Episode 126

Artem Danilov on Challenges and Dynamics of Hiring in the Crypto Market

TIM: We are live on the Objective Hiring Show. Today I'm joined by Artem. Artem, welcome to the show. Thank you so much for joining us.

ARTEM: Thank you very much, team. Yeah, I'm really excited to be here.

TIM: and I'm excited to have you here, and we're pumped, and we've been teeing this up for a while. And where I love to start with guests is just to hear a little bit more about yourself because it just helps to frame the conversation of who we're speaking to. Who is Artem, and what are you up to?

ARTEM: Yeah. My name is Artem; I'm head of data at p2p.org. P2P.org is one of the largest staking providers on the crypto market. I've been here for the last two and a half years. So previously, I had a pretty big experience of working in different branches, but all of my working experiences are connected with the data. From the data tech positions to the head of data in a bunch of different digital organizations. , I have had a pretty exciting journey for the last two and a half years. With the crypto and this. It was a pretty new branch for me with new challenges and new problems that arose across this journey. Yeah.

TIM: And is there anything about hiring in the crypto market that is somehow inherently different from hiring in any other industry?

ARTEM: Yeah. To be honest, it's a pretty, pretty different hiring experience because usually you don't have a lot of restrictions with the business area that you are working in; you can easily, for example, jump from the e-commerce experience to some other branch and so on. But in crypto, you have. A lot of specific knowledge is needed to be able to make this analysis. And this is a pretty big challenge for the hiring. Because at the same time on the crypto market, there are not a lot of really classical analysts with a big crypto experience, so we have to hire guys from the other branches and spend a lot of time to bring crypto expertise for them.

TIM: And so it's different to, for example, let's say you've been an analyst in, I don't know, banking, and then you make the switch to retail where there's, you change industries, there's always new stuff you have to learn. Is it that it's just a much bigger jump in crypto? And so, like, the learning time is a lot higher or longer.

ARTEM: Yeah. Yeah. Definitely. The amount of knowledge you have to learn to be able to do. Analysis in crypto is pretty high. And the branch, the industry by itself, is pretty unique. We saw a lot of new technologies, and the speed of the appearance of the new technologies in crypto is crazy. And so you have to learn a lot about yourself when you're switching to working in crypto.

TIM: Yeah, for sure. So in AI and data, things seem to be moving so quickly. Large language models seem to be getting updated all the time. The capabilities are improving so much. You're saying in crypto as an industry, the technology is also progressing. So then I guess if you're working in crypto, in analytics, then it must feel like things are just changing every single day.

ARTEM: Yeah. That's true. That's true. And moreover, usually you can't have deep specialization in all the blockchains that exist. Usually you have some specialization, for example, Ethereum or Solana. Because to make the deep research, you have to know how these networks work in detail and what differences they have from the other networks. And this requires time. That was, for example, one of the challenges: why we can't switch easily analysts internally from one blockchain to another blockchain. Because when you are working for the two years in Ethereum, you become a highly qualified Ethereum specialist. You know everything about Ethereum, but it's difficult for you to switch on the Solana. You know how Solana works generally, but not in detail, and that's a problem. And that's a real challenge when we are trying to find someone. The market is because new blockchain blockchains appear every day. New technologies, new main narratives, and initiatives appear every day. And when you have to find someone, you have to, you, you have some small change to find someone with a good blockchain background. But usually this is not impossible. By the way, we didn't have such cases for the last two and a half years. Successful cases.

TIM: Is there also then a case or a discussion around hiring people? who are really inquisitive, who are able to learn new things very well and very easily, because almost anyone you hire is going to have to be able to do that and has to get over quite a substantial kind of learning curve very quickly to become productive.

ARTEM: Yeah, that's one of the major requirements from our side when we are looking for the new data analyst or analytics engineer or somebody else, some other data role, because we expect that the guys will be able to learn the domain. Maybe in three or six months to be able to bring some value. And to be honest, you can't do this without a lot of curiosity, without spending a lot of time outside of work. In learning crypto, we can see that basically the guys who have passion for crypto and who are excited about crypto technologies are 10 times more successful in making this onboarding and learning all this stuff because we had successful cases. With the guys who are really excited about all the new technologies that are appearing in the crypto branch. But at the same time, we had other cases when, for example, we hired data analysts from the classical retail businesses and e-commerce businesses. And they were able to learn the crypto at some level, but they didn't have real interest in it. And on, in the long term, it was a loss for everyone.

