our task is to test
59:38our hypotheses using various tools,
59:40fundamentally with a potential consumer
59:42. Specifically, the person who will
59:44become our client. That’s it. And
59:46Java development. I won't be telling
59:47you about that. You know about it no
59:49worse than I do, perhaps even better.
59:51That doesn't concern us at all.
59:53Konstantin, you are not leaving me
59:55alone.
59:57>> Yes. Everything is very cool, very
59:59great, but there is one thing. I would
1:00:02gladly use this myself, if not for one"
1:00:04but. "
1:00:06>> Well, uh, to be more precise, I already
1:00:08use it, but there's one point I'm not
1:00:11satisfied with regarding customer
1:00:13development, specifically at the stage,
1:00:16>> uh,
1:00:16>> of testing this hypothesis, we are
1:00:18still working with a person who hasn't
1:00:20paid any money for it. They have a
1:00:22value proposition in the form of an
1:00:25idea, yes, or some problem, yes, they
1:00:27can give us feedback like:" Oh, I need
1:00:29this, I don't need this. "But in the
1:00:31end, they won't pay money for it.
1:00:33They'll just go to alternatives, say,
1:00:35log into ChatGPT and ask what to do
1:00:38with their pug,
1:00:39>> and they won't pay separately for
1:00:41anything else. Simply because there is
1:00:43no financial value in it for them. That
1:00:45is exactly what, it seems to me, breaks
1:00:47through in
1:00:48>> Stop. Stop, stop, stop, stop. I just
1:00:51want to establish that the approach you
1:00:53are using, from what you're saying, it
1:00:55follows that you're not using that
1:00:56approach, right? I mean, I understand
1:00:58that you are testing a potential
1:01:00consumer, but you say they aren't
1:01:01paying. And they say they need it, but
1:01:03they don't pay, so it's not valuable.
1:01:05It doesn't matter to us at all what
1:01:06they say. We aren't interested in that.
1:01:08We don't need them to tell us:" Wow,
1:01:10how great this is. "We don't need that.
1:01:12We aren't interested in that at all. We
1:01:14don't care what they say. That's not
1:01:15what we're interested in. We are
1:01:17interested in facts. We'll get to that
1:01:18a bit later. Well, it'll actually be in
1:01:21the next episodes. Technically,
1:01:22technically, when we talk about CustDev
1:01:24, it's not about asking whether you
1:01:26need it or not. Well, in the original,
1:01:28okay, it was close to that, right? I
1:01:30mean, it really was something like that
1:01:32. We won't be doing that. We aren't
1:01:33interested in what they tell us. You
1:01:35are right. Until they have paid with
1:01:37their own money, their words mean
1:01:38nothing. Therefore, the only thing that
1:01:40interests us about them, about each
1:01:42specific potential consumer, is the
1:01:43facts from their biography. Well, like
1:01:45specific facts in their life, specific
1:01:47events that we can rely on. Which we
1:01:49can, one way or another, interpret to
1:01:51confirm or refute our hypothesis. This
1:01:53just needs to be established. Their
1:01:55opinion doesn't interest us. It
1:01:56fundamentally doesn't interest us at
1:01:58all. We will never ask them for their
1:02:00opinion. Such a question should not be
1:02:02asked in any interview, neither by you
1:02:04nor by me. There. Uh, well that's one.
1:02:06And two. I am not saying that you need
1:02:08to use all of this. I am just talking
1:02:10about how it was. Next, we will talk
1:02:12about how we will work with you. So,
1:02:17this is our approach. What is the point
1:02:19? Everything here is trivial, right?
1:02:21Those of you who work in product, you
1:02:23probably already do this. You might use
1:02:25a slightly different toolkit, so it is
1:02:27super clear. The idea is simple. First,
1:02:30we formulate a hypothesis. We will talk
1:02:32about the hypothesis later, about how
1:02:33to formulate it correctly. Today, we
1:02:35will sort of gallop through it, but
1:02:37essentially, first we formulate a
1:02:39hypothesis. We are essentially trying
1:02:40to apply a scientific approach here. We
1:02:42talked about this earlier, right, that
1:02:44someone mentioned a scientific approach
1:02:45. Technically, we are trying to apply a
1:02:47scientific approach to product
1:02:48development, and we claim it is
1:02:50possible. And there is a mass of facts
1:02:52that testify to this. We formulate a
1:02:54hypothesis and design the study, that
1:02:57is, its design. A study that confirms
1:02:59or refutes this hypothesis. We conduct
1:03:02this study. Based on the results of the
1:03:04study, we understand whether the
1:03:05hypothesis was confirmed or not. One.
1:03:08That is, a binary answer, just as we
1:03:10would have if we had simply launched a
1:03:12product. Plus, we enrich ourselves with
1:03:16various kinds of knowledge. And our
1:03:19course will be about what tools we can
1:03:21use to test these hypotheses in various
1:03:23situations. Interviews will also be
1:03:26there. Interviews are part of customer
1:03:28development, but our interviews,
1:03:30although classic problem-based ones,
1:03:32will have some features; we will not
1:03:34use those that have been worn out 100
1:03:36times and are known not to work very
1:03:38well. We will, of course, use
1:03:40everything fresh. Konstantin, shoot.
1:03:46>> Like, in the study design, roughly
1:03:48speaking, will we change the toolkit at
1:03:50each iteration that we use to gather
1:03:52these facts?
1:03:56>> Well, look, if you formulate it so that
1:03:58the hypothesis is part of the study
1:04:01design, which is true, then at each
1:04:03iteration the hypothesis will change,
1:04:05but the tool can also change. Depending
1:04:08on what question you have or what
1:04:10information you want to get next, you
1:04:12will choose one tool or another. Well,
1:04:15>> yes, I just mean that potentially we
1:04:17can change the ways of obtaining
1:04:19information at each step.
1:04:21>> Well, theoretically, of course, you can
1:04:24do that. In fact, it will almost
1:04:25certainly not be the case, but
1:04:27fundamentally, of course, who is going
1:04:29to stop you?
1:04:29>> No, we won't change it like that. Well,
1:04:34I'm almost certain that you will have,
1:04:37well, a series of iterations using the
1:04:39same tools.
1:04:41>> Well, well that's understandable. I
1:04:43just understood the research design as,
1:04:45roughly speaking, a set of ways to
1:04:47conduct this research.
1:04:49>> No, you know, research design is
1:04:50something we will talk about separately
1:04:52. Research design is about who I'm
1:04:54going to, or whether I'm going to
1:04:56anyone at all, right? So, conditionally
1:04:58speaking, what experiment exactly do I
1:05:00need to set up right now to answer, to
1:05:02confirm or refute my hypothesis. That
1:05:04is, in general, a creative question
1:05:06every single time, right? So in that
1:05:08sense, the research design stage, when
1:05:10I sit down and I'm not just trying to
1:05:12endlessly conduct those wretched
1:05:13interviews, I'm actually formulating
1:05:15each time which question I don't have
1:05:17an answer to, right, and what I want to
1:05:19prove, confirm, or refute. Based on
1:05:21that, I choose the tool, design the
1:05:23research, and launch the research,
1:05:25whatever it may be. Then I get the
1:05:28result of this research and,
1:05:30accordingly, get the answer to my
1:05:31question. Plus some additional set of
1:05:33data. Yeah. I hope I answered your
1:05:39question. Well, come back if anything.
1:05:41You are our most active listener so far
1:05:43. Regarding the Business Model Canvas.
