Episode 49

Nicolaas Wagenaar on The Value of Networking and Experiential Insights in Data Hiring at Startups

TIM: Nicolaas Welcome to the Alooba Objective Hiring Show. Thank you so much for joining us.

NICOLAAS: Thanks, Tim. I'm very happy to be here.

TIM: We're delighted that you're with us, and the first area I'd love to ask you about is what it was like to be the first hire into Growblox because I think that's a really unique position that I've never been in myself, and I think a lot of our listeners wouldn't have been in even if they joined an early-stage company. It's very rare that someone is the first person, and so I'd love to hear more about that experience, particularly from the hiring side of things.

NICOLAAS: Yeah, of course, it's a very interesting experience, one that I kind of discovered along the way and definitely did not know the full extent of as I jumped into it, but I kind of gradually ended up in that position, so the founders of Grow Blocks—I knew them quite well from previous jobs. I've worked with them very closely, and I knew that some of them were out of work or at least looking for new opportunities, and suddenly they started having a chat with me and said, Hey, I have this idea. I think I'm going to build this kind of product, and they just asked me for some advice on the tech side of things. Like, how would you build this? Is this even feasible to do? I was very keen on thinking about that and helping. What helped in this case for my interest was that it was very close to what I've been doing my whole career but with a slightly different twist of productizing it instead of building data assets for internal use cases. So I was just very intrigued and talking about it; probably they already had a hidden agenda of getting me excited about this and kind of luring me in deeper, but we did have a few conversations when I started realizing, Hey, this, you know, I think they might want to get me involved here, and it was really just literally a wine-and-dine experience. Eventually they took me out to dinner. We had a good chat. We talked about all the possibilities of what this could be, and then I think it was either that same night or a different day they came to me with a proper offer and then allowed me to join. I didn't really hesitate too much; the biggest hesitation came from the fact that I had recently started a new job. So it felt really bad giving that up so early on before I even had a chance to properly make my mark there. That was the biggest trade-off, but ultimately, once I made that decision, I was just very excited to join a company that early stage and see what it's like to actually try to work towards getting a company off the ground when nothing's already there.

TIM: Yeah, I personally find those beginning stages so exciting because you've got, like, an idea in your head, it's nothing, and then somehow, eventually, through a lot of people's hard work, you then have a product and customers and feedback and data, and it's amazing to see and experience that feeling of creating something out of nothing. I'm interested in what you were hired for originally and did your role develop through time, and yeah, how has it evolved?

NICOLAAS: I would say originally it was building up just the very first pieces of tech that would ultimately be part of the product, so you could say I was the first tech hire, even though my profile isn't traditionally that of a software engineer. I have some of those skills, but there are many, many people out there that are much better at that than I am. I think why they approached me specifically is because I had the experience of working with building revenue models, which is what product is all about—productizing the building of revenue models—and I have done that similar to the founders at a small scale within a single organization, and they knew that if we're trying to productize this, we need at least those skills just to know how to build a revenue model. and then together we'll figure out that other part and how to actually productize this and scale this up so the first kind of iteration of our tech stack was heavily based on what you would do for a single company, and slowly we looked at different ways you could run this in a multi-tenant environment. and scale that up to actually handle multiple customers

TIM: So you were effectively the domain expert as well as offering the tech skills. Is that a fair summary?

NICOLAAS: I would say yes.

TIM: And so you've been productized. Just maybe a little bit more background for people who might not be in the know around grow blocks, so you've been productizing revenue models to something you might have done initially in what an Excel spreadsheet or Google sheet or something but creating a fully fledged bit of technology out of that.

NICOLAAS: Yeah, yeah, I think it's a good summary. The Excel spreadsheet, I think, would typically be the kind of the final stage of the process. In many cases, there is a big kind of data pipeline and the data team that sits in front of it to make sure that you first get accurate data and helps work with the operational team to make sure that the processes are are working well so you produce good data, then make it ready, then you give that over to the people building the revenue model, but you typically need Excel as the final piece because BI tools would fall short in the forecasting of data and the causal models that you typically want to build to be able to look at the future, not just at the past.

TIM: I'm interested also in how you arrived at this role because I feel like at the moment, in the current hiring market, we'll get into this a little bit later if you're fighting it out with everyone else on the job boards, you know, applying for jobs. You're probably seeing the LinkedIn post where there are a thousand other applicants that you're competing against. Like, it feels like the job market for the average person right now is a bit brutal; it's also very difficult for hiring companies because they're struggling to deal with these volumes as well. But what I feel like is worth discussing is the other market, which is leveraging your networks and knowing people, being able to get a foot in the door through another way, because I think if you had the ability to do that, I would recommend doing that, especially in this kind of climate, so it'd be interesting to unpack a little bit more there how you got the role in the first place. You said you'd worked with the founders before. You'd obviously done a good job laying the groundwork maybe for this opportunity for a long time before that, so I'd like to hear a bit more about the history of that.