TIM: Yes. So then I guess the question is, how do you evaluate that interest in the interview? And that kind of keenness to learn.

ARTEM: That's a difficult question. First, we are trying to look at any crypto background, even the small crypto background. For example, when I have an interview with someone who. Even didn't try to learn what was taken before the meeting for me. It's a bad sign. By the way, it's not the showstopper. And of course we continue our interview and the next cycles in our hiring process, but even if you try to learn something in crypto before the interview. It's a big plus by the way, as well, we are looking on the level of the energy of the of the potential of the candidate because as we can see in highly with the speed of the change of the crypto, you have to have a lot of energy just to be on the cutting edge on what is happening. And it's difficult to describe some. Defined mathematical criteria in hiring, but these informal criteria's level of energy is really important for us.

TIM: And then, so this is something that you try to gauge during the interviews. Do you take a data-driven approach? Do you try to have a scorecard where you measure candidates along certain dimensions, or is it more intuitive? How do you think about interviewing the candidates?

ARTEM: Initially, we didn't have anything, so it was really an intuitional process. By the way, we are trying to be mature, and so we have created the competence card of hard skills and soft skills of the analyst that we are looking for, and we are using it in a bunch of the processes in performance evaluation, in hiring processes, and so on. Even we made it in the way when you can pass the marks to this model. And it says you the grade and the level of the analyst. And yeah. So basically, in the hiring process, we have the data-driven approach. We, all the interviews. evaluating the candidates across all these criteria we have. And so the hiring manager has a final decision he can make. For example, if everybody estimated the candidate as in the red zone, but he's sure that this candidate will be there. The coolest guy, the coolest fit in the team. Usually we made some additional meetings across all the interviewers to discuss this decision. The hiring manager has the last decision in hiring the data guy.

TIM: And so there's like several interviews with different interviewers, and you're saying they're all on the same page looking for the same competencies, but with their own slightly subjective opinion based on what they see in the interview. What about, how do you think about stages? So let's say, I don't know, there's three interviews. Okay. If a candidate does quite poorly in the first one, are they out, or is it almost like you give them three chances and then you take the average or you take the best score? Like, how do you think of scoring across the interviews?

ARTEM: That's a good question. Usually we had the following stages. The HR screening interview, and then we have the first interview with the hiring manager, where we are looking at the basic experience and background of the candidate. And on this stage, we don't formally pass any survey and fill any criteria. It's a decision of the hiring manager if the candidate will receive the home case. The home case is standard for all positions, even for team leads. It's pretty The home case that we are trying to do is connected with the crypto, so you have to learn. Somewhat about the crypto, and this is a real case that happened in the past, and we are trying; we are given the opportunity to research and bring some insights about the case. So it's pretty, it's not a deterministic task, and we are looking mostly at the ways the candidate is thinking and how he's working with a lack of information and knowledge about the crypto. And so it's our turn in the test case interview; usually we are asking questions, and we are trying to determine how the analyst is thinking and what his hard skills are. So after this, we are making the interview with the key stakeholders of the position. And after this, we are making the final decision where we are filling the survey for all the participants of the interviews within the survey about the candidate. And we are making the last decision.

TIM: And once they've done the take-home, sorry, do they come back and present anything in like a case study kind of interview, or is it just the take-home and then it's graded and then a kind of separate interview after that?

ARTEM: Basically, it's a separate interview where the candidate is presenting his home test case and the way he solved it. So what we are usually looking for in this interview is the way the candidate presents information and the insights and what conclusions he is making. The way how he. Try to solve the issue, this research issue, and what approaches did he use? How did he make it? Did he, for example, make any public research on the test case? Did he? What Python libraries did he use, and so on? So we are trying not to limit the candidates. on ways how they can solve the test case. Because usually we have a lot of undeterministic tasks when we have to research something in the crypto that require a quick data extraction of blockchain data or something like this. And that's why we are trying to look for pretty autonomous guys who can solve the cases from the start till the end without any additional

TIM: Is part of the benefit of the follow-up interview also because you can validate that they've actually done it? I guess one of the challenges with any take-home is always maybe they got their friend to help them, or now these days, maybe Chachapiti has been involved, but if you've got them there and you drill down on the details, I think, I guess it would become very obvious if they haven't done it all themselves.