1:05:48We won't use the Business Model Canvas;
1:05:50we will use a fresher version. It’s
1:05:51not super fresh, but it is fresher. It
1:05:53has more data. It's the Lean Canvas. I
1:05:55think many of you have already touched
1:05:57it. For those who haven't, you will get
1:05:58to touch it. And today we are
1:06:00interested in three specific blocks in
1:06:02it. That's problem, customer segments,
1:06:05and solution. So, we are interested in
1:06:09the problem, customer segments, and the
1:06:11solutions you would like to pour into
1:06:12them. We will talk about this. I will
1:06:17certainly send you the template for
1:06:19this canvas, and it will say exactly
1:06:21what to write where. Now we will talk
1:06:24specifically about two blocks from this
1:06:26canvas that you will need to think hard
1:06:27about. I said this lecture would be
1:06:30light, but it's important to think a
1:06:31lot during it. So, we are interested in
1:06:34the problem and customer segments. It
1:06:37is fundamentally important for us at
1:06:38the very beginning to forget about the
1:06:40solution. It doesn't matter to us at
1:06:42all how or what we are going to develop
1:06:43there. Our task is to focus on two
1:06:45questions. Who is our potential
1:06:48customer and for what problem would
1:06:49they pay us? In other words, which of
1:06:53their problems should we be solving?
1:06:54Two questions, right, or a combination
1:06:57of these two questions. Let's start
1:06:59with the first one. Clients. Why? Well,
1:07:03let me put it this way. Why determine
1:07:05who to approach? I think there's no
1:07:06need to frame this question, it's
1:07:08obvious, right? If we don't know who to
1:07:10sell to, we won't go to anyone, or
1:07:11we'll go to just anyone. But our task
1:07:13now is to talk in detail about how to
1:07:15formulate and describe your client
1:07:17segment, especially when we're sitting
1:07:19in a vacuum all by ourselves. There are
1:07:23several segmentation criteria. You've
1:07:25100%seen them. I'm showing them now
1:07:27more as an anti-example. What do I want
1:07:29to say? There really are a ton of them.
1:07:31You can segment a client, describe your
1:07:33client, yes, give their profile.
1:07:34Someone here might say:" Let's describe
1:07:36the ICP. "We won't be describing the
1:07:38ICP, that's too much. We just need a
1:07:39basic, short description now, but one
1:07:41that is specific and clear for starting
1:07:43, for getting up to speed, for
1:07:44conducting research. So, what can you
1:07:46use? You can say that my client is, say
1:07:51, women from, I don't know, 45 years
1:07:54old, not married. In short, start
1:07:58tossing out these cliché demographic
1:08:01traits. What do I want to tell you? You
1:08:03can use them, but only on the condition
1:08:04that they are at least somewhat
1:08:06significant, meaning they actually have
1:08:07some meaning. What do we need? We need,
1:08:09well, okay, we'll talk about that now.
1:08:11In short, there are many segmentation
1:08:13criteria, that's just important to
1:08:14understand. You can come back to this
1:08:16slide later, after the lecture, when
1:08:18you're doing your homework, look at it,
1:08:20and understand what else is worth
1:08:21considering. But I want to say this. So
1:08:24, apart from these demographic criteria
1:08:26, like when describing your target
1:08:27audience and your client, there are
1:08:29many others. Yes, geography might be
1:08:31important to us; there are products for
1:08:33which geography matters. Yes, meaning
1:08:36if we are making a product that affects
1:08:38, for example, I don't know, American
1:08:40legislation, then the region is
1:08:41important to us. But it's important
1:08:44that it's in America. Then we write
1:08:45down America. If we are talking about
1:08:49the fact that our user is, well, we are
1:08:51making a product, for example, for, I
1:08:53don't know, dog groomers, for instance.
1:08:56Well, for some reason we're dancing
1:08:57around a pug today. Let's go with
1:08:59groomers. So, we're making it for dog
1:09:01groomers. This is a product literally
1:09:02for them. We'll help them groom more
1:09:04easily, well, I don't know, yes. Then
1:09:06we specify the field of work. If the
1:09:07field of work doesn't matter, you don't
1:09:09need to write it down. Well, and so on,
1:09:11right? That is, you can use all the
1:09:13attributes, provided they actually have
1:09:15some significance for your specific
1:09:17idea. So, there are many criteria. And
1:09:21what is needed, what is the task? You
1:09:23will need to segment and describe a
1:09:25segment of your target audience. This
1:09:26will be your homework assignment, one
1:09:28of two. So, describe a segment of your
1:09:30target audience. You will need to come
1:09:32up with a product, join into teams to
1:09:34work on this product, or conversely,
1:09:36join into a team first, then come up
1:09:37with a product together. Just think of
1:09:39some idea that you like and formulate a
1:09:41target audience for it. Who is your
1:09:43target audience? Who are the people who
1:09:45will pay you? Who are your users? These
1:09:46, by the way, are not necessarily the
1:09:48same people. One might pay, and another
1:09:49might use it. It is important to keep
1:09:51this in mind. If you have a complex
1:09:52multi-actor model, you will also have
1:09:54to deal with this. But basically, you
1:09:55need to describe your target audience.
1:09:57So, someone has their mic on, please
1:09:59turn it off. Thank you. Requirements
1:10:05for segmentation. You know, it's quite
1:10:07easy to get lost in segmentation,
1:10:08especially when you do it for the first
1:10:10time. The requirements are simple. So,
1:10:12first of all, your segment must be
1:10:14measurable at least to some extent.
1:10:15That is, well, it should be possible to
1:10:18measure it somehow, right? Meaning it
1:10:21has a clear size, it is well-defined.
1:10:23The size is sufficient, right? That is,
1:10:26it’s not just measurable, but it’s
1:10:28not like 3.5 people for the whole world
1:10:30, right? It should be a segment you
1:10:32describe that is large enough to earn
1:10:34money on, so that if they give you
1:10:35money, you can make a living from it.
1:10:38Next, it must be accessible. That is,
1:10:41it is a segment that you could
1:10:43theoretically reach. Maybe with some
1:10:45effort, maybe through acquaintances,
1:10:47maybe by expanding your staff, it
1:10:48doesn't matter. What matters is that it
1:10:50is possible to reach them. The
1:10:52representatives of this segment, the
1:10:54people within this segment, are truly
1:10:55homogeneous, that is, they share the
1:10:57same characteristics. And at the same
1:10:59time, these characteristics differ from
1:11:01those of neighboring segments. And also
1:11:02stable. What does that mean? It means
1:11:04that while you are developing this
1:11:06product, the segment won't change.
1:11:08People won't drift away from it, it
1:11:10won't transform, it won't vanish,
1:11:11nothing will happen to it. It's just
1:11:13important to keep that in mind.
1:11:16Detailing is also an important point I
1:11:17want to get across. I want to save you
1:11:23from trying to define your target
1:11:25audience as, like," my target audience
1:11:27is everyone, "you know? You might say,"
1:11:31Well, Instagram's target audience is
1:11:33everyone. "And that would not be true.
1:11:35Or you might say," Well, like, VK,
1:11:37their target audience is also everyone.