NICOLAAS: Yeah, I think it's a very good point you bring up there because indeed I personally have found that the value of my network has been the most valuable asset in trying to land the jobs that I've had in my career. I think only the very first job I had was maybe not from my network, but all the others eventually somehow the network was involved in me landing that opportunity. So yeah, very much the same here. I don't think I can pinpoint any particular things I've done along the way to get me there. I think it's the culmination of working with various people along, you know, for various years in various different companies and apparently having a good experience doing so. I think that that's really most I can say about this. It's definitely a mutual thing as well, right? It's not just, you know, they were liking me and my profile. I thoroughly enjoyed working with them, so that's also why when they initially reached out and said, Hey, do you want to just help us think through this without any job opportunity involved? I was like, Sure, absolutely right. This sounds really interesting. It sounds really cool, so yeah, again, I'm not sure if this is the best answer that you're looking for, but I feel that the best thing you can do if you want to leverage the value of your network and find new opportunities is just making sure that you keep those relationships warm. So even if you leave a job, it doesn't mean you never have to talk to these people, and I think that's ultimately because I hadn't worked with a bunch of these people in a while, but we still talk quite frequently just to talk about all the different things that they're working on that I'm working on. just interesting topics related to our work

TIM: As you were describing that experience, it made me think of a few things. one was that I feel like particularly if you're a junior candidate, let's say you're just getting into the workforce for the first time, if you're a graduate, maybe you're about to graduate, and you hear the word networking, you probably get the wrong idea about it, and you probably view it as this Oh, I'm going to go to this event, and I'm going to hand out a bunch of business cards, and this has got like an almost cringy transactional feel to it. Whereas really, if you turn up to work every day, be a decent human, be very competent, and be interesting and engaged on a consistent basis, then that's the value of the relationships you build that eventually might turn into some opportunity that you have no idea how they will or won't. It's almost maybe it's almost going to the gym. It's like eating well—just those micro habits that, in the short run, are going to do nothing but, in the long run, are incredibly valuable. I wonder if that's a lens that we should view networking through.

NICOLAAS: Yeah, I think absolutely, because, as you said, it's not one thing. You cannot just wake up one day and say, Now we're going to network, and that's going to yield me an opportunity, you know, two weeks from now. It is, as you said, those kinds of micro habits. It's something you nurture over a period of time. That's where people who are more experienced in their careers have richer networks than the people who are very junior and just come right out of high school or university, so it is absolutely the right answer to look at it.

TIM: And I think also, yeah, again, if the more junior candidates who might not have had that experience yet think of this more openly, so I'll give you a quick example. So I spoke to someone recently on this show who told me Oh, isn't it interesting how the more senior you get, the shorter the hiring processes and the simpler it is, at least in his experience? and he said he basically hadn't had a job interview for 12 years; like the last two or three roles he'd gotten were through someone in his network who either suggested how you should hire him or somebody met with him for a coffee at McDonald's, and then they gave him a job, and I feel like if you were a junior candidate with that much experience, you might interpret that in the wrong way. The takeaway for them might be Oh, this is unfair; he's gotten a foot in the door through some kind of cronyism, but I feel like that's the wrong takeaway because he's really good at his job. He's been working out for 25 years; he took time to build relationships, and eventually they're bearing fruit for him. and I'm sure he's delivering value now, so this is just, yeah, again, some maybe like a reframing that I think people might need to tap in their heads around this.

NICOLAAS: Yeah, it's an interesting topic, right? Because when I get the other perspective, I get the perspective that it might not be fair, and even from the company's perspective for hiring, it might not actually be their best move to just go with the first person that you know through your network because maybe there are other candidates out there that would actually end up being better. and similarly, if you turn it around, if you are in the market for a new job, going with your network is very comfortable, but it's not But maybe there are actually other opportunities out there that could be better for you but that then allow or require you to have to go through the ringer of the traditional hiring process. but that doesn't mean that it's not valuable to either of those parties, right? I think there's also an argument to be made that there's value in not having to go through many different hiring processes. Maybe you don't look for the perfect opportunity, but this is a very good opportunity that's right in front of you. and you know it'd be an easy start because you know the people that you're going to be working with have the exact same mindset for the company, right? I think you just have to balance that. You need to figure out what it is that we're actually looking for right now. Are we looking for the perfect candidate? Are we looking for the applicant? Are you looking for the perfect job? Are we looking to be extremely objective in this and trying to give people a fair shot, or do we just need someone in this role right now because otherwise our business is going to suffer? You have all these perspectives to keep in mind. and obviously that leads to a lot of different hiring practices and a lot of different ways you can either find candidates or land jobs.