ARTEM: Yeah. That's it. That's it. Basically what we can see is that for the last year, a lot of candidates have been using ChatGPT, and this is not bad by itself. It's a good tool to use in everyday work. By the way, we are trying to check the hard skills. So if the candidate can't explain to us how he made this decision, we can't move further with him. We can really easily understand if he used ChatGPT, if it is his thoughts and code, and yeah, this can be really checked during the conversation. And that's why we have this test case interview. Because we can ask anything on his test case,

TIM: Can you think of any candidates, like over the last year, where they've done the test case, they've come into the interview, and it's been quite obvious that they don't really understand what they've submitted?

ARTEM: I think maybe 10, 15%. Yes, we had such interviews, 10, 15 percent of the overall amount. When the candidate just blindly used ChatGPT or somebody's help. And when we try to get some details, he isn't able to explain, or sometimes we're using the technique when we are giving the candidate some random SQL or Python task just to double-check if he really knows how to write the SQL or Python code. Yeah, sometimes we are doing this, and we can pretty easily understand if he, if the candidate really knows SQL. Or Python or something like this.

TIM: I have been trying to think about recently, with all the stuff that's happening with large language models, what to use them for and what not to use them for. Yeah. And I was wondering if I could share a thought with you, which is that I wonder if the benefit is that they're like a leverage tool where it multiplies from your current skill set. But if you have no skill, let's say you're a zero in a particular skill, zero multiplied by 10 is still zero. So that the real value is making someone who's already a skilled engineer really efficient as opposed to a complete newbie who knows nothing. They're not going to get to that senior engineering level because they won't really be able to scrutinize the output. Is that what it is, or do you see it differently?

ARTEM: Yeah, I'm looking at this totally the same. You have to have some fundamental knowledge of all the hard skills you require to make the analysis. ChatGPT is a good helper. But he can't supplement all of your skills and the efficiency of your work. When you have this fundamental knowledge, maybe 10 times higher without it. And of course you can solve most of the cases with ChatGPT, but you will spend a lot of time, and as well as the results, they can be not really, the results can be not so good as you're making it with the. Good fundamental knowledge of the data. And there are some classes of the task that you consult with charge PT. For example, you can't make a good logical data model. Just with the chat GPT, you can learn from the chat GPT some fundamentals of how to do this, but without real experience, your first, I don't know, 10 logical data models will be bad. And so the experience, your real experience, matters.

TIM: On the candidate side, the candidates who are applying for these crypto roles, are there any specific challenges that they face, any way they should prepare themselves when applying for these roles? Like, for example, you've mentioned that the barrier to learning, like the knowledge gap, is going to be a little bit high in crypto. Things are moving quickly, and that is because that success seems to correlate with genuine interest in the field. Yeah. Any other challenges that candidates are facing? Do you think?

ARTEM: I think mostly only the high level of interest is due to the requirement of specific knowledge of the crypto branch. By the way, as well, we can observe that a lot of crypto startups and companies are quick and dirty companies. Everybody is running, and they're trying to get value as soon as possible. And this is normal for the current maturity of the crypto market. By the way, this means that. If you want to work with the data on the crypto market, you face a huge amount of challenges, and what we're especially looking for in our interviews is the way a candidate can quickly fetch some crypto data, because basically, you have maybe, for example, you're trying to fetch some Ethereum transactions, and you have maybe 15 ways how you can do this. In real life and what we are really like, like when the candidate is able to find different ways of getting the same data, because usually just using the one source of data. will force you to have some data quality and data governance issues. So the good data analysts can quickly encrypt, quickly extract data from the various data sources, compare it, clean it, and use it. And this is a common problem in the crypto market.

TIM: What about when it comes to using large language models or any kind of AI in the hiring process itself? Is that something? You've started to double down on the hiring side.

ARTEM: We are not using any LLM models on the current site right now, maybe when we are creating some job descriptions or some good descriptions of the test case, but basically I don't see the ways how. They can help us by the way, and what I really like is that after the widespread introduction of LLM models across the market, the overall. The way of the description of the test cases and description and conclusions increased in all our test cases. So we can see that usually candidates double-check themselves in, for example, ChatGPT or many other LM models. And I really like that level. The language and conclusions are really higher right now than previously. But it's not on the hard insight. It's on the candidate's side.

TIM: And let me propose then a few potential applications just to get your thoughts on this. So I guess there's like the resume screening stage currently done by a human, let's say someone in talent acquisition, maybe also someone in analytics, and having, let's say, a more consistent or scalable way to do that. It'd be measuring, like, a match score between the resume and the job description scale. Would that process with something like that be valuable, or do you feel like it would be better off doing it with humans?