1:11:40"And that's not true either. There are
1:11:41just many different segments there. In
1:11:45short, your task is actually to
1:11:47gradually narrow your circle down to a
1:11:48size where, when you describe this
1:11:50segment to me and say they have a
1:11:52problem, I can't name an example of
1:11:53someone in that segment who definitely
1:11:55doesn't have it. It is important to pin
1:12:00this thought down. So, how do we do
1:12:02that? Essentially, we choose a domain;
1:12:04for example, I say I'm making a product
1:12:07for pathologists. My domain is medicine
1:12:09, pathomorphology. I've outlined the
1:12:11first circle. So, I formulate their
1:12:13problem. People who are working in a
1:12:15high-volume...Hello again. I got
1:12:56disconnected. I will repeat it now. So,
1:13:01going back to my example, I have
1:13:03pathologists, which is essentially the
1:13:05first circle. The second circle is
1:13:07people with my problem. Meaning people
1:13:09who I assume have this problem, and
1:13:10this will add attributes to them. The
1:13:12problem as I define it sounds something
1:13:14like this: pathologists writing their
1:13:16reports—their pathohistological,
1:13:18pathomorphological reports—anyway,
1:13:20they write their reports and (I'm
1:13:22saying their problem is that they)
1:13:23spend a ton of time pulling text from
1:13:25various files and pasting it into the
1:13:27report form they use in their medical
1:13:29information system. So I say, they just
1:13:31spend a lot of time on this. Then I ask
1:13:33:" Well, who has this problem? "Well,
1:13:35obviously, this problem is significant
1:13:37for those, it exists for those, it is a
1:13:39problem for them. For those whose flow
1:13:42of studies is small—if they write one
1:13:44report a day, they don't have this
1:13:46problem. Okay, I have already narrowed
1:13:48the circle. These are people who work
1:13:50in a high-volume lab where the workload
1:13:52is heavy. Next, we talk about people
1:13:56who are aware of this problem. Who is
1:13:58aware of this problem in my
1:14:00hypothetical scenario with the
1:14:01pathologists? Well, it's those who are
1:14:04overloaded because of it. Meaning,
1:14:05people who are held to high standards,
1:14:07who are required to do all of this
1:14:09quickly. But they just can't seem to
1:14:12get rid of it. And so they end up
1:14:13working overtime, overworking,
1:14:15generally feeling sad, and losing sleep
1:14:17. These are the people who will feel
1:14:18the problem. They are aware of it. This
1:14:20is fundamentally important. Then, among
1:14:22them, I need to find those who are
1:14:23willing to pay. That will be my
1:14:25research task. Yes. Well, actually, all
1:14:27of this is my research task, but
1:14:29basically, my narrative should be built
1:14:30this way from the start. And here is
1:14:33just a kind of example, yes, I'm just
1:14:35sketching out an example of how it
1:14:36could be. Of course, it needs to be
1:14:38more specific than what I wrote here,
1:14:40but just so the approach is clear to
1:14:42you, yes. We don't say that, uh, my
1:14:43product is a product for young people;
1:14:45we need to be specific, yes, we have to
1:14:47tailor it depending on what the product
1:14:49is. We say that these are, for example,
1:14:51first-year students who moved to
1:14:52another city to study and live in a
1:14:54dormitory. My product is somehow
1:14:55connected to this, meaning I am solving
1:14:57a problem related to this. That's it.
1:14:59Well, there will be examples here, you
1:15:01can look at them and read them later,
1:15:02but basically, I think it should be
1:15:04clear. Are there any questions here? Or
1:15:08shall we move on to the segment problem
1:15:09? Is it clear what to do to describe
1:15:12your segment? Vladimir is looking sad.
1:15:18Gloomy, I would even say. Tired. Yes,
1:15:21tired. Okay, let's talk about the
1:15:27problem. Uh, about the problem and
1:15:29everything else. We are actually
1:15:30already finishing up. Uh, there is a
1:15:32definition, right? I didn't invent it.
1:15:34It was invented in 1972 by those New
1:15:37guys. Well, in general, New is credited
1:15:40with this phrasing, but it has an
1:15:42author. So, a problem is the gap
1:15:43between how things are now, meaning how
1:15:45they actually are, and how a person
1:15:47wants them to be, right? Meaning, well,
1:15:49it's like the difference between
1:15:50expectations and reality. It might seem
1:15:52simple, but if we break this topic down
1:15:54a bit, it might seem like a basic,
1:15:56unimportant definition, but it is
1:15:58fundamentally important. What is
1:16:00fundamentally important about this
1:16:01definition? It is important to us that
1:16:03the person wants something. It is
1:16:06important to us that this gap is
1:16:07perceived by the person themselves, the
1:16:09consumer themselves. In other words, if
1:16:12we see from the outside that someone is
1:16:14working inefficiently, but they
1:16:16themselves don't feel any problems,
1:16:18then for them, it's okay. Yes, he’s
1:16:21spending five times more time on it
1:16:22than he could be. He has no problem. We
1:16:27, as outside observers, see this
1:16:29problem, but he won’t buy anything.
1:16:32He won't buy a solution from us because
1:16:34, in his worldview, the problem doesn't
1:16:36exist. A hand is raised. Go ahead.
1:16:40>> Yes, I have a question. I've seen this
1:16:43phrasing often, and I always have the
1:16:45question: what does" wants there to be
1:16:47"mean? I mean, take the taxi example;
1:16:50we used to order taxis by phone, send a
1:16:52text, and the taxi would arrive, which
1:16:55generally didn't take much time. Well,
1:16:59an app appeared that showed you could
1:17:01just order a taxi. But at the time when
1:17:03people were ordering taxis by phone,
1:17:05that solution wasn't obvious. Now we
1:17:09can obviously say that a person wants
1:17:11an app to exist, but in that moment,
1:17:13when people are ordering by phone, for
1:17:15us it's...wait, wait, this is important
1:17:17. We cannot define the problem as...The
1:17:20person wants there to be an app. That's
1:17:22like saying I'll sell you a shovel
1:17:23because you're a person who wants a
1:17:25shovel. No, we are talking about a
1:17:27problem. He has some sort of problem.
1:17:29He doesn't want your solution; he just
1:17:30wants the problem to go away. I mean,
1:17:33say, for example, I'm sitting there,
1:17:35let's just imagine, I'm sitting there
1:17:37and I don't know, joining two parts
1:17:39together by hand. So I'm sitting there
1:17:41on a stream, assembling and joining two
1:17:43parts by hand. That's my business. That
1:17:45is my business. I'm sitting here doing
1:17:47it by hand. I spend 20 hours a day on
1:17:48this because it's my business. Now, for
1:17:50me, that's a problem. Well, for example
1:17:52, I don't actually do that, but that's
1:17:53an imaginary me, from another parallel
1:17:55universe. So I'm sitting there doing
1:17:56that. The problem for me is that I
1:17:58spend 20 hours on this, because I'll
1:18:00spend the next two on reviewing taxes,
1:18:02doing other things, selling stuff, and
1:18:03somehow spinning my wheels. For me,
1:18:05that is the problem: that I'm spending
1:18:07my time on this. That is my problem,
1:18:09which I am aware of. In this definition
1:18:11, it's important that I am aware of
1:18:13some problem for myself. I see it. Not
1:18:15you noticing from the outside that I
1:18:17have a problem. You might assume that
1:18:20since a problem exists in principle,
1:18:21and since there is some time to be
1:18:23saved, and you can offer me that time
1:18:25savings, then I might have a desire to
1:18:26obtain those savings. But until you
1:18:29interact with me, you won't know if I
1:18:31actually have that desire. There. So
1:18:33your task, our task, when we formulate
1:18:35a problem, is to formulate the problem
1:18:37itself. The problem for this person
1:18:40sitting there assembling parts is that
1:18:41they spend 20 hours of their time
1:18:43assembling parts and can’t do
1:18:44anything about it. Accordingly, when
1:18:47you come to them with your solution,
1:18:48you must solve the problem for them.