TIM: Yeah, you're right. Hiring is such a game of trade-offs, isn't it? You can go fast and dirty; you can go slow and a bit cleaner maybe, and everything in between.

NICOLAAS: Yeah, absolutely, yeah.

TIM: What about then at Growblocks when you were first building out your team? How did you approach it?

NICOLAAS: So this, in the beginning, was very difficult, I think, for two reasons. One, we're a startup that absolutely no one knows about, so trying to convince people to quit their job and thus work for something that doesn't actually exist yet or barely exists is a challenge in and of itself. but on top of that, it was also kind of late-stage COVID, and the hiring market was very, very hot, with it being extremely difficult to find candidates at the time, so we did a lot of cold outreach in those early days. We just, you know, we knew that we tried, and kind of we were—we got the confirmation that if you just put out a job ad, that nobody would come to us because there wasn't anything inherently enticing by the store. We could tell through a job ad, so we had to sell people, so I wrote a ton of messages on LinkedIn. It was the only few months in my life where I actually had LinkedIn Pro to be able to send that many messages and to do that many searches. But it did work. It did work. It's kind of grunt work. I think most people realize and know, but some people will reply, and with those people, you can start to have a conversation, and then you have your pool of applicants. It's a different relationship, though, in that it's more about you convincing them that this is good for them instead of them convincing you that they're the right candidate for the job.

TIM: Yeah, inbound and outbound are so different, aren't they? Because you're asking them to join you rather than them applying for a job that's open, I guess almost all of the time the people you were approaching would already have been in a job, probably one that they liked, and so that's already quite a bar. I imagine, though, the ones who replied may be the ones who were already contemplating a move or were actively searching. Did you find that

NICOLAAS: Yeah, at least I think they were at least contemplating something new. I've never actually asked that question specifically, so I can't give you a conclusive answer, but if I go with kind of, you know, my perception of the people we hired at the time, they were people that were maybe a little bit stuck in their current company considering doing something else. or they'd have a decent job but didn't see a lot of growth opportunities, and just something about us reaching out piqued enough of their interest to be like, Hey, maybe this change is good for me. I didn't think that that at least allowed us to have a conversation with them.

TIM: And these people would have been what, within the first 10 employees in the company, I'm assuming?

NICOLAAS: Yeah, yeah.

TIM: Yes, yes, I have hired people at a similar scale of company. I feel like there's a certain character like this; it would be the kind of job that would be dreadful for some people and someone else's dream job. and maybe the segment for whom it would be their dream job is small; if you could somehow identify those maniacs who want to join a company where the product barely exists, then they are like the pioneers—the kind of pioneering spirit that you want. I find, was there anything particular in the screening or the hiring process where you were looking for that kind of startupy, and this is a real startup? This isn't a thousand-person startup; this is less than 10 people. That's a certain type of character that you need to not drown and to actually thrive. How did you identify that in the interview process or in the hiring process in general?

NICOLAAS: I think I'm trying to just kind of recollect my thoughts on this, but we definitely did almost say a subconscious prescreening of people working in startups and scale-ups, so in my searches I wouldn't really go to people working in large corporations, as that would just have been too large of a mountain to climb. There are probably people out there who want to join a startup, but you know from the thousands or tens of thousands of people who work at corporations that it's going to be hard to find so pre-screening is looking for the scale-ups, and we also were hiring at that point; we were hiring local, so scale-ups in and around Copenhagen. What, in my particular case, helped as well is that the data technologies that we work with were mostly adopted by startups and scale-ups, so just looking for people who have those skills within 98 percent of the cases led you to people who have been working in startups and scale-ups, so that was a really good way for me to kind of zone in on a group of people that could be of interest. and then you take it further from there

TIM: And when you were building up this team, what worked, what didn't work? Do you remember what the main overall challenges were?

NICOLAAS: Yeah, it's a tough one. I'd say I don't think I could really pinpoint one or a few things that worked and didn't work. I noticed that it was a bit of a hit and miss, so the approach we took was just kind of zoning in on this group of people, and for some people it was you get an instant reply and they would be very keen to talk, and for others you just hear nothing from them at any point in time. Trying to figure out if there was anything that we did in kind of in our messaging to people, but no, to go on some blanking a little bit, I don't think I can give pinpoints of very specific things that worked or didn't work in this approach.

