0:00:01 – (Matt Widdoes): Welcome to another episode of Growth@Scale. I’m your host, Matt Widdoes, CEO of MAVAN.com. and today I’m joined by Dylan Serota, who is the CEO and co-founder of Terminal, a leading company forging the global talent marketplace. Before Terminal, he was at Eventbrite, where he led platform product initiatives, strategic partnerships, corporate strategies, bringing on board high profile partners such as Google and HubSpot. Welcome to the podcast.

0:00:24 – (Dylan Serota): Thanks, Matt. Thanks for having me.

0:00:25 – (Matt Widdoes): Yeah, I’m excited to dive in. So for people who don’t know you, tell us, like, who are you? What have you done? Where have you been?

0:00:31 – (Dylan Serota): So you mentioned at the top, I’m CEO/co-founder of Terminal, and that has been my baby in my life for the last seven years, since we started. Feels like a lifetime and a journey. But prior to that, as you mentioned as well, I spent prior seven years actually being at Eventbrite leading a bunch of different things. I feel like it was my corporate MBA, kind of rotating across different teams, different problems to solve. But one of the interesting challenges and opportunities I got was as we were exploring how to go global, really in search of a competitive talent strategy that led me to working and building teams in Latin America and Canada and Europe and took all that away to eventually start Terminal.

0:01:12 – (Matt Widdoes): Great. And Terminal was an offshoot of Atomic VC originally, is that right?

0:01:16 – (Dylan Serota): Exactly. So I left Eventbrite and joined Atomic, which is really the pioneer of the venture studio model back in 2017, and I joined as a founder in residence. So that’s kind of the entry point of starting a company is come in. And I had a bit of a different entry point into Atomic, which was really to focus on building a talent solution that eventually became Terminal. But at the beginning, it was just, here’s a major problem that the US is facing, that there is increasing demand for technical talent and we have very limited and stagnating supply.

0:01:50 – (Dylan Serota): And that imbalance causes a lot of challenges from enterprise value loss because companies can’t hire, all the way to frustrating CTO’s who just can’t hire or retain their talent. And so I came in kind of under the broad premise of this is a meaty problem that has pretty big societal impact, and if we can solve it, we can unlock a lot of pent up demand for innovation. Other founders and residents come in with more of a broad base mandate of exploring different areas and testing out a lot of different things. So my path was a bit different maybe than the normal venture studio path, if there’s such a thing.

0:02:29 – (Matt Widdoes): So, yeah, so for Atomic, this is the kind of thing that they would need in any sort of venture project that they were pursuing. Right. High quality engineers and staffing and all of the things that come with that from design, UI, UX, and the management of those teams. Did this start off as a solution to an internal problem that then was that then turned into something that should just be a service product in and of itself? Or did it start off kind of with that service layer in mind from the start?

0:02:51 – (Dylan Serota): No, great question. Absolutely. It started as a problem that they themselves were facing. And I think as a venture studio, when you’re simultaneously building, you know, 510/15 companies, every problem that one company has is compounded by the amount of companies you’re building. And so this was a big challenge for them at the time. Also, they had the studio set up in the Presidio, beautiful place in San Francisco, but also not the most convenient for different parts of the Bay Area. So they just had a very limited pool of talent that they were able to recruit and hire for these companies.

0:03:22 – (Dylan Serota): And what they saw was actually some of the best talent that was coming into their teams and companies was not from the Bay Area. It was from places like Canada, like Latin America. And it made them kind of pick their heads up and say, hey, how can we double down on this strategy so that we’re not competing against every company here? We actually can bring our opportunity to these different markets. And so phase one was, let’s build something that solves the problem for the portfolio of Atomic.

0:03:52 – (Dylan Serota): If we do that really well, it should then scale to other companies who are going through the same challenges. And so year one was really prototyping the model around the portfolio companies, and year two and beyond was taking it to the broader market.

0:04:06 – (Matt Widdoes): Well, you also have all of these potential inefficiencies that you see now in the standard corporate structure of. I’ve got a bunch of really amazing people that I don’t want to let go of because I know I have something coming. I might not even know what that something is. I just know there will be more things to work on. They are potentially bored at work because they’re not being utilized. I have all of these inefficiencies in that underutilization.

0:04:30 – (Matt Widdoes): In a company like Atomic, you can kind of redeploy them as needed. Fairly, I was going to say fairly easily, but just maybe more easily than a singular product company, like a traditional verticalized venture. But at the same time, this is an ongoing challenge. It’s the demand for world class engineering talent is far higher than the supply so why don’t we just kind of gobble up that engineering talent, keep them in one place, keep them active on a lot of different things.

0:04:57 – (Matt Widdoes): And large gaming companies can do that. Companies, you know, venture studios can do that because they have a lot of different products. But for something like maybe Eventbrite, where you kind of have a singular product, maybe some expansions outside of that, but by and large, it’s a hard way to capture that. And you don’t want to be in this endless cycle of turning over engineers. You want that consistent throughput and thread.