1:18:50How exactly? Basically, they don't care
1:18:52. They will start to care when you come
1:18:55to them, and I come to them, and
1:18:56Vladimir comes to them, and Artyom
1:18:58comes to them, and we all come to them
1:19:00offering different solutions. That is
1:19:02when they will start choosing. That is
1:19:03competition. That is where competition
1:19:05starts to play a role. But it is
1:19:06fundamentally important that there was
1:19:08a problem at the very beginning. They
1:19:09have a desire, and their basic desire
1:19:11is to get rid of the problem. They feel
1:19:13it. If a person understands the problem
1:19:15, they are aware of it. This right here
1:19:17is our target audience. This is our
1:19:19target audience. Those who are aware of
1:19:20the problem. I'll go back to this slide
1:19:22. What is this? To this slide, yes?
1:19:26People who are aware of the problem, of
1:19:28its existence, they realize they have
1:19:29it. This is fundamentally important.
1:19:32There. That is the essence of it. When
1:19:34we talk about taxis, about how people
1:19:36used to order by phone, and now you can
1:19:38call a taxi. When ordering by phone,
1:19:41there was some kind of problem, there
1:19:42was actually a whole set of issues,
1:19:44which is why this Uber format took off.
1:19:47And then for another 10 years, everyone
1:19:49said," I have an Uber-like solution. "
1:19:51Well, in short, there was a problem
1:19:53there. It wasn't that people didn't
1:19:55want Uber and then suddenly did. Of
1:19:58course, nobody ever wanted Uber. It
1:20:00wasn't like someone was sitting there
1:20:01wanting it. They had a problem, and
1:20:03they were offered a solution to that
1:20:05problem. That's all. So, Uber simply
1:20:09offered a solution to that problem in
1:20:11some way. That's why I say we don't
1:20:13need the solution at the very beginning
1:20:15. We only come up with a solution with
1:20:17you just to have something to start
1:20:18with. Actually, we will abandon this
1:20:21solution later, well, like 99.1%, the
1:20:23solution you come up with will fall
1:20:26apart after the very first research
1:20:28attempt. I am sure of it. That is why
1:20:30the solution doesn't matter, it's
1:20:32important to find the problem. A
1:20:33problem is what the user themselves,
1:20:35the person, the client perceives as
1:20:36their own problem. Signs of a problem.
1:20:41Well, if we try to formalize it, first
1:20:43of all, it shouldn't contain a solution
1:20:45, right? Meaning," Uber "shouldn't
1:20:47appear anywhere here. Furthermore, it
1:20:49is recognized by the person themselves,
1:20:51which is what I emphasized. Next, it is
1:20:53measurable. Well, as to how to measure
1:20:56a problem, if we keep it simple, a
1:20:58business problem is always measured in
1:20:59money. Almost always it is measured in
1:21:02money. In some businesses, it is
1:21:05sometimes measured by the founder's
1:21:06time. That works too, right? I mean,
1:21:09it's significant to them as well, but
1:21:10it's measured either by money or by
1:21:12time. Sometimes by both, as one is
1:21:14easily converted into the other. We can
1:21:16easily convert money into time, and
1:21:17time into money. So, it's easy to
1:21:19calculate. In general, the problem is
1:21:21measurable. And the problem is, what
1:21:23you call, repeatable, right? It recurs;
1:21:25it's not a one-time situation. It's not
1:21:27that the problem happens to them once
1:21:29in a lifetime. You come up and ask:"
1:21:31Have you had such a problem? "They say:
1:21:32" I have. "Like, did you have
1:21:34chickenpox?" I had it, "or" I missed
1:21:36out on that. "There. So, it's important
1:21:39to us that the problem is frequent for
1:21:41the person. And the most important sign
1:21:43is that they are already trying to
1:21:45solve it somehow. They are already
1:21:47doing something to get rid of it,
1:21:48anything at all. If a problem exists
1:21:52and they tell you it's significant, but
1:21:54they aren't doing anything to solve it.
1:21:57Almost certainly, the problem is made
1:21:59up. They invented it themselves, or it
1:22:01just doesn't matter to them. Therefore,
1:22:04if they haven't even tried to solve it
1:22:05themselves, they almost certainly won't
1:22:07pay you to solve it either; they must
1:22:09have at least tried to solve it in some
1:22:11way. A problem, well, is what I just
1:22:17mentioned; for example," doctors need
1:22:19an assistant for reports "is not a
1:22:21problem. The problem is that a doctor
1:22:23spends a certain number of hours on a
1:22:25specific task. Uh, a hand again, go
1:22:27ahead and ask your question. Ah, yes,
1:22:29Vadim, Duma, I have a question.
1:22:31Listening to you right now, I’m
1:22:33experiencing a bit of a paradox. Okay,
1:22:37you said we start with a problem, but
1:22:39to pick a problem, you at least need to
1:22:40define the domain where you’ll look
1:22:42for those problems and people. It feels
1:22:46like the circle, I don’t know, gets
1:22:48broken. I mean, we search for a problem
1:22:50, segment, define the people, and
1:22:52describe them. But in fact, how do you
1:22:55even get to that problem? You initially
1:22:57start by thinking," Okay, where and for
1:22:59whom do I even want to do something? "
1:23:03And it seems like you immediately skip
1:23:05the stage where you've defined
1:23:06something and then started segmenting.
1:23:09You know, do you understand what I’m
1:23:11saying? How do you get to that problem
1:23:14first, how do you define it, and at the
1:23:16same time
1:23:18>> not break this logic?
1:23:20>> I’ll jump straight to the homework
1:23:21and explain it right there. The essence
1:23:23is simple. So, what needs to be done
1:23:25now? You’re right. You’re right
1:23:27that, well, I haven’t researched
1:23:29anything yet, right? I haven’t
1:23:30researched anything. I’ll just pick a
1:23:32segment now, I’ll just choose it and
1:23:34invent a problem for it. That’s your
1:23:35thought. Say yes or no. If not, correct
1:23:38me.
1:23:38>> Well, yeah, yeah. I mean, for sure, I
1:23:40can’t just, you know, take it out of
1:23:42my head randomly." Oh, now I’m going
1:23:44to do something for the mobs, "and
1:23:45that’s it. Well, no, first of all,
1:23:47you can, who’s going to stop you? Do
1:23:48you want me to come up with it for you?
1:23:50We’ll do it for the mobs. It’s not
1:23:51about whether you can, like, who’s
1:23:52going to stop you. The idea is simple.
1:23:54First of all, all of us, one way or
1:23:56another, encounter some problems
1:23:57ourselves. Like, everyone has some kind
1:23:59of problems. One way or another, we try
1:24:01or don’t try to solve them. That’s
1:24:02one. Two, we observe problems
1:24:04externally. Something might seem unfair
1:24:06or wrong to us. In this sense, the task
1:24:08for those launching products is, in
1:24:10general, to solve such things, right?