TIM: And what about, let's say, more broadly then, in your hiring experience for data roles? Have you experienced the pain point of a candidate who, on paper, looks amazing? Their CV seems to have the right skills, seems to have the right experience, but then you get into that first interview and you think, How is this the same person? I don't even understand the gap here; is that part of your experience, either in building out this team or in other roles?

NICOLAAS: Yeah, it's a notoriously difficult one to navigate because it's very easy to put a skill on paper, right? That takes basically no effort at all; you just need to have the confidence to put it there and be willing to talk about it, but this is where skills assessments can help out a lot. I don't think they're silver bullets, but they can definitely help you kind of weed out the ones that are just blatantly lying; that helps a lot. I don't think I've ever gotten in a place where we ended up in an interview with someone who just completely lied on the resume or had exaggerated so much that it was just so far away from reality. I think that's also because, apart from GrowBlocks early stage and the other companies I've worked for, there was someone in talent acquisition internally that would always do a prescreening, and they might ask a couple of problem questions just to see what experience is there, and then by the time I'd have a call with them, I know at least, you know, what skills they roughly have. and then you can I couldn't ask them deeper questions just to assess the level or the extent of the skills that they have.

TIM: You mentioned talent and HR there; yeah, often coming in is that initial phone screen step, which is the way most companies will probably set it up. Some feedback we receive is that a lot of hiring managers feel like that process is not particularly accurate in the sense that they'll end up still interviewing candidates who they feel like they are wasting their time with a little bit. and I feel like part of the challenge is that for recruiters and talent acquisition, they are not really domain experts, and we're asking them to do screening for a data analyst or a software engineer or a performance market or some role that they themselves have no real experience in. It must be very challenging or impossible. To really be an effective screener for a role that you yourself haven't done—that's my view. I feel like there's a kind of a fatal flaw in the process. How do you see it?

NICOLAAS: There's definitely that risk, yeah, a hundred percent. The way I've seen it mitigated is by having recruiters internally that focus on the various areas of the business, so, for example, at Falcon.io, now Brandwatch, we had a recruiting team that would have some of them specifically focused on hiring for sales and others specifically focused on hiring for product. and I'll be specifically focused on hiring for customer success, et cetera, and I can absolutely see the value there. A recruiter might not have software engineering experience, but if they have been working with hiring people in software engineering long enough through feedback from the hiring managers and the other people you work with, the candidates, as well, of course, You develop a skill in understanding what a good candidate will be. Again, it's not that you don't get a perfect understanding of how a software engineer works and things, but I do think you can get to a good enough point where you can do a screening call that will weed out kind of the initial bunch of candidates that would really just not be a good fit for the role or have exaggerated their resumes too much just so that the process can continue and that the hiring manager basically has a smaller pool of candidates to work with.

TIM: So there's some learning through specialization, and yeah, if I think now about the way a lot of hiring managers are described when they've worked successfully with their talent team, it's this kind of back-and-forth, almost iterative process, like building a machine learning model actually is almost like this training data. You presented these CVs. Here's why I don't think they're relevant: back and forth, including then on the interview step, and okay, you mentioned you evaluated them as X in the screening call, but I evaluated them as Y. Here's why, and they iterate through it that way, which, yeah, I guess after a while you develop a bit of an intuition for it, even if you didn't necessarily understand the details of what they're doing. so that makes

NICOLAAS: Yeah. Yeah, and I've also seen it work in a way where the hiring manager might actually look at all the candidates coming in and all the resumes and all of the questions answered in the application, and they would actually make kind of an initial list of people they think are interesting or not interesting at all. But then the talent acquisition specialist would still actually go in and do an initial screening call just to save the hiring manager the time of not having to make, you know, 20-30 minutes of screening calls.