ARTEM: That's a really interesting question. I'm not right now. I'm not doing a lot of research work for the new candidates. Mostly it's our recruitment team that is making these, and I know that they're using a bunch of the LLM models, but basically these models are embedded in the. Recruitment tools that they're using. So it's not the separate tools; it's just the current tools that they're using to track all the CVs and tackle the hiring process. But then what I can see is that it seems to me that the amount of open positions on the market is decreasing in most of the markets and most of the branches. The number of the candidates is growing. And right now we are facing the issue when For the candidate, it's really hard to reach the recruitment specialist, because when you're trying to apply on the website or LinkedIn or any other source, it's really hard to make sure that your CV will be really read. Specialist, and what I have noticed as well, maybe for the last six months, is that a lot of candidates are trying another approach, which is to increase their chances to get the job. So they are trying to reach the hiring managers directly or indirectly via friends, and they are making some strategies. For example, if you want to work in crypto, you have the list of the companies you would like to apply to, and you are trying to find ways to really reach some life men in these companies just to share your CV with them. And this is really interesting. I'm pretty sure that even 10 years ago, it was a really good way to find work. But nowadays, when the other ways of just applying on the website or linking it stop working. That's a really good strategy to find the job.

TIM: Yeah, I feel like if I were to apply for a job now and I was looking for a job, I would probably not go to the job boards. Because, as you say, there's just such a huge volume of other candidates, and it's so hard at that early stage because you've got maybe, I don't know, 10 seconds to impress someone with a resume. It's so easy to fall in with the noise of everyone else. I would certainly be trying to use my networks. What about you? If you were suddenly going for a job in a month, would you still be applying on LinkedIn, or would you use your networks, or how would you approach it?

ARTEM: I think that I will try to use my network. And direct connections with somebody in the companies I would like to apply. So the way I would like to find the job is to make the list of the companies. I really like to work in, to work with. Then I will try to find the strategies for each point in the list. How can I reach somebody here? How can I efficiently apply to these companies? Because yeah, job boards don't work right now. Yeah,

TIM: Yeah, I think any candidate working in analytics who's also got some experience in sales will have a much easier job at getting a job because most early-stage business development is doing exactly that: coming up with the target list of accounts, figuring out who you want to reach out to, crafting a LinkedIn message, and doing a phone call. And so it's just that process, but applied to getting a job. And so yeah, if anyone has some experience in sales, I think it'll really help them.

ARTEM: That's right. And I think that the power of the personal brand is increasing each year. I saw a bunch of the good and bad cases when the candidates were investing a lot of time in their personal brands, and some candidates were really good candidates. So the market was trying to hire them. They. Something like the stars of the market, by the way, we had at the same time bad cases when the guys just were, when the guys just were pushing their personal brand without real investment in their hard skills and soft skills. And just personal branding doesn't work as well. So you have to have the balance of the soft skills, hard skills, and the ability to. To reach somebody and to impress somebody.

TIM: Of those skills you've laid out, those kinds of segments of skills. How do you think about yourself along those segments? Is there one that perhaps you'd be focusing on that you will focus on yourself in the next year?

ARTEM: Yeah, I think definitely. Yes. And I have personal experience when I invested a lot of time previously in conferences, participation, and making public speeches, and this really boosted my career, to be honest. And I think that we shouldn't underestimate this way of building a personal brand. But as I initially invested mostly in the Russian labor market and my brand was created there, when I switched it to the worldwide companies, it became really hard to work on this, on the personal brand. And yeah, I would like to boost my skills of building a personal brand and making some community around me and around, I don't know, crypto data or any other topics.

TIM: I have one question for you, and if you don't feel comfortable answering it, it's completely fine. But I was interested; your Russian background, is that correct?

ARTEM: Yeah. Yeah.

TIM: Have you faced, do you feel like, have you faced any discrimination, do you think, like any kind of bias when going for roles with non-Russian companies?

ARTEM: Definitely not, to be honest. I didn't face it at all. And I think that in the data market, the information technologies market, it's difficult to find. If you're a good specialist, then everything, everything else doesn't matter. And yeah, so I have a pretty good Russian immigrant community. Who are working across Europe, the USA, and other markets.

TIM: Yep. Okay. That's good to hear. I'm relieved to hear that. What about when it comes to evaluating candidates? So we've spoken a little bit about their technical skills and also about just finding candidates who are really motivated and interested in crypto because that's clearly going to affect how quickly they can learn. How do you think then about cultural fit? How important is that? And is it ever at odds with the concept of diversity of thought or diversity of mindset? I sometimes feel like they're slightly contradictory, but I'd love to hear your thoughts.