1:24:11You might see this problem somewhere
1:24:13outside, you might see something and
1:24:15think:" Well, damn, like I see people
1:24:16who are suffering, I can help them
1:24:18somehow, yes, I see this problem,
1:24:19something here is unfair, I understand
1:24:21how it is, but I want to try to solve
1:24:22it somehow. "Yes. That's how you can do
1:24:24it. You can just go from yourself, it
1:24:27turns out, starting in any case. A
1:24:30product is an ugly copy of the one who
1:24:32makes it. It's always like that. Well,
1:24:34that is, we are talking about the fact
1:24:36that we are going to work on a startup
1:24:38now. That's all. We are going to work
1:24:40on a startup now. We are going to
1:24:41launch our own product. Essentially, we
1:24:43are going to develop our own product. I
1:24:44choose which product to make. What is
1:24:46important, look, let's look at my
1:24:48homework. I write item one: get into
1:24:50teams. If you want to go solo, you can
1:24:52go solo. But ideally, well, my
1:24:54recommendation is to gather a team so
1:24:55that you can split up tasks, yes, it
1:24:57will just be more convenient to move
1:24:58forward that way. But solo is also
1:25:00possible. And come up with a product
1:25:02idea. Come up with a product idea that
1:25:04ignites you. You need to choose what
1:25:06will drive you. Well, there is no point
1:25:08in doing it. Like, you can't, you know,
1:25:10just Google the most significant
1:25:12problems not solved by other products.
1:25:15It won't work that way. You must choose
1:25:17something, a solution you are ready to
1:25:19work on, which will drive you, which
1:25:20will motivate you. Therefore, yes, you
1:25:24choose, you choose the target audience
1:25:26for the problem you want to solve, and
1:25:28you choose the problem you already know
1:25:30, or see, or assume you will solve.
1:25:32I'll tell you right away that after the
1:25:34very first cycle, not only will the
1:25:36solution likely fall apart, but most
1:25:38likely, the problem itself won't be
1:25:39confirmed. Or you will realize that the
1:25:41segment is described incorrectly and it
1:25:43is a little different, something will
1:25:44change, and that's normal. We will move
1:25:46in cycles. But you need to enter the
1:25:48cycle. To enter it, you need to
1:25:49formulate at least something. Therefore
1:25:53, at the very beginning, we must
1:25:55formulate in the Lean Canvas, which I
1:25:57will send to the chat, we must write
1:25:58down the segment, the problem of this
1:26:00segment, and how we would like to solve
1:26:02it. These are the three blocks we must
1:26:06fill out. If someone knows how to fill
1:26:08out the whole thing, they can fill out
1:26:09everything. But our task at the current
1:26:11stage is to fill this out. Next, I
1:26:15think we will have a consultation on
1:26:17Friday. So, my recommendation is for
1:26:19you to fill this out by Friday; that is
1:26:21, draft an idea and think about what
1:26:23you would like to write, following the
1:26:25rules I formulated earlier. Don't just
1:26:27write it haphazardly, don't describe
1:26:28the segment haphazardly, and don't
1:26:30describe the problem haphazardly. The
1:26:32criteria are provided, and I have given
1:26:34examples. I expect you to correctly
1:26:35formulate both the problem and the
1:26:37target audience. Importantly, you might
1:26:40be wrong. The problem might not
1:26:42actually exist. That doesn't matter.
1:26:44What matters is that this is simply a
1:26:45well-founded assumption. I need you to
1:26:48get together with your team and, as a
1:26:50team, reasonably assume that a target
1:26:52audience—which you've written down—
1:26:54has a specific problem—which you've
1:26:55written down—and then consider how it
1:26:57could be solved and write that down.
1:27:00This needs to be done. We will proceed
1:27:02to research from there. For the
1:27:06consultation on Friday, I suggest you
1:27:08sign up with mentors and hash it out
1:27:09together with them. I will write more
1:27:13about the consultations in detail later
1:27:15, but that’s the idea. Right. Hand.
1:27:19>> How many people should be in a team?
1:27:21How are we supposed to get together
1:27:22anyway? What is the point of all this?
1:27:25>> What is the point? The point is not to
1:27:26do everything alone, right? You can—
1:27:28No, I mean, okay, we'll have three or
1:27:30four people there, and we just—
1:27:32>> Let's limit it to three, up to three. A
1:27:34team of up to three people, yes? Are we
1:27:36just discussing the idea, with someone
1:27:38throwing out suggestions? Yes, that’s
1:27:40the work system. And by the end of our
1:27:42course, are we just going to refine a
1:27:44hypothesis into something decent?
1:27:47>> Yes. Well, in general, you could put it
1:27:50that way, yes. I spent a long time at
1:27:52the beginning explaining what we'll be
1:27:53doing, but, essentially, if you boil it
1:27:55down, you could say that; but overall,
1:27:57yes, you join a team and formulate your
1:27:59starting point now. This is, in essence
1:28:01, your starting assumption. I won't
1:28:04call it a hypothesis, although in a
1:28:05sense it is, because we will refine the
1:28:07hypothesis later, but it is your
1:28:09initial assumption that we will start
1:28:11with, and at some point, it will
1:28:13materialize into something proven. For
1:28:16now, it’s just an assumption. Let’s
1:28:18describe the assumption, pull it out of
1:28:20your head, from observations of friends
1:28:21; can we ask our moms? In short, any
1:28:23way you can.
1:28:26>> And please send us the presentation as
1:28:27well.
1:28:28>> I will, of course, there will even be a
1:28:30recording. Well then, ladies and
1:28:35gentlemen, any more questions? Comments
1:28:38? Who is going to provide hate? Hate,
1:28:39someone will come up to you and say:"
1:28:41Well, Vadim, this is such nonsense, I
1:28:43just wasted an hour and a half. "" What
1:28:45kind of garbage is this? "Raised hand.
1:28:48Go ahead.
1:28:49>> No, Vadim, there won't be any hate here
1:28:51. Look. And by
1:28:54>> Okay. Alright, I'll prepare some
1:28:56specifically next time. Look, you said
1:28:59that you define what drives you, right?
1:29:02Things that you like, and you just sort
1:29:04of validate them and test them out.
1:29:06Well, look
1:29:07>> I didn't say that I do that. Wait, I
1:29:09didn't say that I do that. I said that
1:29:10I am suggesting you do that, because
1:29:12you will be working through it.
1:29:13>> Wait, wait. Why do we need to do that?
1:29:16We need a verified option right away,
1:29:18because we want to refine it.
1:29:20>> But if you pick what you like, then
1:29:21others might not...
1:29:22>> Let's. Fine, let me explain, let me
1:29:24tell you how it actually happens in
1:29:25reality. In my practice, it looks like
1:29:27this. Especially back at Repetixell, it
1:29:28was happening all the time there. It
1:29:30was like this. We take an idea, we are
1:29:31currently working through some
1:29:33hypothesis. So, like, a product idea
1:29:35emerged, someone came up with a product
1:29:37idea, we did some desk research, so to
1:29:39speak—we’ll talk about that too.
