Beyond the Match: How the Market Benchmarks Candidate Relevance

[00:00:12] [SPEAKER_02] Welcome to episode 144 of the Inside Job Boards and Recruitment Marketplaces Podcast. I am Steven Rothberg. I'm the founder of College Recruiter, Job Search Site. And we've got a bit of a change in format coming up with starting with this episode. We used to publish every other week with the occasional breaking news episode dropped in here and there. And each episode would run for roughly 14-15 minutes.

[00:00:40] [SPEAKER_02] Peter and I have been talking and I think that half of our audience wants more Peter and less Steven. And the other half wants more... No, nobody wants more Steven. No, but in all seriousness, what we're doing is we are going to move to a monthly schedule, but each episode is going to be roughly twice as long. So it's the same total amount of content, but this format gives us a much better opportunity to really dive in deeper with the guests for what it's worth, share our own opinions about the topic, and really

[00:01:10] [SPEAKER_02] get into a much better conversation. So second Thursday of each month is going to be our new schedule. And this is the first one. Peter, with that said, how the heck are you?

[00:01:24] [SPEAKER_01] I'm doing well. Just came back from Istanbul. Beautiful, beautiful city. Great time there. And I'm delighted to be here with you on the podcast. Today with us, we have Prachik Ajamiyan, which is not a Russian name or an American name, but an Armenian name.

[00:01:43] [SPEAKER_01] And he has founded four or an American name. And he has founded four or five companies. And currently he's the founder of Perch AI. That's P-E-A-R-C-H dot A-I, a Bay Area based company that does matching and AI relation for candidates and employers.

[00:02:06] [SPEAKER_01] So let's talk a little bit about matching because there are so many issues right now with matching. There are the resumes that have all been scrubbed through AI and match the job description perfectly. There's the flood of thousands, millions of resumes. There are the job descriptions that are looking not only for a purple unicorn, but a purple striped unicorn.

[00:02:33] [SPEAKER_01] That's Steven. That's Steven. He's the purple striped unicorn. But talk about matching a little bit. Talk about how you see it and how it's changing.

[00:02:43] [SPEAKER_00] Yeah, it's interesting. I think that everyone kind of agrees that the market is changing, but how exactly is something that no one knows? It's so rapid right now. I see companies emerging and dying every day. All the products get replaced with new things. And then it's the best time to be alive, honestly. We're on the edge of it right now.

[00:03:08] [SPEAKER_00] We personally are mostly focused on sourcing part. Like how do you actually go and find the right person? Because I kind of believe that in this world where AI is generating like perfect resumes, as you said, how do you make sure that you're really getting the best candidates, right?

[00:03:26] [SPEAKER_00] Probably like outbound sourcing is my belief that that's the way to do things. But yeah, the matching part is hard because we have all the biases. It's something that we're navigating right now.

[00:03:41] [SPEAKER_01] Where's it going to end up? That's a fun question to tackle. I can guarantee you'll be wrong, but go ahead and answer.

[00:03:49] [SPEAKER_00] I agree. I agree with that. Well, look, I think everyone's talking about replacing recruiters, you know, AI recruiters, recruiters, all that stuff. I don't believe that's going to happen anytime soon. I think that people need people and you still want to talk with a real human being and their judgment and all that.

[00:04:11] [SPEAKER_00] So I think that in the next three to five years, one important thing that's going to change is that new types of recruitment agencies are going to be replacing older recruitment agencies by just applying more of technology, different kind of AI and all that. But replacing recruiters is not something that's happening in the next couple of years. That's what everyone expects, but I don't believe that's going to happen.

[00:04:39] [SPEAKER_02] You've mentioned matching a few times. Peter did as well. I want to dive into that because what I've seen from a lot of talent acquisition technology companies, including job boards, recruitment marketplaces, but also applicant tracking systems, is that the matching that they do is typically just comparing two different documents. That the resume or the resume or the resume or the resume or the resume from the candidate, which tends to be backward looking.

[00:05:08] [SPEAKER_02] This is what I have done. And the job posting or sometimes job description from the employer. This is what we want you to do in the future. Right. I had a demo with you and a member of your team several weeks ago, and it was pretty clear to me. Yes, that's part of the process, but you go well beyond that. Maybe you can talk about the well beyond that part.

[00:05:33] [SPEAKER_02] Like why is matching important to do not just from those two documents, but bringing in outside information as well?

[00:05:42] [SPEAKER_00] Yeah. Well, it's all about context, right? It's one of the things that we see all the time is on paper, the profile can look right. But if you dive deeper into the meaning of things, there is no real match. That happens all the time. And for technology, it's not that obvious. For experienced recruiter, it is, or for hiring managers.