0:05:18 – (Matt Widdoes): And so it’s a no brainer, it seems like, to kind of also offer that up as a service, but it means you also have to then scale it as its own independent business and you’re still dealing with some inefficiencies or some gaps there. I’m curious, how has that changed over from the beginning until, you know, where we sit now? If we kind of think about the past and present for a second, and then I’d love to touch on the future of Terminal as well.

0:05:39 – (Dylan Serota): Yeah, if we look at the past and we look at the model that we prototyped, really around the original Atomic portfolio, it’s very different than what we’re doing today. So the original model was to set up remote teams, end to end as a service, but really from the pre-Covid perspective of setting up a satellite office somewhere. And so we were looking at the best cities. And within those cities, what are the best areas to open an office? And then we were building out these terminals or these campuses that were multi-tenant. So kind of like a large WeWork, but solely geared around housing different engineering teams. So in one terminal, let’s say in Waterloo or Toronto, we would have 20 different companies, engineering teams that might have anywhere between three desks, up to 20 desks within one common building with a cafeteria and a stage and events and. And meeting spaces.

0:06:35 – (Dylan Serota): But at that time, that was the perspective of CTO’s, was, if I’m going to expand, I want the team to be together in one location, I want to be able to travel to that location, and so I want everyone co located. That had to change when Covid came about, because all of our campuses, all of our terminals were empty, everyone went home. And what we saw was a change in perspective on both the demand side or the hiring managers, as well as the engineers.

0:07:01 – (Dylan Serota): Engineers were not motivated and were not eager, excited to go back into an office. They actually were loving the flexibility of work from home. And it obviously gears very well towards engineering, especially engineering culture, as well as just the product development lifecycle to be remote rather than in office. But hiring managers then said, hey, look, if they’re not at the office, why do I need to limit my search to this one geography, this 30 miles radius around, around the office? Let’s just get me the best engineer wherever you find them.

0:07:31 – (Dylan Serota): And that opened our eyes of, all right, we are no longer a vertical service platform for building teams. We need to evolve into a global marketplace. And so that pivot really took about two years probably, to go from what was working for us, and it was working very well at that time to a completely different model that was location agnostic. It was much more built around the matching of supply and demand, the merchandising of candidates in the marketplace, while still retaining some of the vertical services around employment payments that make the rails of the marketplace work.

0:08:10 – (Matt Widdoes): Makes sense. Yeah, it’s funny, I think it was probably my first week at King was meeting with the CEO, and we’re talking about the early days at King, which they were actually in dating apps initially, and then they found that all of the conversations around dating were kind of happening around games. And that was like a common thread where they decided, why don’t we just go into games? But he mentioned that one of this is Ricardo Zaccone had mentioned that one of the early things that they learned, and this conversation was taking place in probably 2017 or something like that, and they had been founded in like ’11 or something, but they had come to the realization, which is so obvious when you hear it, but essentially, if you want to hire the best people in the world, you have to be willing to hire all over the world, because they don’t all live within 15 minutes of the Presidio or 15 minutes of Waterloo or whatever.

0:08:57 – (Matt Widdoes): And in many cases, and yes, there are like lots of diamonds in the rough that you have to kind of discover, but you have this amazing set of resources distributed globally who maybe don’t have access to the same types of projects or whatever. And again, it’s not easy to find these people, but when you do, it’s very valuable both for them professionally and from a, you know, personal fulfillment, and of course, very valuable for the companies that are able to utilize those types of resources.

0:09:20 – (Dylan Serota): Yeah, I totally agree. And I think what gets me excited and the rest of our team excited is this kind of dual aspiration that we have. One is from a very mission driven perspective, we firmly believe that talent is equally distributed, opportunity is not. And the way in which we can actually deliver more opportunity to talent that you know, wasn’t lucky enough to win the geographic lottery. That makes us feel good. And I think that does produce a lot of good in the world. And the second is really solving a problem that we think is fundamental for just human progress, which is if we can unlock the bottleneck on innovation, which usually is the shortage of product engineering, design resources, we can actually accelerate a lot of the solutions to problems that I think we all face. And I think the combination of the two that we could do both of those within one company, gets us all really excited about the work that we’re doing.

0:10:08 – (Matt Widdoes): Yeah. And how do you think about the future now that you’re in a post-Covid world? You’ve got the lessons from the last few years. You’ve made these kind of somewhat minor, but sometimes maybe more medium or major pivots, like, how are you thinking about the future of Terminal as it sits today?

0:10:21 – (Dylan Serota): I think there’s never been a more optimistic time for Terminal. And I think you’ve got a couple things really building that groundswell. The first is toothpaste is out of the bottle. With remote work, people aren’t going back. Most companies are going to follow a remote default, a distributed or hybrid model that leverages the strength of geographical distribution of talent. And tapping into that in a way is requires, I think, a marketplace model in order to organize the supply and provide access to that supply in a meaningful way for companies to scale. So I think just where the trend is going with global talent, you need a platform or a marketplace to really be able to access that. The second is understanding that cost of capital is not zero.