1:29:40Well, generally, we dug up info
1:29:42suggesting the idea isn't absurd in
1:29:44principle. We start researching it. We
1:29:46start researching it and immediately
1:29:47stumble upon a more adequate, more
1:29:49verifiable idea. With every iteration,
1:29:51we collect more info, but it comes
1:29:52about like this. An idea can be taken
1:29:54from absolutely anywhere. I’m just
1:29:55suggesting you choose the scenario that
1:29:57you like. I mean, you’ll like working
1:30:00on your own idea more than on mine. An
1:30:03important nuance. If any of you don't
1:30:05come up with anything and say:" Vadim,
1:30:06give us a product and some hypotheses
1:30:08to work through, "I will give them to
1:30:10you. There. But I would like you to
1:30:12take something you are interested in
1:30:13working through. That is, a problem
1:30:15that you would be interested in solving
1:30:17for real, like, truly, because you have
1:30:19every chance to actually solve it
1:30:21during the course. Uh-huh. So, the main
1:30:25point is that we just need to start
1:30:27with something, and in the process of
1:30:29these iterations, we can arrive at
1:30:30something adequate. Right?
1:30:32>> That’s how it will be, yes? That’s
1:30:33how it will be. Basically, you are just
1:30:35going to choose a domain, a target
1:30:36audience, and a problem right now. And
1:30:38these things can be changed, so to
1:30:40speak. With each new iteration, you can
1:30:41say:" Damn, I just realized the
1:30:43solution is cool, but, in short, I
1:30:44shouldn't go to these people, but to
1:30:45those others in a completely different
1:30:47domain, and you go there. "That's also
1:30:49allowed.
1:30:49>> There you go.
1:30:50>> Well, that's beer, yes, understood.
1:30:52Okay, good.
1:30:53>> Okay.
1:30:57>> Sofia,
1:30:58>> yes. Vadim,
1:31:00>> hello. And a question, well, I don't
1:31:03know, not hate, but I would like a
1:31:05survey to be conducted on us too. Uh,
1:31:09>> maybe some materials to be provided as
1:31:10additional links to what can be studied
1:31:12in more detail. It's just that when
1:31:17you've theoretically done this many
1:31:19times, you just want there to be a
1:31:21variety of levels, say, something to
1:31:22read or maybe some best practices, or
1:31:24other interesting methods. Well, just
1:31:29for myself, so for those who want, like
1:31:32, extra materials for extra
1:31:33familiarization, just to have them on
1:31:35top. Because I'm also personally
1:31:38interested in digging deeper into this
1:31:40topic. That's it. Okay. Well, Sofia,
1:31:45you realize now that since you've said
1:31:46this, all the mentors will hear it,
1:31:48they will watch the recording, and when
1:31:50you bring them a problem or a target
1:31:52audience, they will expect it to be
1:31:54just" wow. "
1:31:57>> Well, I will try it on a new product,
1:31:59on the idea that I want to develop for
1:32:01my diploma. So, in essence, it will be
1:32:04an important story for me to work on
1:32:06over these two years.
1:32:07>> Yes. And, look, an important nuance is
1:32:09that in the lectures we essentially set
1:32:11a vector, we break down some important
1:32:13things, but the real highlight, in my
1:32:15opinion, based on previous experience,
1:32:16is actually the consultations. Because
1:32:18during the consultations, we all
1:32:19cross-pollinate, and by the way, you
1:32:21can go to other people's consultations,
1:32:22you don't have to be the one consulting
1:32:24, you can come and listen to what
1:32:25others are being told. That's also
1:32:27useful, so we will split into rooms,
1:32:29sign up for slots, and you can wander
1:32:30back and forth to listen to it all.
1:32:32It's super cool. It's like, you're
1:32:35observing live how one idea or another
1:32:37is analyzed and you hear the feedback.
1:32:39That's also very useful. In this sense,
1:32:41Sofia, something might come up that's
1:32:43directly from practice, something you
1:32:45won't read about, but rather, well,
1:32:47only a real person who works in this
1:32:48field can tell you that, for example,
1:32:50things don't work like that in this
1:32:52domain at all—they work differently.
1:32:54>> Yeah.
1:32:54>> But I understood about the additional
1:32:56material, yes, I will try to provide
1:32:58some. No problem. Nikita,
1:33:02>> hello. I have a question about the
1:33:05assignment, uh, like from the session,
1:33:07that there must necessarily be a
1:33:09problem we want to solve. But looking
1:33:15at these steps, I understand that it
1:33:17should start with an idea, and while
1:33:19working on the idea, we should find
1:33:21some problems. Well, what if it turns
1:33:27out that this product we came up with
1:33:29doesn't solve any problems at all?
1:33:32>> Wait, wait, I'm saying that right now
1:33:34we are not coming up with a product
1:33:35without problems. We fundamentally do
1:33:37not do that. What do we do? We sit down
1:33:39at the very beginning and, again, with
1:33:42a sound basis, we essentially formulate
1:33:44an assumption, describe this segment.
1:33:46There is this problem, we describe it
1:33:48and estimate that we could solve it
1:33:50roughly like this. Then we sort of
1:33:53clean it up, think about it a bit, and
1:33:55start researching. We do not research
1:33:57the solution. We don't care about the
1:33:59solution for now. We are researching
1:34:01the problem and the target audience.
1:34:03That is what we will do right from the
1:34:04next lecture. We will research the
1:34:06problem and the target audience.
1:34:08>> Research and finding problems is part
1:34:10of the second step. Yes,
1:34:12>> but we are essentially starting to
1:34:13verify it right away, yes, we will
1:34:15verify it with you right away; meaning
1:34:16by conducting research, in the first
1:34:18iterations, we will investigate whether
1:34:19the target audience actually has the
1:34:21problem. We will change hypotheses
1:34:22about the target audience, hypotheses
1:34:24about the problem, and we won't get to
1:34:25the solution hypothesis for a while. We
1:34:27don't care about the solution for now.
1:34:29It doesn't matter at all how to solve
1:34:30it. We will definitely figure out how
1:34:32to solve it. I mean, I'm sure that any
1:34:34of you will decide and find a solution
1:34:35to the problem you unearth. The task is
1:34:39to find a significant problem for a
1:34:40large, large audience. Or at least some
1:34:43sort of significant problem for a
1:34:45significant size of audience. Find it
1:34:48all. This is step number zero. This is
1:34:50what we need to start with. We will be
1:34:52stuck here for a while. That's how it
1:34:59is.
1:35:03>> Ivan,
1:35:05>> yes, Vadim, look. And here's the killer
1:35:08question:" How do you yourself
1:35:10personally, when it comes to you, well,
1:35:12how do you look for ideas? "Let's keep
1:35:14it simple,
1:35:15>> No, look, let’s talk about ideas,
1:35:17products, and so on.
1:35:18>> Let me, let me explain. Right now, I am
1:35:20focused on developing a specific
1:35:21product. I have a large product in
1:35:22particular.
1:35:23>> Well, generally speaking, you had
1:35:24previous experience. You
1:35:25>> Go ahead, go ahead. No, I’ll tell you
1:35:27. Wait, don’t rush, don’t rush.
1:35:29I’ll explain everything. Look, here
1:35:31is how it happens. This is, this is,
1:35:32this is the answer to your question. I
1:35:34am currently developing, I am engaged
1:35:36in product development. An example of
1:35:37how we are going to make another
1:35:38product now. Here is how it appeared.
1:35:40We are busy developing this product,
1:35:42right? Along the way, requests are
1:35:44pouring in all the time; everyone asks
1:35:45for something, everyone wants something
1:35:47, everyone is basically pushing
1:35:48hypotheses for us to work on. We, of
1:35:49course, score them in a certain way,
1:35:51validate them preliminarily, and do
1:35:52something with them. That’s not
1:35:54really the point. Then we start
1:35:55researching. In the course of research,
1:35:57it may happen—and it often happens—
1:35:59that we start doing something
1:36:00completely different from what we were
1:36:02asked. That almost always happens. We
1:36:03don't do what we were asked to do. We
1:36:05figure out that the problem is
1:36:06elsewhere and needs to be solved
1:36:08differently. And then we make everyone
1:36:09happy, right? I mean, everything we
1:36:11need grows there. But, for example,
1:36:13just to understand, how can that be?