[00:06:07] [SPEAKER_00] For machines, it's not. People might be overqualified, but on paper, all the criteria are met, right? And then like the machine says, perfect candidate. And you're like, we'll never hire this guy. It's just impossible for us to hire this guy. Okay. They're probably expecting something way bigger than what we have, say, in our company as a small company, right? For those reasons, we're trying to dive, not just us. I think it's like what happens on the market right now, right?

[00:06:36] [SPEAKER_00] You start with some basic matching thing, and then you get feedback from your customers like, hey, looks good, but this is not what we need, right? And that's where you have to go deeper and deeper and deeper. And the final idea is, how do you actually predict that this person is going to be a great fit for this company, right? Culturally, they're going to be a good part of it.

[00:07:01] [SPEAKER_00] And that's something that has not been solved by anyone, you know? Right now, most of the companies are trying to cut the first layers of the most boring work. You know, recruiters spend a lot of time on totally irrelevant resumes, but to make sure they're irrelevant, you have to look at them. Right now, the market is solving that first step.

[00:07:26] [SPEAKER_00] But eventually, we'll be able to, when I say we, I don't mean me or my company, I mean the market, right? Eventually, we'll find ways to actually create beautiful matches where everyone wins, you know? Because there are always great people and companies, if they meet the right time, they can build great things.

[00:07:51] [SPEAKER_00] But that doesn't happen because there is not enough good technology, you know? And that's why I think that people matching, like finding the right person is a lot more important than finding information. What, like, say, Google solved. Google did a great job with information. But people are creating things. People are the ones that move things forward. And finding the right matches of people is genuinely a huge mission for us as PCs.

[00:08:21] [SPEAKER_02] Quick follow-up, if it's okay, Peter. You did a pretty good job, I think, of sort of laying out your vision for sort of where matching is going. If you could, like, look at, like, one or two things that job boards, recruitment marketplaces, the ones that are doing matching today, what are they messing up with? What are their biggest mistakes?

[00:08:39] [SPEAKER_00] I don't think it's their mistakes. It's that things are changing fast right now. And there are so many companies that are messing things. Like, you know, now you can apply for a thousand jobs just by clicking one button, right? And every time you'll apply with the right adjusted version of your profile, which is not fair, right? But that's killing the whole process.

[00:09:04] [SPEAKER_00] Now, the companies get flooded by thousands of incoming applications that have nothing behind them. You know, that's not real. It's all pretty much fake.

[00:09:14] [SPEAKER_02] And they all look the same.

[00:09:16] [SPEAKER_00] Yeah, yeah. Generated by the same models, right? And then you're starting, like, hey, we cannot now, we cannot, like, take a look into every of those profiles. Let's use AI to analyze them and sort them. And then those AIs would do mistakes as well. And sometimes you would miss a really good candidate, right? So I think there is a lot of mess happening right now. Now, it's not because someone is doing the wrong thing.

[00:09:42] [SPEAKER_00] It's just the technology lets people, you know, do things. And it's just a phase. You know, when you just created electricity, it would kill more people than today, right? Like there would be some regular devices that would be killing people. And that's what happens with AI right now. It's messing with regular processes. We'll fix those in a year or two.

[00:10:06] [SPEAKER_01] And there's your perfect opportunity. Tell us about what Perch is doing. P-E-A-R-C-H dot A-I for people who are listening. There's your opportunity to tell us how Perch is doing everything right to fix the world. That's what we do. Okay. Thanks. We'll move on to the next question.

[00:10:31] [SPEAKER_00] So Perch stands for PeopleSearch. And we've built a technology for candidate sourcing, outbound candidate sourcing. The idea is that the best candidates are out there. You only need to find them. And Google, as I said, did a great job with information. But no one has solved the PeopleSearch, you know. So whenever you have some idea of a great person you'd like to find, they do exist out there. Our question is, how do we find them?

[00:11:01] [SPEAKER_00] And we're building right that. We're a search technology company that's collecting a lot of publicly available information about people and companies. And then we've built a technology that's delivering, like surfacing highly relevant people out of hundreds of millions of profiles that we found out there for very nuanced natural language searches that our customers provide to us.

[00:11:27] [SPEAKER_00] And the way we use this and the way we help the market with this is that we're not selling this solution to recruiters. We're not selling it to end customers that need to hire. We're only white labeling this technology to HR tech platforms, applicant tracking systems, AI recruiters, all of the different HR platforms. They plug us in, and now they have a high quality candidate sourcing as part of their products.