0:11:08 – (Dylan Serota): And I think that’s an awakening that we’ve had over the last few years. And I don’t think that’s going away. And there are innovative talent strategies that you can now employ for the first time ever. Leveraging global talent in different areas of your company as different parts of your strategy. To be able to lean into a low cost solution by having lower cost in your labor, you can tap into more specialized talent in different markets that are less competitive, like AI talent in Canada.

0:11:39 – (Dylan Serota): That allows you maybe to have leverage on certain aspects of your products and services. And so I think that bringing that into the forefront has really eliminated a lot of the fear, uncertainty, doubt that companies or investors especially, had previously around hiring people outside the valley or outside the United States, for that matter. So the fact that the stigma is gone, the workers and the talent are here to stay in a more flexible work environment, and then you have cost of capital kind of increasing the opportunity and really, the rationale for why you should do that, I think puts us in a really interesting position.

0:12:16 – (Dylan Serota): And again, kind of going back to just like where we settled on from a business model perspective. The marketplace model, I think makes a lot of sense in this context where it previously didn’t exist. And so I think we’re one of probably many talent marketplaces that will evolve and will be created over the next decade. And I’m excited for where we’re positioned now, especially in the realm of, of engineering and technical talent.

0:12:40 – (Matt Widdoes): Yeah, we believe deeply here as well, that this is kind of the future of how teams will be built and augmented. I think there are, especially in roles like engineering, where somebody usually starts with a known discrete need that maybe isn’t worthy of a full time role, or is, but they can’t find the talent that they want, where they need this person to sit and maintain that context over many months or years, as well as just specialized talent, where maybe they need a Golang developer in this example, where they’re hard to come by.

0:13:09 – (Matt Widdoes): Even the ones in the Bay are no longer in the bay as an example. Right. It’s like they’re, they’re everywhere. So if you’re. Yeah, I think the kind of toothpaste being out of the, out of the bottle, as you mentioned, is, is there and people are. It’s almost funny that we didn’t operate this way before, but it made, it made everything kind of come to the forefront where people, you know, outside of like sales teams being face to face, probably. Right. That’s a, that’s something that still needs to happen, but they don’t need to be flying out of the same hub.

0:13:32 – (Matt Widdoes): And so I totally agree that really, almost every seat in the can be remote and could technically be fractional when people need it for the time that they need it. And sometimes just as a bridge between them finding their forever person and what they’ve got, you know, today. So I’m curious to hear your thoughts at a high level on venture studios generally, and incubators, maybe we can lump in there together, but their kind of role in contributing to early venture success as it’s similar in some ways to the kind of fractional start, if you will. But I’m just curious to hear your own thoughts, having worked inside one of the largest venture studios on the planet.

0:14:08 – (Dylan Serota): Yeah, I mean, I think they’re, they play a very integral role. And I think for me in particular, you know, if it wasn’t for Atomic, I don’t know that I would be an entrepreneur like I am today. And part of that was I was at a critical juncture in my life. My career at Eventbrite was going up, I was getting promoted. So I was seeing kind of pay increases, but I didn’t have a payoff where I felt like, all right, I’ve got a cushion where I can go really start something and take that risk of maybe six months, up to a year salary, and the model to have some money coming in while I was building and prototyping and understanding if this was even viable, that to me, really helped mentally de risk the opportunity to jump in.

0:14:54 – (Dylan Serota): And I think there’s a lot of people who come from very different points in life where they are making the jump into starting a company. But I think they play one of those roles in offering kind of an on ramp to those individuals that otherwise might not take that jump. And so I think in one way they are helping to create more entrepreneurs, whereas VC’s are just funding entrepreneurs. They are creating them through this model. But it’s a very difficult process to find the right balance and mix of those individuals who have really the core competencies to do, to do the job and do it well, but then also those individuals who are willing to trade off on the economics, because it is different. When you are building with a studio, you’re building with another co founder or set of co founders alongside with you. So you’re immediately splitting the pie different ways than you would otherwise.

0:15:45 – (Dylan Serota): And so I think finding people at that right time with the right capability, but then with the right mix of incentives, it’s very difficult. And I think most studios are still trying to figure out really the right way of recruiting and identifying that talent, bringing them in, and the right incentive model that doesn’t push too much economics one way or another. And so studios have this, I think, very interesting challenge of you need founders, you need capital, and then you need talent, and you need those three to all be really in sync so that you can scale up.

0:16:21 – (Dylan Serota): Because a studio that doesn’t get to portfolio scale is unlikely, just probabilistically, to get enough of an outcome that can return a fund, can secure another fund. So I think it’s a very, very difficult job in terms of getting the studio model right. But I do think they play a very interesting and important role in hopefully creating more paths to entrepreneurship for more people.