TIM: That makes sense just to take a bit of the load off of them. What about from your perspective? So when you were building out this team at Grow Blocks or in other roles, did you have a view over the soft skills versus technical skills? Do you feel like one is more coachable than the other? like how do you prioritize those

NICOLAAS: Yeah, so this has actually been an eternal balance or challenge, you could almost say, in the roles I've been hiring for. I'm not sure if it's specific to data or if it's just my perception of what I think a good data person should be, but I found that when you hire for data roles, unless it's a very diehard data engineering role, these people will be fed JIRA or linear tickets, and they just do the work, and they deliver a product. Unless it's that data, people will always have to interact with other folks across the organization, so you need to find this right balance of technical skills to be able to produce data assets and soft skills so that ideally you don't have to hold their hand all the time when they do need to talk to a VP of sales or a marketing manager. So when you look for candidates, I always find it a tough balance because some candidates are really good and strong on the soft skill side. They might come from a commercial background; maybe they've come from a consultancy, so they clearly have an analytical mind; they can deal with stakeholders. They're organized, but they just have very limited tech skills, maybe just Excel, sometimes VBA, or a little bit of SQL on top, and you have the other ones, right? The people that actually come out of data engineering are completely strong in the technical side, but you have great doubts about whether they're going to be able to actually have a conversation with the sales manager. and that I've mostly done this on a case-by-case basis to have making that assessment. I don't have one system where I assign weights or values to certain parts of their profile, and then out comes a number, and I go by that. It has always been based on the candidates that I had available to me for that particular hire. and then also obviously the requirements for the role, and then I had to make the assessment: Am I going to go for someone who's a bit stronger on the technical side, and do I believe they are able to develop those soft skills, or am I going to go for the one who is stronger on the soft skill side and will teach in the tech skills along the way? What I can say just to round this off is that I found it easier for people to develop technical skills than for them to develop soft skills, so I've typically erred on the side of hiring people with the stronger soft skills and then helping them learn these technical skills along the way than the other way around. There might be an interesting one to kind of peel back a little bit more on why that is, but if I think back on it, that's typically the direction I follow.

TIM: And when I've asked this to other hiring managers, I'd say 80 percent would say the same thing as you, which is that the technical skills are easier to teach than the soft skills. I wonder if it's worth digging into a slight caveat, which is I assume you would say that on the basis of a candidate already having a baseline of skills. So, for example, you mentioned a candidate who knew Excel and VBA; then, for me, if I looked at them, I'd go, VBA is harder than SQL. I'm sure they could learn SQL if they already knew VBA, but if they knew no programming language at all and they were the sort of person whose eyes glossed over when you showed them some data, I would probably be quite dubious they could learn SQL. I'm like This is just not your thing at all, so I assume it's like on either side they already have to have a baseline set of skills, and yet the soft or technical is for you to believe that they could then upskill. Is that a fair comment?

NICOLAAS: Yeah, yeah, 100%. Yeah, it's I. I would never be in either, or I don't think I would ever hire someone in a data role that doesn't have any experience from a technical level working with data. I would probably even go as far as to say, of course, depending on the role, but let's say you're hiring an analytics engineer. So they'll be working with SQL DBT a lot. I probably wouldn't hire someone who only has Excel skills; they need to have a baseline of SQL here, and also, depending on the seniority of the role, that baseline goes up and up and up.

TIM: Speaking of that baseline, have you had any thoughts about how large language models might change this equation? So if we're at a point soon where maybe you could just be prompting the LLM to write the SQL for you, maybe you still need to interrogate it, and you still need to understand the data model. and there are lots of other details to be fair, but it feels like maybe the barrier to entry to starting to code, or at least code via a prompt, has now reduced. Yeah, would your views over this trade-off and change over the next year, do you think?

NICOLAAS: I do think so; like if you're going to put a gun to my head and say yes or no, I would say, Yeah, it will change.I struggle a little bit to see to what extent because in my experience, knowing the syntax of, say, SQL is one thing, and I know that LLMs are very good at writing SQL, also way beyond just knowing the syntax. But I know that a lot of the skill of, say, a good analytics engineer is not just in writing syntactically correct queries that produce the result you want but in writing good, composable, easy-to-read queries that help not just you when you write it and the person getting the result but the next analytics engineer having to maintain and read this code. and I don't think I'd have the trust in someone without those technical skills to be able to produce SQL like that written by anonyms I think they would look at the output and say this is correct but if anyone had ever at any point in time needs to come in and actually change that SQL without the use of an LLM that might be more difficult but if you if you take that you know one step further maybe you don't need those people anymore maybe the LLMs manage the code and it's just an output driven environment now and I actually don't know how that's going to go in that case I haven't been there yet myself but if you take it to its logical conclusion you might end up in a place where you don't actually need those technical skills anymore and it is more about how do you prompt the LLM to produce the results that your business needs

TIM: So we've just talked about, yeah, the large language model. Helping make the technical skills easier with the lower barrier to entry, but maybe the opposite is also true, which is let's say written communication is a soft skill, then surely that's now drastically easier. If you're a candidate who had patchy English skills and you found it quite hard to know all the perfect grammar and spelling and what have you, this feels like a bit of a level off for them as well because they could just whack in their best efforts text into LLM and say, Hey, can you just tidy this up for me? Make sure this is all grammatically correct, and suddenly they've got the ability to communicate professionally in email and Slack in documentation in a way that maybe they couldn't before, so maybe also it's a bit of an enhancer for those types of candidates as well, but maybe that's when people say soft skills"—maybe they're not really talking about the written stuff. maybe it's more like the interactions the stakeholder management the in-person I can't imagine putting this person in front of the CEO; that kind of vibe is what I normally get when I hear people talk about soft skills, or what are your thoughts?