ARTEM: That's a really good question. I'm totally sure that the more different people you have in the team, the stronger the team is. When we are trying to look. On the team feed, basically we are looking at the two points. The first one is how the person fits himself in the data team, how he enriches it, and if he rises above the team or not. And I think this is the most important question: Is the candidate raising the bar in any particular skill, or is he boosting the team? And the second one is the fit with the key stakeholders, because it's the most important part if you want the communication between the stakeholder and the data specialist. If you want to make it successful, you would like to have this connection, this feed from the first day. Of the candidates, that stakeholder is there, and there are a bunch of cases when the data analyst is changing his stakeholder, and for him, it's difficult to work with a new stakeholder. And, of course, in an ideal situation, we would like to hire the people who can work with anybody, but usually it's not the case. And one of the things that I have noticed hiring in crypto is that it's connected with the passion for crypto. I spoke to her earlier, and what I've mentioned is that usually the most results in energy. We have from the guys who are slightly punks. I don't know why, but in crypto this works because crypto is right now the branch of all the punks. The more you're a punk, the more you're successful in crypto. And this is punk like crypto punks. The guy who is not conventional, and yeah, so I mean that It's not easy to work with the real true CryptoPunks, but it's good to have a couple of such stars on the team because usually they can bring a lot of insights, but we are trying to balance the team. To have, for example, mature guys with the proper soft skills to have the guys with the strong hard skills, so this balance creates a real, really good fit and makes the team stronger.

TIM: I liked how you framed it in terms of finding people who are going to be better along some dimension versus what you already have. And I was thinking about the same thing recently, doing recruitment for my football team, my local amateur football team. And like the number one thing I thought about this season was they just, they have to be better than our current players. Like any new person who comes in has to be better in some way—faster, more skillful, or more committed, or whatever. And. Yeah, I feel like if I think back to our own hiring and our business, maybe sometimes we didn't always follow that, what feels like an obvious rule, because if you keep doing that, and I guess the average keeps going up, which is fantastic.

ARTEM: To be honest, I think this works for any level. of the position, even for the junior specialist, for example, or interns, because even if you're a junior specialist or the intern, usually the hiring manager will ask, Okay, how are you raising the bar across other juniors or across all the team? And if you can bring something special to the team, that's always good.

TIM: One person that I interviewed said to me when it came to this diversity of thought. That he said, Yeah, like it definitely leads to better outcomes. Like you, you arrive at a better place because you've had this wider view of things. But he said, maybe a bit slower, though, because there's a bit more debate and you're not just all thinking the same way. So it becomes a bit slower. Have you noticed that there's like maybe a bit of a trade-off in terms of speed versus quality of the end decision when you have a broad set of views?

ARTEM: I can agree that's usually the trade-off. The quicker you are at making something, the less time you spend on the conclusions and insights, the less quality you have overall. By the way, it really depends on the branches you work with. In crypto, you can't spend half a year on the data research. It's just not acceptable. And so you have to be quick. And in other branches, it works in another way. And I think that, for example, there are a bunch of the fundamental researchers in crypto that usually are performed via grants from some foundations of the networks. And for them, it's okay. To make the research for half a year, but for the internal researchers, you want your stakeholders to be happy. And so usually you have to find some balance. In terms of the time and the quality,

TIM: And you'd mentioned, yeah, getting people who get along with their stakeholders and that being important. And so that's assessed in a stakeholder interview. Is that the main way that you evaluate that in the hiring process?

ARTEM: Basically, yes, basically, yes, our last stage for the interviews is the stakeholder interview. And usually we have two aims for this interview. The first one is to check the feet of the stakeholder and data specialist. And the second is to make the final. Selling the position for the candidate because usually this is the most positive interview because we know that this is a final interview. We are ready to make an offer, and so is everybody else. Everybody is trying to be the best at these interviews.

TIM: Do you have any thoughts around hiring quality that is Like, how, if you were defining this or if you do define it or you try to measure it, how do you think about it? How should we be thinking about it? Because this is something I've been asking lots of people, there doesn't really seem to be a consistent definition. Some companies, or most companies, don't even measure it at all. And so I'd love to get your thoughts on hiring quality.