0:16:46 – (Matt Widdoes): Well, I think, too, it allows, from our perspective, there’s just knowing our network, there are a bunch of what I would call latent entrepreneurs. These are people who have some idea that is probably viable that they’ve been thinking about for some number of years, but they’re locked into the economics of a full time role at a major company. They’re on that same career progression where they’re seeing the cash and they have equity in something.

0:17:09 – (Matt Widdoes): They have equity in something meaningful, in spite the fact that the equity maybe isn’t outsized. And they have to look at that economic trade off. And in many cases, they don’t even necessarily want to pursue the idea in and of itself, even if, well, if you could make all the money the same, they would, but you kind of can’t. And then on top of that, the traditional model means they have to go put together a deck, put together a team, put together, get all these pieces done, and have everybody else work on equity because they typically don’t want to put in a quarter million themselves, as an example, and then they have to go raise, and then they have to take this huge pay cut and all these other things. And I think the studio model offers an opportunity to kind of fast track some of that stuff. Skip the finance side in some ways, skip the team building phase in some ways, but it doesn’t remove from the fact that you still need a good idea. You still need capital, you still need talent.

0:17:59 – (Matt Widdoes): And sometimes it’s worth probably splitting the pie a little bit different because you’re saving. 80% of the hate goes away through a studio model, in some cases, for the entrepreneur.

0:18:10 – (Dylan Serota): Totally. And we didn’t touch on it, but, I mean, I not having to deal with running payroll, setting up a cap table, all of the, you know, talking with lawyers all day to get the right, you know, copyright, trademarks, like having a lot of that taken care of so you can focus 100% on the product, I think that’s a big other aspect that turns people off of it, of like, having to, you know, sit down and run payroll at 10:00 p.m.

0:18:33 – (Dylan Serota): one night, you know, and have that offloaded and be able to focus just on the product, the customer and building is, I think, one of the other obviously big advantages that they bring to the table.

0:18:42 – (Matt Widdoes): Yeah, there’s just a ton of efficiencies, too, with people, teams that have done it before together, right. Because you can assemble five killers, but they don’t play well together. And so having an engineering team, a product team, a creative team, a finance team, a whatever team, all kind of not reinventing the wheel every single time on a new venture and allowing those teams to pivot as needed to where the most signal is coming from.

0:19:06 – (Matt Widdoes): Um, seems to make a lot of sense, but it’s in some ways, it’s almost identical, actually, to the game studio model where we look at a bunch of different games. We prototype. They have to pass certain green lights, people come off of projects and onto this one. Um, and you’re just kind of using similar internal, like, centralized resources to solve some of the same problems over and over again. So you’re getting some of that efficacy and efficiency with the headcount you’re already carrying and. And all of the systems that need to exist for all these things. The submission to iOS is the same every single time. And there’s a clearly defined process and there’s clearly established relationships within those platforms so that you can get that done quickly again, versus me at a new company with none of those same resources or whatever doesn’t always get the same.

0:19:48 – (Matt Widdoes): The phone calls are oftentimes very different than going back to a King or a Zinger or a VC, like Atomic or, I should say, totally. So you mentioned, as you were thinking through both the past, present, and future of Terminal, a lot of how remote work has changed the landscape in many ways, but in particular for your business. I’m curious, as remote work becomes more and more cemented as just a part of how we operate, and I wonder if that’ll change much if we fast forward 25 years, what things we’ll still be seeing as a byproduct of Covid and everyone being remote for so long.

0:20:29 – (Matt Widdoes): Um, but, you know, and given these kind of pushbacks with people being called back into the office, you see headlines every day of, I think it was Dell most recently, or somebody who’s like, half of their employees are like, nope, we’re not going back. And then the corporations need to deal with that. But I’m curious, like, how have you been able to kind of meet demands of remote work and global talent acquisition for tech companies? It’s part and parcel of what you guys do, but how do you guys think about solving for that?

0:20:53 – (Dylan Serota): It’s a good question, where to start. So I think one thing that I heard a while ago that stuck with me, I think it was, might have been Parker Conrad. I’m not sure who, but it was that remote is the worst way to build a company that’s better than all the rest.

0:21:06 – (Matt Widdoes): Okay.

0:21:07 – (Dylan Serota): And when you unpack that, it basically says, look, building a company, there’s no easy way. So don’t think that, like, there’s one model or one silver bullet, that if you build it this way, you’re going to have a successful company, successful team, successful culture. But what remote provides and advantages, I think, outweigh a lot of the negatives, but it has to change the way you think about building companies. And I think we relied a lot on the osmosis and collateral that would happen with people in one office that you just, you know, you don’t have to define a lot of things because you just let people collide into each other and talk and figure out things and tap each other on the shoulder.