NICOLAAS: Yeah, yeah, I was going to say I'm a little bit more skeptical on that front because indeed when it's about soft skills, the way I think of it is it's not about what you write at what point in time or even how you write it; it's about making the right decisions in a conversation, right? Whether that is face to face or whether it is on Slack, I don't think you can run your whole Slack with an LLM in the background. That'd be very impressive, and maybe that in itself is a skill that is worth hiring for, but when someone writes you on Slack, I think you need to be able to understand what they mean. You need to be able to read between the lines and then be able to provide an answer even if the question is a little bit vague. I feel that the people who understand sales, for example, are able to answer that to the satisfaction of the sales manager because they have some sort of intuitive understanding of what they might mean. and if not, they might also just be confident enough to say Hey, can you actually be a little bit more clear for me? Because I think this has been ambiguous, and I think those micro-decisions on how to interact with people, whether that is written or verbal, are very difficult to sort of hide behind an NNM.

TIM: What about AI? I'd love to get your thoughts on the use of AI and hiring. Have you already started to see candidates use it, for example, in optimizing CVS or taking tests? Have you dabbled in any use of it, for example, in screening CVS or anything else? What have you seen so far?

NICOLAAS: So I've definitely seen it used in making applications, so yeah, I need to optimize your CV writing cover letters, especially on the other side, though I haven't personally used it to screen candidates. I also have to say that the past couple of years at Growbox we haven't been hiring as much. So this, through the whole AI boom, I actually haven't had the opportunity to hire for many roles and had to use that where I do think it can help a lot, though, is parsing CVs and presenting the skills in a more fact-based approach or a fact-based way, and that's because when you read a CV, your eye is drawn to the way things are organized. So the way the skills are listed, the skills that are listed first When I read that, I think, okay, they're most proud of these skills, and these are the ones I want to highlight, but maybe there are other skills in there that can actually be valuable to me, and I might want to probe a little bit. So I think an AI could help normalize all of the resumes that come in and give you a much more factual understanding of each candidate that's there instead of reading the way the candidate wants to present themselves to you, and whether that's a good or a bad thing, I think is a fun discussion to have because there's also value in assessing how well the candidate can present themselves to you through a resume. but it's nice to know that there is the option of, say, turning that off, and if you want to look at it very objectively, then I definitely think AIs can help there to give you, in the same format, the facts about the candidates or in front of

TIM: Yeah, I feel like a big bit of the benefit would also be removing the noise from the CV. We could have a debate over what is signal and what is noise on a CV. For me, things like hobbies are not helping me make a decision. I'll give you one specific example. I remember so I was hiring a few years ago a product analyst, and we had hundreds of CVs. I'm screening for these CVs like a madman, like everyone else. By the way, it becomes tedious after a while. There's no way you treat the 100th candidate the same way you treat the first one because after you've read the 100th CV of the day, you're pretty bored, so there's such a level of bias that's already creeping in there anyway, and I remember seeing this one guy's CV, and I looked at the hobbies section, and it said he was a semi-professional footballer in Brazil. and we had a five-a-side team at work that used to compete on Tuesdays at lunchtime at Sydney University, and we were just narrowly missing out on the tournament every year. We would just lose the grand final by a goal, and so we desperately needed one more good player, so when I saw this guy's CV, I was like, Oh wow. killed two birds with one stone I hope this guy's amazing. We interviewed him; in the end, we didn't offer him the job, but he got an interview, at least partly because he was a footballer in Brazil, which obviously has no correlation at all to his ability to do A/B testing and user behavior analytics, I would have thought. And yeah, there's all that stuff on a CV that, yeah, if we did a pass through a large language model to begin with, there are some clever prompts I'm sure we could use to remove all those things that could be used to make a clearly unfair hiring decision and make

NICOLAAS: Yeah.

TIM: a bit more objective

NICOLAAS: but I do think it's worth having a discussion on in which cases or with which contacts that's desirable because I would also argue that the fact that you saw that he was a good football player meant that probably in the back of your mind it was like, Hey, this could be at least a good culture fit. Right, so maybe that box has been checked, and now you need to, of course, do a skills assessment and make sure that they're offered the job as well, but it at least helps you get an understanding of who this person is apart from the skills they possess.