[00:11:57] [SPEAKER_00] That's what we do. And who are some of your clients right now? So it's always white label. We're not mentioning any of them. There are over 140 HR tech platforms now that use us as their candidate sourcing back end. But it's all white labeled.

[00:12:12] [SPEAKER_01] And does that include job boards, recruitment marketplaces?

[00:12:16] [SPEAKER_00] Yes, yes. There are a few job boards as well. Yeah.

[00:12:19] [SPEAKER_01] Just a few you want more, right?

[00:12:21] [SPEAKER_00] I mean, we're not proactively going there and trying to sell this. We're mostly trying to partner with the platforms that are the right fit for this technology. If it's part of their strategy that they want to deliver high quality candidates and match them with the jobs, we're always open to those conversations.

[00:13:10] [SPEAKER_02] Yeah. Like why this candidate ranked higher than that candidate? How much of a better fit? None of that on the surface is all that different from what I've seen at other job boards. And that's not a criticism. It's a compliment because the learning curve is then lower. It's not hard then for people to understand how to do that searching.

[00:13:32] [SPEAKER_02] But the results, the different twist to me is that the buyer, the recruiter or whoever, then gets a mailing list. And so, you know, if you've got 12 matches, you get the contact information and the resumes, et cetera, for those 12 people. You can put those into our CRM. You can email each of the 12. And that's a bit of a different twist. And that works at scale, too.

[00:14:02] [SPEAKER_02] It's not just those purple squirrels. If you're hiring five field salespeople and you need 200 people to reach out to, you can do that. So that's interesting. That was interesting to me.

[00:14:15] [SPEAKER_02] Prachik, when we were talking before, weeks back, one of the things that I was unclear of whether you're doing, whether you want to be doing it, whether you think it's a bad idea is, does your system kind of learn that recruiter Cindy, when she types in this, she's actually looking for X?

[00:14:41] [SPEAKER_02] Because Jeff, who sits beside her, when he types in the same thing, he tends to look at different applicants. So does your system over time kind of learn what the recruiters are actually looking for and then adjust its results?

[00:14:58] [SPEAKER_00] Yeah. So that's a deep question. That's definitely something we want to build. We're not doing that right now. The reason is that, as I said, we're providing this to HR tech platforms, right? Right. So in most of the cases, we don't really know who exactly is requesting the search for us, right? We see it as an incoming request from the platform. Like the ATS itself says, find me 25 people like this. And we do that for them, right? But that's one of the things we want to add.

[00:15:26] [SPEAKER_00] So they'll be able to send us specific signals about who exactly is this, or at least some unique ID of who's searching so that we can learn and make things better for a specific person.

[00:15:40] [SPEAKER_02] Yeah. And Peter, like when you search Google using exactly the same keywords that I use, you're going to see different results than I see. And that's kind of where I was thinking is like Google learned a long time ago that personalized results at scale is something that is very helpful to the user. Search Google? Does anyone do that anymore?

[00:16:02] [SPEAKER_01] I thought you'd just ask Claude or chat GPT. I mean, search Google? Come on.

[00:16:08] [SPEAKER_02] Google's going to figure it out one of these years. They're going to figure out how to make money. I'm pretty confident.

[00:16:13] [SPEAKER_01] Tell us a little bit more about Perch. Funding, how many people, where are you based, that kind of stuff. I saw in your resume that you had a Y Combinator background. Are they behind Perch or are they behind one of your earlier companies?

[00:16:28] [SPEAKER_00] Yeah, YC is not behind Perch. With YC, once you're in, you're in. I mean, I'm part of the community forever. They did not fund Perch. Perch is an early stage company. We raised a little over a million dollars of pre-seed money. Got to break even pretty quickly in like seven months after the launch. We got to break even. So haven't raised more money now.

[00:16:58] [SPEAKER_00] Feeling pretty confident with this. I probably will raise some money next year. We're a team of six. And that's the amazing part. With a team of six, we're able to do things today. I've been building tech companies for 25 years now. With a team of six, I think like 10 years before, we would need like 60 people team for what we do now. A lot of AI, a lot of AI employees do things for us.

[00:17:27] [SPEAKER_00] Today, this night, one thing that happened was one of our customers sent us a report. Something went wrong with their experience. Our AI employee read that, found the problem, fixed it, and reported back while the whole team was sleeping. We fixed a bug for our customer and reported it. And that's what's happening right now.

[00:17:52] [SPEAKER_01] The white label aspect of it, if you can't name specific companies, give us some sense. Are they the biggest ATS systems? Are they the biggest middle size? Give us some sense of who you're working with.