0:21:46 – (Dylan Serota): Remote doesn’t allow for that. You have to be a lot more thoughtful and intentional about what’s your culture, what are your policies? How do we do things, why do we do things and make sure everyone is really locked in and in tune? Because you don’t get the benefit of just letting things happen when you get people in a room. And I think that actually forces a lot of positives, but it takes more work to set up. But the advantages you get, I think, are immense.

0:22:14 – (Dylan Serota): Any company, once they really breach, call it 50 employees, they start opening up, maybe a satellite sales office, all of a sudden they’re pretty much operating in hybrid mode anyways. And if you’ve really built that fabric from the ground up, I just think you’re going to see companies be a lot more resilient at much bigger scale and less of those breaking points or those challenging points when kind of the traditional model of a company goes through those.

0:22:41 – (Dylan Serota): So I think it’s harder. But the advantages you get by putting that into the fabric at the beginning, but then also obviously, the talent advantages that you’re not geographically constrained and very small things can have tremendous impact on one market. But if you are agnostic geographically, world is really your oyster as far as things you can do, where you can hire, how you can build your team, and I think that’s a lot of strength that we’re going to see play out in the next generation of companies coming up.

0:23:08 – (Matt Widdoes): Yeah, I think to the systems and processes that come out of a largely remote workforce create some advantages as well. I mean, it requires a certain type of person to be able to work within the remote kind of framework. Right. Some people need that face to face interaction. They need the water cooler stuff. They need to be able to quickly jump in to rooms and whiteboard, although you can still do that remote, but there is kind of a nuance there that is different, but it does create, I think, a more.

0:23:36 – (Matt Widdoes): I guess I would say there’s probably, like more rigor that is required to make that work and that there are major benefits of that if you find the right people that can play within that framework.

0:23:45 – (Dylan Serota): 100% agree.

0:23:46 – (Matt Widdoes): We hear a lot from startups on how hard it is to find and retain quality developers. This is true, though, inside and outside of remote. I’m curious, any advice you have for startups looking to scale as it relates to building a dev team generally, and if there are clear advantages or disadvantages of taking a full stack approach?

0:24:08 – (Dylan Serota): A lot of thoughts there. I think one of the things that I hear get thrown out a lot is every company needs to hire top engineers and we want to hire the best developers. And I think that misconstrues a lot of what you should actually be thinking about when hiring and what we have seen just through our data, which I find is interesting. All of our companies come to us and say we want to hire the best engineers you have, nobody wants, and we have really high bars for talent. No one’s going to come in saying I want your mediocre engineers and we’ve got a low bar for talented and.

0:24:43 – (Matt Widdoes): We want to pay a lot for it.

0:24:45 – (Dylan Serota): Exactly. So we hear the same thing from everyone, but interestingly, we’ll have the same engineer interview at four companies that all say the same thing, and two of them might give them a strong no rejection. Two of them will give them a strong yes, and one of them will end up hiring them. And that happens time and time and time again. And really what it shows is that the best is very dependent on the company, the stage, what they’re solving for.

0:25:14 – (Dylan Serota): And I think that the idea that there is this hierarchy of just from best to worst is a really bad mentality, because I think it actually forces hiring managers to think about not the job that has to get done and who is going to be the right component as part of the team to get that job done. They are the really anchoring themselves to some false heuristics that they have just heard play out. So we want to hire someone who’s had Google experience, or who has made it through the Facebook interview process, or someone who has had this type of experience.

0:25:53 – (Dylan Serota): And I think what I have just seen is that the best companies, the best candidates, are really thoughtful around what they want and block out a lot of the, a lot of the superficial aspects of hiring. And so I think that it’s easy to get caught up in that idea that you want the best, put that out of your head and think, what is the best for my company? And what do I need to get to? And we see that probably most often happens when you have a hiring manager who is hired for one of the high flying thousand person companies where they have thousands of inbound applicants. They’ve got a huge recruiting and sourcing team. They’ve got tons of data to sift through to find the people who most map to those engineers that have had success prior.

0:26:41 – (Dylan Serota): They get moved to being the VP of engineering at a small startup, no name brand, very unknown. You know, aspects of their, their technology don’t know how viable the company is going to be long term, but they hold the same hiring bar and they think that they are going to get the same crop of people. And I feel like that is one of the biggest challenges I see is people reassessing what is their value proposition as a company and then map that to what is the talent that we can actually access and what’s the job that we actually have to get done for those individuals. So I think there’s just a big breakdown that has to happen when companies are thinking about building and scaling their team. And that’s outside of thinking about whether they hire that person in San Francisco, New York, or Guadalajara, Mexico.

0:27:27 – (Matt Widdoes): Well, we see this a lot too, sharing similar aspects of business model. But there’s also a really big advantage that doesn’t exist in the traditional hiring framework of a traditional full time hire at a major company, which is the ability to just get to work. And usually, I think within the first, conservatively, the first six to eight weeks, you can tell very quickly whether they’re a full time hire or a fractional hire or a one off random person coming to the business if they fit within the mold.