TIM: Yeah, yep, this is where we get into this tricky area of subjectivity because you're right. I, yeah, the fact that he plays football and is Brazilian, already in my mind, I like him. Okay, I'll just be very blunt with you and very honest: I have a bias to liking Brazilians and liking footballers. Is that fair, though? Yeah, he's going to fit in with our team. A few of us played football, and we had the lunchtime comp, and we had a few Brazilians in the company, so it may be a similar kind of vibe, but it's a slippery slope, isn't it, to just hiring people that, like you in some way, know

NICOLAAS: No, it very much is. It very much is, but that's what I also find. I find it very difficult because I think you want to strive to be very objective in this. Ultimately, you have a role open, and that means you need people with a particular set of skills to be able to provide value to the company. But then on the other side, the chance of them succeeding in that role isn't a hundred percent based on the skills they have; it's also based on how well they interact with the people giving them work or the people that that might benefit from the work they produce, and that's where it just becomes, for me, a bit of a struggle on how objective do I want to be knowing that I will probably even also after hiring them look at them a little bit subjectively just to see if I think this will be a continued success.

TIM: Let me throw one at you so I feel like, yeah, there's always this element of just, do I like this person? Like, at the end of the day, some of the valuations are coming down to as simple as, can I imagine myself working with this person, which I think is fair enough because we've all got to get along. I also think it's maybe especially fair if you're hiring a consultant because their ability to drive revenue for the business is directly linked to how much the clients are going to like them, so that's absolutely fine. I think that's a very explicit, important bit of their job. I would just like to see those kinds of subjective gray things codified in some way to say, Listen, we're hiring this person; there are 10 things we need. We need them to have SQL skills, statistics, visualizations, and things that are obviously more easily measurable. They're a bit more objective, and here are these other things we're going to assess them for: how likable they are. That's going to be worth, I don't know, 20 percent weighting, and we're getting everyone who interacts with them to grade them on a scale of one to 10 and how much they like them. It's obviously subjective, but it's still at least making it an explicit bit of the process, so, for example, if one person said that they liked them at nine out of ten and someone else said they liked them at two out of ten, Oh, okay, that's an interesting discrepancy. You can always dig into why that's the case and ask the question, like, why did you like them? Why did you not like them? And maybe with that process, at least we would not be doing this unconscious filtering in our heads, and at least we'd have it on paper and have a real discussion about it and get it all out in the open. Yeah, you think of that approach.

NICOLAAS: No. I think it's very good. It makes me think that, yeah, is the resume stage the right place to apply the filter of do you like them or not, or is it better to do that at the end once you've at least picked the candidates or narrowed it down to the candidates that you know possess the skills to do the job? so kind of, and that I think ultimately comes down to the question of how important is that element of do you and do your colleagues like this person, right? It's a very subjective measurement. If you put it all the way in the beginning, apparently implicitly you make it a very important part of the hiring process because you may just reject candidates based on that initial assumption, which can be on a resume. But if you do it all the way at the end of the hiring process, it's more of a check marker; it's a tiebreaker if there's a couple of candidates that are equal in all other aspects. We at Roblox and in companies prior have typically put it all the way at the end because we want to be as objective as possible. and we also realized that hiring only could be a light in the long run; it's probably not going to result in the best company, so we, as much as possible, try to do a more objective assessment first. You know, do they take the boxes? Do they have the experience that we're looking for? Do they have the skills that we're looking for? but you then typically have a couple of candidates left who would then have a conversation with a founder or with another senior person and also actually often appear just to see how well they would integrate in the company and how much work is there going to be to make them successful.

TIM: If you're just trying to hire someone who's like the most extroverted, most friendly, and most make-sure-they're-comfortable, like, whatever these proxies are, then I think if you had no ability to understand the technical skills, you would have to filter on something else, and without that being really clear, I can just go bad so quickly. I think

NICOLAAS: Yeah, but I find it personally very difficult to not let those biases creep in. I think we've all, if you've ever hired for your role, you've probably Probably you've also been in a situation where you have a chat with someone or a conversation with someone, and just it's a really nice conversation. You know, it's fun to talk to this person, and it's very difficult to not let that influence whether this would be the right person for the job. You can always argue that, you know, having a good conversation is a qualification for that particular role, but in many cases it's probably just a very small part of the whole picture. So you need to really force yourself to kind of step away from those biases. In my personal experience, the best way to do that is just to be very clear about the process up front and kind of say, These are the things we want to know, this basic checklist of things. You know, how well did they score on those things? And then that conversation becomes a little bit moot because it's not on the list, right? Or maybe it's just one of the items on the list, and it helps you put that bias into context, you know, in the context of, you know, the larger pool of candidates but also the requirements you've set for the role. Yeah. I absolutely agree that is the best way to do it because without that you'll often get into these conversations where you'll be reviewing candidates and going, Oh, you're going through the pros and cons of the candidates, and then the feedback will often be quite reasonable, be like, Oh, they demonstrated more of a growth mindset because they said this, or I felt they were friendly because of this, or whatever.