1:36:17Our platform, well, it—I’ll have to
1:36:19tell you a bunch of domain nonsense,
1:36:21but the essence is simple. Our product
1:36:24is installed in certain types of
1:36:25medical institutions, where, for our
1:36:27product to provide maximum benefit,
1:36:29there must be a few other products.
1:36:31Specific ones, that is, of a different
1:36:33category. So, our product, in symbiosis
1:36:35with them, provides the absolute
1:36:37maximum benefit. Our goal is to ensure
1:36:39that it delivers its maximum benefit.
1:36:41All the power of the benefit embedded
1:36:42in it, all the power of the value it
1:36:44should deliver. That is what we want.
1:36:46But for that, we need to integrate with
1:36:48two neighboring products. Problem
1:36:50number one. These products are not
1:36:51always present in the medical
1:36:53institution. One. Two. These products
1:36:55can be so crappy that, on one hand, we
1:36:57still need to integrate with them, but
1:36:59it leads to a total disaster. That’s
1:37:02how an idea emerges; we go," Why the
1:37:04hell not? Let’s figure out if we want
1:37:07to make this product, and then we start
1:37:09making that very product. "That is how
1:37:11it works. I mean, a product idea can
1:37:12come from anywhere at all. There isn't
1:37:14just one methodology. Well, don’t.
1:37:16There is a methodology for finding
1:37:17products. I mean, in reality, people do
1:37:19it in different ways. You can look at
1:37:21which business models have worked in
1:37:23other markets and launch the same one
1:37:24here. But this is about the approach of
1:37:27taking a business model that works in
1:37:29other markets, locking it in here, and
1:37:30seeing what comes of it. That works too
1:37:34. There are other life hacks for how,
1:37:37like, how to take an idea that is more
1:37:39likely to succeed. But we are here for
1:37:41something else. We are here because we
1:37:43want to learn how to develop products
1:37:44systematically, right? That is,
1:37:46launching businesses systematically,
1:37:47developing products systematically, and
1:37:49adding features systematically. We want
1:37:52a system so that it's systematic, so we
1:37:54aren't just borrowing other people's
1:37:55ideas all the time, trying to test them
1:37:57, we need to start research and find
1:37:59problems during that research. And what
1:38:02is important, why we start this way,
1:38:03and how I did it last time, I already
1:38:05explained. In my previous company, the
1:38:07experience was like this: we were doing
1:38:09research within our current product,
1:38:11stumbled upon a problem we hadn't
1:38:12thought about before, and started
1:38:14building products to solve it. It was
1:38:16always like that. It was always like
1:38:17that, without exception. You're just
1:38:21doing some research and you stumble
1:38:22upon something that, you know, hits you
1:38:24like a flash of light from above, and
1:38:26you start working on it. Sometimes a
1:38:30founder flies in on a broomstick and
1:38:32says," I've come up with a genius idea,
1:38:34"which is okay, but you can research
1:38:36that too, and you might find a real
1:38:38problem there. But it doesn't really
1:38:39matter where it came from. That's why I
1:38:41say take ideas that you personally find
1:38:43significant and solve problems that are
1:38:45meaningful to you, not just random ones
1:38:47. If you want some random ones, I'll
1:38:48give you some. Right. Can I ask another
1:38:52question?
1:38:53>> Of course. If we already have a fairly
1:39:00developed idea, say even with the first
1:39:02sales, can we continue to develop it in
1:39:04this format, just to keep the focus on
1:39:06one thing? How do you feel about that?
1:39:14>> I'm perfectly fine with that. Moreover,
1:39:16I will say, point number one. I
1:39:18probably shouldn't say this, but I will
1:39:20. Nothing will happen if you suddenly
1:39:23stop attending the lectures. It doesn't
1:39:26matter to me at all, you know? I just
1:39:28need you to complete the assignments,
1:39:29okay? We formulate a task, for example,
1:39:31to conduct customer development
1:39:33interviews. You can conduct those
1:39:34interviews for a product that you are
1:39:35already much further along with. That's
1:39:37okay, right? I mean, if you've already
1:39:39gone through these iterations and
1:39:40you're working on a product, just fill
1:39:42it out—you should already have this
1:39:44data. Put it here. If you don't have it
1:39:46, it'll be useful for you, so fill it
1:39:47out—yes, do it from scratch. You
1:39:49don't necessarily have to change to a
1:39:51completely different direction if
1:39:52you're already working on a product. So
1:39:54, the essence is simple. You can just
1:39:58continue developing your current
1:40:00product, put the results you've already
1:40:02achieved here, show them to me, and
1:40:04I'll accept it. That's okay. There.
1:40:11>> Thanks.
1:40:13>> Well, if there’s really something
1:40:15there—I’ll even go so far as to say
1:40:17that...well, never mind, I won’t say
1:40:18it. That everyone will just scatter,
1:40:20who needs that? Nobody needs it. Anyway
1:40:22, uh, so, okay, ladies and gentlemen,
1:40:32no more questions. We've gone over time
1:40:34, I won't keep anyone any longer. If
1:40:36anyone wants to stay and ask something
1:40:38else, feel free to stay. If anyone is
1:40:41ready to head out for the rest of the
1:40:43day, I wouldn't dare keep you. It was
1:40:45nice to meet you. I hope you liked it.
1:40:48Subscribe, rate, and like. Yes, thanks,
1:40:51you too.
1:40:56>> Thanks.
1:40:57>> I have one more question. Is it
1:41:00possible to work in a team of one?
1:41:02>> Yes, I said it's possible. Solo,
1:41:04absolutely. I will.
1:41:06>> That's all, thanks.
1:41:10>> Will there be some mechanism for team
1:41:12division, or?
1:41:12>> Uh, well, I’m counting on you to
1:41:14self-organize. If you need a mechanism,
1:41:16write in the chat. And let me suggest
1:41:18this mechanism. This is what we did
1:41:20last time. You can write and say," Guys
1:41:22, I don't want to do anything myself,
1:41:24take me into a team. "Ready. Ready to
1:41:26join a team. That
1:41:28>> is the best mechanism, just shifting
1:41:30the responsibility.
1:41:32>> Contact Vadim. Shifting the
1:41:34responsibility. Yes. Yes. There is none
1:41:36. I have no mechanism, to be honest. So
1:41:39I suggest you just self-organize into
1:41:41teams. Therefore, yes, just write in
1:41:43the chat, say:" Guys, I'm looking to
1:41:45join a team. "Or let me organize it so
1:41:48you don't look completely lost. I'll
1:41:50just write:" Come on, comrades, whoever
1:41:52already knows which team they're in,
1:41:54sign up here. "Whoever is looking for a
1:41:56team, sign up here. Anyway, let me
1:41:58organize it, don't worry.
1:42:00>> I'll post a site, put your ideas here,
1:42:02and we'll sign up. That's it. Yes,
1:42:04>> yes, that's it. In short, I will take
1:42:06this responsibility on myself. It will
1:42:08be so. You've shamed me, I'm almost
1:42:10feeling ashamed. Pers.
1:42:12>> Oh, yeah.