0:28:01 – (Matt Widdoes): And there’s so much that the traditional hiring process doesn’t capture in that some of the heuristics you mentioned. But like, again, will they play well within our team? Can they meet our standards? We can talk a lot about for both sides, how do you like to work? And this is how, these are some of the problems that we’re dealing with. How would you go about solving those and kind of get some sense of where their heads are at? But there’s no substitute for actually just working on stuff. And so I think that that model allows people to kind of quickly sanity check and then move, and then oftentimes redefine what the full time role is that they’re hiring for, because they may find that somebody came in and essentially solved, or will solve soon 80% of what they thought they were hiring for.

0:28:40 – (Matt Widdoes): And they’re like, whoa, this is going to be way different than we thought. And now it totally changed our perspective on who we would hire next or if we even want a full time person in this role or need it, or what this person may do next. And so it allows people to be way more nimble and it gives more variety to the people doing the work as well, which can benefit all sides. From a burnout perspective, from a productivity standpoint, from a fresh eyes on something that nobody else has looked at, versus somebody who’s going to in a full time seat, it may take you six to twelve weeks to question whether somebody’s going to work out of another six to twelve weeks to try to pull them along, another six to twelve weeks to decide are we really going to potentially rehire for this role and put this person on a PIP and then more time to work through the traditional HR process of off boarding them or giving them one more chance or whatever. And that cost that like nine month wasted effort even just ignoring the cost, just thinking about the opportunity costs of not getting the thing done that you needed done and having to go again is also tough. Not just financially and from opportunity cost, but from a cultural standpoint, nobody wants to see people coming and going very quickly in and out of orgs on any side.

0:29:47 – (Matt Widdoes): It’s bad for the team, it’s bad on so many levels. And so there’s, I think, a huge opportunity and advantage of speed to hands on keys still going through somebody’s own checklist of what they want to see. But you’re not making, you’re not forced to make a forever, essentially a forever decision, three to four year decision out of the gate. And that can be super advantageous for companies that are trying to move quickly, especially with very discrete, specific needs.

0:30:11 – (Dylan Serota): I agree we didn’t touch on this at the beginning, but one of the unique things about Terminal is we’re one of the only platforms where you can hire both full time and contract talent. And by virtue of that, you actually have this very natural contract to hire model built into our hiring platform. I think what you touched on is really well said. And one of the things that I see with a lot of hiring processes is the good ones are built to get to a higher, the bad ones are built to reject.

0:30:38 – (Dylan Serota): And I think that when you are thinking about a full time, forever hire, you start to think very tribal around. Well, this person has to go through all of these steps. They have to complete these two tasks and missions before we get to this step. And they have to pass all of this before we’ll let them into our tribe. Even if the job they’re doing has really nothing to do with the interview process that you just put them through, but you feel like you have to put them through that, that type of rigor in order to bestow full time employment.

0:31:07 – (Dylan Serota): And I think the smarter companies are like, you know what? This person checks a lot of the boxes. Let’s bring them on six month contract. Let’s see how they perform. If they perform well, let’s give them an offer and convert them to being it full time. And that try before you buy for talent, I think is, is novel, and I think people are coming around to it. And I agree that. I think, like the next decade, we’ll see all these different types in terms of fractional time, contract, contract to hire, where there’s a little bit more actual testing on the job rather than creating these crazy, you know, obstacle courses to get in.

0:31:39 – (Dylan Serota): And I think that. I think that those are going to crumble in the face of just speed and efficiency and actually seeing how people perform and jive with the team. Because there’s nothing more unnatural than an interview process.

0:31:54 – (Matt Widdoes): Yeah, well, and it doesn’t an, and it doesn’t answer the core question of what will they be like on the team? Because you can actually end up with a lot of people that interview really well and say all these things, and then you have them actually implanted in teams where the people are saying, how did this person get hired? Like, they didn’t know this or this or this, or they took 20 hours to do this thing that should have been 20 minutes. And there’s like, all these question marks that can just be rooted out. And I think that there are.

0:32:18 – (Matt Widdoes): Any large company has significant data on how well that hiring process is working. And some of that is actually muddied because you have people who get hired. I I’ll spare any major brand, but where they’re like, they’ve been there seven years and they were never good, but their managers have changed. They’ve always somehow escaped layoffs. They’ve always whatever, and they’re just floating around. And these people are like, this is fairly abundant, and so half of that’s not even captured. So even if you’re just looking at how many employees are ahead of people and you’re looking at how many employees are still here 18 months or a year or six months or whatever, their yardstick is probably just 90 days post higher.