TIM: And then you'll be thinking, Hang on, but these aren't criteria that we're searching for. Like, we already have the list of 10 things that we're meant to be looking for. Like, how did we score them on those things? Where are all these other things coming from? Which sound fine, and I feel like because they sound fine and sound reasonable, that's how easy it is to then be swayed off for one candidate or another before you're selecting for all these new criteria you've just pulled out of nowhere. So yeah, having a decision from the start and trying to stick to it, having some kind of scorecard certainly helps because you're sitting there. Going well, I have to give them a score in this interview. These things I want to score them on aren't here; we should have created a better criteria in the first place, or it's not important. It can't be one or the other, I guess.

NICOLAAS: Yeah, exactly.

TIM: Nicholas I'm going to throw one random question at you, which is if you had the proverbial magic wand, and you could change anything about hiring, you could fix it in any way. What would you do with that magic wand? What would hiring look like in some idealized world, which maybe we're not too far away from with the borderline magic of AI that we're seeing at the moment?

NICOLAAS: All right. That's a difficult question. The first thought that jumped into my mind is, Okay, better for whom? Better for the company or better for the applicant? Maybe there's a way you can make it better for both at the same time, but I do often find that what's better for the company is worse for the applicant and what's better for the applicant is worse for the company. So, a very concrete example of those very early days at Groblox was, I think, ideal for the applicants that we had. I wouldn't even call them applicants; I would call them, you know, candidates—people that we were talking to that we had reached out to and were trying to lure in, which meant that they had a fantastic negotiation position. and they had a very good time in the hiring process. We had to be very careful to not scare them away. It was very good for us, and this was at a later stage. We opened up a software engineering role, and we got tons of applicants. Now obviously you have to deal with all that noise, but you have the confidence that in all those applicants there are definitely a few good ones. and you're in a luxury position where you can ultimately choose between several good candidates, but that's obviously a very bad experience for the applicant because we can allow ourselves to make the process for them a little bit shittier, you know, make the ask a few more questions in the application, have them go through a few more rounds, all that stuff to make the assessment easier for us, but it's ultimately at the expense of their experience. So if I were to wave a magic wand and kind of let that loose, I would try to find something that would make it better for both, and I think what that would look like ultimately is a better process of matching applicants with companies because the reason you have so many applicants for that one role is probably because there's a bit of a mismatch between what they think would be a good role for them, right? and what's actually good for the company, and similarly for the company, there's just so much noise of all these different people to look through. It'd be much better if they could just get their top five immediately instead of having to go through all these filtering stages, so I probably focus on that part. How can you make sure that you just get a few applicants, but those are the right applicants, and then if you're applying for a job, instead of seeing 300 different jobs you can apply to, you can apply to five, and you can invest a lot of time in those five, but they're going to be the right ones? what that means I don't know yet; that's probably where I would focus my efforts.

TIM: You've painted a great picture there of what I imagine is a product that hopefully will be built in the next few years. Someone will make a lot of money matching candidates to jobs in an accurate, consistent way, and so I'm excited to see that idea come to fruition because I feel like we're not too far away from that.

NICOLAAS: Not exactly. I think it's doable. I mean, if you think about it, that's really the role that a recruiter tries to play, right? They approach companies and say, Hey, we're going to give you candidates that are valuable; you don't have to deal with noise, and they approach candidates saying, Hey, we have a couple of companies lined up for you. and then often in many cases this can actually work really well, so if there is a way you can productize this, you know, leveraging AI in the process, I think that could indeed be a very interesting product.

TIM: Exciting times ahead with AI and hiring Nicholas; it's been great to chat with you today about a variety of different topics, and thank you so much for sharing all your insights with our audience.

NICOLAAS: Yeah, of course. Thank you very much, Tim, for inviting me. I really enjoyed the chat. It's always interesting to reflect on past experiences and look at them from a new perspective, so it's been fun doing so, and yeah, thanks again.