1:42:15>> Well, at least a little bit. You know,
1:42:17you wanted some hate. So here is a
1:42:19little bit of shaming for you.
1:42:20>> Weak, still weak. Weak, weak. Need to,
1:42:24>> need some more. Need to level up.
1:42:26>> Vadi, is your life so good that you
1:42:27want some negativity in it?
1:42:30>> Uh, no, I'm, you see, I'm a product
1:42:32manager. You can't do anything to me,
1:42:35you know? You it's like, it's like when
1:42:39I finish this course, I'll also be a
1:42:41pro
1:42:41>> No, wait, let me put it this way. I'm a
1:42:44product manager in medtech working with
1:42:46the state, you can't touch me.
1:42:48>> That's one. Two, I'm in another city.
1:42:50>> Two, and in another city. Well, that's
1:42:52not a fact, by the way. We might be in
1:42:54the same city.
1:42:55>> Could we have a white presentation, or
1:42:57will it only be black like this?
1:42:59>> Well, I can make it white, but do you
1:43:01really need to? You don't like black?
1:43:03Well, I just had to hate on something,
1:43:05you know.
1:43:06>> Oh, so that's what it is. Well, that's
1:43:08not hate.
1:43:09>> I like the color, the squares, and the
1:43:11triangles. I want stars. Di,
1:43:14>> can I
1:43:14>> now? I I will answer quickly. I will
1:43:16answer quickly. Tell him he has no
1:43:18taste. All right, now you can go.
1:43:21>> You just mentioned this, but I wanted
1:43:23to clarify. On Monday, there's a slight
1:43:25overlap with the lecture. If anything,
1:43:31will it be possible to watch the
1:43:32recording, will there be any
1:43:34interaction, or if there is something
1:43:36super important that shouldn't be
1:43:37missed? Will the format always be like
1:43:42this, right? Meaning, I'll be talking
1:43:43about something, then pause at some
1:43:45point to ask something, right? So, it's
1:43:46going to be set up something like that.
1:43:48Today we had more of a dialogue, right?
1:43:51Other times it will be less
1:43:52concentrated, just spread out, well,
1:43:54throughout the lecture. But in general,
1:43:56it will be roughly like this. Anyway,
1:43:58well, you can watch the recording, plus
1:44:00if there are any questions, there are
1:44:01always consultations for that. You can
1:44:03miss a lecture, watch the recording,
1:44:05come to a consultation, and ask
1:44:06everything you want. We'll break
1:44:08everything down.
1:44:09>> Okay. Thank you so much. So, there
1:44:11won't be any punishment if I'm not
1:44:13there? No, I, yeah, I said right away
1:44:15that I have no restrictions, like, if
1:44:17you don't show up to the lecture, then
1:44:19it's the end of the world. Zero
1:44:20complaints about that. I understand
1:44:22that some people take two courses at
1:44:24the same time, they overlap, yeah, that
1:44:26happens. It's okay. So, there is no
1:44:29such requirement.
1:44:31>> Thank you so much. That's it, thanks
1:44:33for the lecture. Got it,
1:44:35>> brother. Yes, Pavel.
1:44:38>> Yes, hi. I might have a quick question.
1:44:42If it's too big, just say you can't
1:44:44answer it yet. Basically, I'm a
1:44:49developer, I'm a developer at Sber
1:44:52right now, before that I was at Yandex,
1:44:54and a small group of us started working
1:44:57last year on an idea, creating small,
1:44:59fast, well, AI SaaS projects that we
1:45:02build. And one of my acquaintances, who
1:45:08we are starting this with, has a
1:45:10problem, well, we are doing this as
1:45:12self-founders, from scratch, but the
1:45:14issue is that we are very strong on the
1:45:16development side, but less so on the
1:45:18product side. And so he had this
1:45:21question. How do you know that a
1:45:23hypothesis has failed? For example, he
1:45:27has an idea, he tries to drive traffic,
1:45:29buys ads, it doesn't take off, and he
1:45:32has a product that specifically needs
1:45:34traffic to show that someone needs it,
1:45:36but he can't find the target audience.
1:45:39And so he
1:45:40>> Yes, I, I understand. Let me, let me
1:45:42stop you there. Look, I hope that's
1:45:44okay. You can take, you can take this
1:45:46thing, bring it to a consultation, and
1:45:48we'll run through it there. Why?
1:45:49Because I'm going to start asking
1:45:51questions now. We'll be sitting here
1:45:52for 30 minutes.
1:45:53>> Well, yeah, yeah. Because the domain is
1:45:55important, the core is important, how
1:45:56it launches. There could be a lot of
1:45:58problems there, right? I mean, it's a
1:45:59very long chain of a bunch of things.
1:46:01All of that needs to be analyzed in a
1:46:02consultation. Let's run through it, no
1:46:04problem at all.
1:46:05>> Yes. There. But a quick question. So, a
1:46:07month has passed, let's say, and how do
1:46:09you know it's time to quit and do
1:46:10something else? Well, because you can
1:46:12do things quickly. Or continue and push
1:46:14through. It's just
1:46:16>> uh, it's simple. Either you've run out
1:46:18of hypotheses, you no longer believe
1:46:19anything can be done here, you stop
1:46:21doing it, or you have a pool of
1:46:22grounded hypotheses. You continue to
1:46:24test them. Like, you see, the question
1:46:25is exactly how they were tested.
1:46:27Meaning, maybe the problem isn't the
1:46:29hypothesis, but how it was tested.
1:46:30Maybe it was just done incorrectly. So
1:46:32if you're saying that he's constantly
1:46:34trying something and nothing works,
1:46:35then I hear that as him just getting
1:46:37binary rejections, right? Exactly. I
1:46:39mean, he isn't enriching himself with
1:46:40data, that's strange. Meaning he could
1:46:42have pivoted many times already, so
1:46:44that needs to be analyzed.
1:46:45>> Uh-huh. Yeah, I see. Okay. Thanks.
1:46:49>> Anytime. Are you just going to keep
1:46:54sitting there? It's nice with you, of
1:46:56course, if there's a question.
1:46:57>> One more small question, sure,
1:47:00>> regarding how the course is graded? I
1:47:02think I missed that.
1:47:04>> I won't lie to you, I don't remember.
1:47:06I'll tell you next time. Okay,
1:47:09>> thanks.
1:47:13>> Not that I've been lying all this time,
1:47:15but now I won't.
1:47:17>> Yes, have a good evening.
1:47:18>> Yeah, that's it, thanks. Well, that's
1:47:20all then. Let's everyone
1:47:22>> bye. Have a good evening.

About this transcript

This page contains the full transcript of 1_Продуктовые исследования_07.10.2026 by AI Talent Hub, generated from the public captions YouTube serves with the video. The transcript has 19,705 words across 2,998 segments, with the original timestamps preserved so you can click any line to jump to that moment in the embedded player.

What you can do with it

Use the transcript to take notes, quote the speaker, build a study guide, generate a summary with ChatGPT or Claude via the YouTube Summary tool, or export it as a timed subtitle file with YouTube to SRT. You can also re-open it in the transcriber to choose a caption track available on the video.

Free YouTube transcript tool

YouTube2Text is a free YouTube transcript generator — no signup, no daily limit. Paste any YouTube link and get the full transcript instantly, with timestamps, click-to-jump, caption language selection, AI prompts for ChatGPT, Claude, and Gemini, and exports to TXT, SRT, VTT, or Markdown.