0:32:56 – (Matt Widdoes): They can pat themselves on the back, but they probably are still seeing attrition because of the most obvious cases. But, like, all of that still took six months to hire that role. And in the meantime, that six months of work was not being done. So anyways, it’s a topic I feel strongly about and we’ve just observed, which is why don’t we just start working and we’ll see. And a lot of those things that I’ve found can be fleshed out often are actually defining the role they thought they needed. Because once you put an expert, they’re like, well, what about all these things? And someone’s like, oh, that’s a good point. And then they’re like, oh, okay, well, we almost just hired somebody where we made a commitment to them for the next, I’ll keep using 18 months because it seems like a relatively, you know, it seems kind of where most people’s heads are probably at, where they’re like, oh, okay. And then you’re in this spot where you’re trying to reposition that person internally because you made a commitment to them, you still like them, but you found out your problem was different because they finally looked at it and said, that’s not what I do. This is different. You never told, we never talked about this.

0:33:51 – (Matt Widdoes): All we talked about were I, how many golf balls can fit in a 747 or whatever that might be. And then it’s like, okay, and now you’re reframing teams or going through layoffs, and all these things are bad culturally as well, and not to mention the finance and the business opportunities.

0:34:05 – (Dylan Serota): Totally agree.

0:34:06 – (Matt Widdoes): So I’m curious, you know, AI is on everyone’s mind. I know you guys have leveraged it in some capacity internally. I’m curious, how are you guys utilizing AI right now? And like, where have you been able to find efficiencies that you think are meaningful?

0:34:19 – (Dylan Serota): Yeah, the rise of AI for us couldn’t have come at a better time. And transparently, we were in a pretty tight position. Being a tech hiring platform, we definitely saw the impact of hiring going down in our core market, which is growth stage companies. As interest rates went up, there’s less funding and obviously less hiring, a lot of layoffs. That definitely affected our business. And so we really had to pull back from a cost perspective. At the same time, AI large language models were coming out proliferating, and we were trying to think about ways that we could leverage it. And we had to do two things. One is we had to become much more efficient. We realized that we had scaled with a lot of people to do a lot of jobs that actually could be done more efficiently and effectively using a lot of AI tools that were coming to market.

0:35:10 – (Dylan Serota): So, for example, we had huge teams of talent curators who were looking. You know, we get thousands of engineers who build a profile on our platform every month. And they were going through profile by profile. And the first step was just, is this an engineer? Like, is this a software engineer? Is this a civil engineer? Is this a materials engineer? Like, we’re in the business of finding software engineers. So that first screen had to take out actually a huge portion of candidates that were coming to us, but that was humans and human time to go through that.

0:35:39 – (Dylan Serota): Once they were approved, we had to input a lot of the data so that their profile was merchandised the right way on our platform. Long story short, we realized we’ve got seven years of hiring data and we have first party data. We have candidates giving us all this information. We have hiring managers going into those candidates and saying, pass, reject. Here are the reasons. And we were able to compile all of that into a scoring engine so that a profile came in.

0:36:04 – (Dylan Serota): We could easily pass the ones that we knew were high quality software developers, we could reject the ones that were not a fit, and we could just do human review for a small portion of the ones that needed that closer look. And then all of the actual marketplace operations of getting that candidate from applying to actually having a profile on the platform entailed all these very meticulous steps of information gathering, keyword classifying, what type of engineer they were, all of that we’ve actually been able to automate as well.

0:36:36 – (Dylan Serota): So when we have a screening call, they’re sharing all this information with us about their prior experience, a lot of which they never put on the resume. We can capture that and we can identify. Here’s three other skills that they didn’t have on their resume that we can now append to them. So we can take their transcript, run it through a large language model, append their profile, and publish it. So we’ve been able to scale down our team, reduce our cost basically by 50% year over year.

0:37:01 – (Dylan Serota): Meanwhile, we’ve actually increased the speed and velocity and throughput of the marketplace by about 2-3x. And a lot of that is through adoption of AI in a really practical way, especially for us, because we’ve got a lot of structured, yet unstructured data coming in, and we need to have a very structured, standardized output of that. And before, we relied on literally over 100 humans to do that. And today we’re doing that with the team sub 40.

0:37:27 – (Matt Widdoes): And what are some of the things that you’re looking for there? Like not to peel back the onion too much. I don’t want to reduce the efficacy of your, of your model, but what are some of the things that you look for that are like early indicators or high level indicators? Because, you know, many companies are using AI to screen candidates and have, or have used at minimum, like a rudimentary keyword tool on things. This is where you see lots of professional copywriters for people’s resumes, just stuffing it with keywords. How do you, what are you guys looking for? Either not necessarily in search, but just like, what are some of their indicators that you guys see are indicative of.

0:38:04 – (Dylan Serota): Of high performers in the actual candidates themselves?

0:38:06 – (Matt Widdoes): Yeah. Like what are, when looking at kind of like software developers or, you know, people in your sphere of influence that you believe will be, have a high likelihood of being in demand? What are some of those kind of flags that come up? And maybe we don’t need specifics, but just kind of generally, like, how do you guys kind of think through weeding through so you don’t end up with too many false negatives?