ARTEM: That's an interesting question, to be honest. We are not systematically measuring the hiring quality; of course, we are measuring the number of overall employees that left us due to the bad fit or the bad hiring quality. But it's pretty low because usually you spend a lot of time on the hiring. It usually takes at least two months to find really good specialists. And so you are mostly interested not to make the mistake. So the way that I'm thinking about it, it is really important to build this defense on the entrance. To make the interviewing process, to spend more time on the hiring, to make the interviewing process more complex, to bring the data-driven approaches in the estimation of the candidates, and then to solve the problem when you hired the wrong guy. And that's why, for example, we had a lot of. Thoughts about how we should build the hard skills check of the candidates. Should it be a home case? Should it be a live coding interview? And we decided that we would have the life, small life coding part in our test case interview because it takes a lot of time, but it's really important because you really need to be sure that the. The candidate has proper heart and soft skills. And right now we are only decreasing the quality of the hiring process, the requirements for the hiring process, only if we are in a real rush. When we need to hire someone in one month, something like this.

TIM: Yeah, I think it's so difficult in hiring because, of course, you want to hire the person as soon as possible. And I don't know about you, but. Whenever I've done a lot of hiring, I get almost like a hiring fatigue, where I think, Oh my God, can we just hire this person? Like we've been interviewing all these candidates, let's just make an offer to someone who feels good enough. But I think that's such a bad trap to fall into. Because as soon as you've hired a few regretted hires, bad hires, or whatever you like, it's way worse than just waiting an extra week to hire someone, unimaginably worse for everyone, the candidate first. If they end up having to get fired after three months or something, that's like a disaster for them. And then, unless you're a very strange character, I think most managers don't really like having very difficult, negative conversations, like every day with someone, like that's just stressful. And I think, on balance, it probably is worth waiting or having an extra step just to be as sure as you possibly can, I think.

ARTEM: Yeah. I think it's just the experience of the hiring manager because. The hiring without strict requirements for working in the short term, you can make it a couple of times, but in the long term, it's not working because you have properly balanced the team. You. Should have the proper feet of every employee of the team. You should the team should be complimentary. Each new employee should add some value to the overall team. And this requires time and. It's good to make a couple of such mistakes because then you will definitely understand how to balance the requirements on the entries for the candidates and the overall hiring speed because usually the hiring and recruitment team is pushing hiring candidates to speed up everything. Just to be able to hire more guys in a shorter amount of time. But it's really important to push back the recruitment team in such cases.

TIM: Yeah, and I guess then, at least in theory, it would be the benefit of having some kind of clear quality of hire metric that maybe is almost like a shared metric, because I can imagine, yeah, most talent teams, probably their main KPI would be time to hire or hires made, like bums on seats, we call it in Australia. And rarely, from what I've seen, would they be motivated by quality of hire. And the pain and misery of a bad hire almost always fall on the hiring manager and the team, not really the talent team. And so having some kind of clear metric, I don't know what that would look like exactly, but a first pass would be, I don't know, average performance review rating, or did they pass their probation? Like, I could imagine a few basic metrics like that would at least be a good place to start.

ARTEM: Yes, maybe. I think it's really difficult to define the metric of the quality of the hiring. We are usually only looking. If the candidates work at least one year at the company, because if you joined and lived earlier than one year, this looks like we fucked up somewhere.

TIM: Actually, I should share one insight I got. So I did speak to someone yesterday who does hiring at a ridiculous hyper scale in Europe. And they, the way they started to measure the quality of hire, was they did like an NPS survey to the hiring manager at like day 30, day 60, day 90. It was very simple. It's just, would you hire this person again if you knew what you knew? So a yes, no, and average that out across thousands of placements. And I found that interesting because they're getting like an early indicator. They're not waiting for the six-month performance review or probation. They're getting this kind of just little, simple metric early on. And then they start to notice an average decline, and that's a conversation for them to then have with that hiring manager about their team and then readjusting the process as a result. So I found that an interesting way to do it.

ARTEM: Yeah, it's really a good approach, to be honest. I think I'm sure this metric should work because we have, for example, for the first three months of the candidate, we have the recruitment, not the recruitment; the people manage the screening of the new employees. It's not the direct manager; it's the HR team doing this survey each month of the first three months of the employees. By the way, I think if we can add the question for the hiring managers in these surveys, it would be really great. Because yeah, it can; it's proxymetric, and it can somehow highlight if the overall hardening process is working well or not. It should be the balance of this metric versus the metric of the speed of the hardening. Yeah, really good.

TIM: Artem, it's been a really interesting conversation today. We've covered a lot of different ground. I've really appreciated hearing your insights, and I'm sure our audience has as well. Thank you so much for joining us.

ARTEM: Thank you very much, team, for the opportunity to discuss all these topics. This was really a pleasure for me.