0:38:26 – (Dylan Serota): Yeah, it’s a great question. So one is relying on our data, our past data. So with, with the seven years of hiring data, we basically have over 500,000 inputs on profiles. And so using that to build a model, we can very probably probabilistically tell whether this profile, what likelihood are they to pass a first round, a second round or a third round interview? And that is looking at things like what type of engineer they are, how long is their work history, what types of companies did they work at in the past? What is their computer science program?

0:39:05 – (Dylan Serota): What ranking is that? Computer science school. Let’s see what else. We can look at the transcript and we can grade it on a scale of English proficiency. And one thing especially for global talent, that’s important is communication ability. What’s their vocabulary? How clearly can they articulate certain concepts and talk about where they had ownership? And we can even extract things around motivation.

0:39:27 – (Dylan Serota): And then one of the things that we do is, again, going back to our prior conversation, it’s not hierarchical. What we do is we just try to create the most robust view of a candidate, leveraging all of these different tools and sources. And then when a job gets created on the platform, we use a vector database and a distance model to actually say, how close is this job description, the combination of skills, the language that we’re using to describe the individual, and how close is this profile?

0:39:57 – (Dylan Serota): And the closer they are on vector database is how we create our shortlist recommended matches. And so it’s not about, here are the best front end engineers. It’s, here’s the actual what we’re looking for. And how close are these different profiles resembling the specific skills, the language, the experience levels, and maybe even the location and communication and motivation aspects. And so I think that’s a, I find, just interesting approach that we’ve taken, but I think it also is the most compelling for finding, again, the best matches versus just the best people.

0:40:31 – (Matt Widdoes): Well, I’m back testing that against your quality score and assumption with what you actually see come out of the model. You know, we’ll further enrich that. I’m curious, have you guys been able to automate that initial screening call? Because that’s, that could be a. It probably is a company in and of itself, but I think there’s probably people trying to solve that, but have some list of AI prompts that are like, cool, tell me about this. And then I just read it and I say, okay, well, yeah, let me give you the answer to that, or.

0:40:56 – (Dylan Serota): Whatever we actually do. We have two options for candidates. When they come in, they can schedule a human screening call or an async screening call. And so the async one is we’ve got someone based on the role asking questions that are pre recorded. And so they can do this at, you know, midnight their time, and they can go through and answer the questions. We still get the transcript. We can still get all the information needed.

0:41:20 – (Dylan Serota): And what we do find, though, is good candidates. And we actually find that the better candidates often times choose the human screener. And so we don’t want to get rid of that because there’s something, especially when you’re choosing your livelihood. You want to talk to an individual, you want to know that you’re legit on the other side, and you want to have a conversation because you know that’s going to bring out the best in you.

0:41:38 – (Dylan Serota): And so we still give them that option. They can book a screening call with, with someone from our team, gives them an opportunity to ask questions, but also to get comfortable with someone from the other side. So we’re giving them both options.

0:41:51 – (Matt Widdoes): Got it. That makes sense. I’m curious, any other kind of emerging tech that you think will be significant with global market like talent marketplaces moving forward over the next decade?

0:42:01 – (Dylan Serota): I think there’s so much, frankly, I just think that the model of talent marketplaces, the nature of it is that you just never have homogeneous supply. Every person is a little bit different. They might all be front end engineers, but based on what experience they have with different frameworks, different languages, different industries, it’s all different. And so I think one of the challenges with marketplaces, unlike some other marketplaces, like a car marketplace where, you know, it’s a red Ford this year, this mileage, how do you take all of those nuanced aspects and then be able to build a candidate database and a model around understanding a lot of that unique structure of data?

0:42:42 – (Dylan Serota): And so I think with the way I’ve seen us implement LLMs and the different generative AI technology, it’s going to open up just a huge field of opportunity where economically it didn’t make sense because you’d have to have so many humans doing all of this custom work. If you can actually automate 80% of that now, I think it drops the cost, the operations so low that you’re going to actually see a ton of interesting innovation on talent marketplaces that didn’t exist before.

0:43:11 – (Dylan Serota): So I’m excited to see what the future looks like and hopefully we’ll play a big hand in defining it.

0:43:16 – (Matt Widdoes): Well, if internally Terminal or internally at a company who’s hiring, if you can also repurpose the time of your best human interviewers or whatever you want to put on that kind of HR or people Ops role, you can have a higher touch, you can have a higher quality with the candidates that you think have the best fit versus spending 95% of your time with candidates that are potentially not good fits. So it can, it can enrich the lives of the people hiring as well. Totally.

0:43:43 – (Matt Widdoes): And the likelihood of a quality outcome. Cool. Well, Dylan, thank you so much for the time today. There’s probably a bunch of other things that we could unpack in the future, but appreciate your insights and context on where you guys are heading and ways that teams can manage remote work and find the best talent. Really appreciate the time.

0:44:00 – (Dylan Serota): Thanks, Matt. Appreciate you having me on.

0:44:03 – (Matt Widdoes): Absolutely. Look forward to next time. Thanks again.

0:44:04 – (Dylan Serota): Bye, Matt.