[00:18:09] [SPEAKER_00] So all sizes are there. We have a bunch of small companies. In terms of larger ones, the ATS, I cannot mention them, but the ATS that everyone's considering the best ATS on the market today is one of our new customers. There's a company everyone again knows. They have hundreds of thousands of SMB customers in the U.S. And they are piloting with us as well right now. Started a few weeks back.

[00:18:39] [SPEAKER_00] And there are a bunch of smaller companies that just started, just got their first funding with a team of three companies like that as well. And while we're in the U.S., and I did not answer that part of your question, by the way. I live in California, in Palo Alto. We have team members in different countries. So the whole team is remote. But we gather from time to time to spend some time together. Right now, I'm on one of those offsides with the team.

[00:19:06] [SPEAKER_00] And we have customers from Middle East and India and Brazil and all of the countries. Canada, a bunch of European countries. So it's a global thing. Our whole philosophy with White Label is that we know that candidate sourcing is really tough. No one has been able to build good candidate sourcing technology before. And ours is not great as well. On my personal meter, it would be 5 out of 10 search.

[00:19:35] [SPEAKER_00] That's where the search quality is for me personally. Right now, it does outperform other search tools, but it's not what we need on this market. We need a really high-quality search. And it's not there. The White Label helps us help the right teams build really good products for the end market. That's what we do. If we were to go and sell this to end customers, we would do this impact.

[00:20:01] [SPEAKER_00] By partnering with the best companies out there that are really great with the rest of the HR tech, we're focusing on one small but really hard problem while our partners build the rest. And together, we finally are able to deliver really, really high-quality results to end customers.

[00:20:19] [SPEAKER_02] If a job board were to partner with Perch, what guarantees does the job board have that a year later, two years later, that Perch wouldn't just basically go directly to that customer, to that employer, and cut out the job board? You know, there have definitely been some organizations in our industry that kind of built their business that way. We might as well be upfront about it. Indeed.

[00:20:46] [SPEAKER_02] Indeed, we can be upfront about it. But they were pretty transparent. We partnered with them, you know, way back when. And they always were upfront and said, you know, as long as the ride is good, we should all enjoy the ride. And when the ride no longer makes sense, then we'll end it. They were never, at least with us, anything but fully transparent that at some point, they are going to go around us and go directly to the customers, to the employers.

[00:21:16] [SPEAKER_02] So we went into it with our eyes wide open. I'm wondering about your partners and what guarantees they may have maybe done, that you won't restrict yourself to white labeling, but that you a year from now, three years from now, you might be a direct solution for them.

[00:21:31] [SPEAKER_00] Well, they have no guarantee, obviously, for that. It's just not part of our mission. You know, they have to trust us. That's pretty much it. Well, I think that the philosophy we selected works really well for us. You know, we're not competing with anyone on the market. The whole market sees us as someone that can add value and not steal customers. That's helping a lot.

[00:21:59] [SPEAKER_00] We're getting all this, like most of the sales we have, it's inbound because of that. No one sees us as a threat. And this also helps us stay focused. You know, one thing that I learned in Y Combinator is that in order to deliver something really great, you need to stay focused on a really narrow problem. And once you start serving end customers, they'll be asking things that so many companies have done before.

[00:22:27] [SPEAKER_00] Like, do you have integrations with this and that ATS? Can you do this kind of outreach to candidates? All those things. It's been done by so many companies. I don't need to do the same thing once again. You know, I want to stay focused on the problem, solve it, and help the whole market. I think we'll make a lot more money over time with this philosophy by not selling to end customers.

[00:22:52] [SPEAKER_00] We are like what Stripe or AWS, Amazon does with, you know, you are the infrastructure layer. So many companies rely on you and Amazon doesn't need to build all those products on there. They just provide the platform. Or Stripe, right? You just want to deliver the best platform. And the rest will be done by teams that know what they do. We're that. We're our infrastructure in our DNA, I would say.

[00:23:22] [SPEAKER_00] And our partners seem to trust that it's not a mind game. You know, we really do believe in this mission.

[00:23:30] [SPEAKER_01] So you're the best there is at matching, of course. And you're a five on a scale of one to ten. Here's a tough one. What does it take to get to ten on a scale of one to ten? And who and how are they going to do it in 90 seconds? I don't think that ten is possible.

[00:23:54] [SPEAKER_02] And we need the code for that. And it needs to be open source. Go. No, but seriously.

[00:24:01] [SPEAKER_01] I mean, it's a great question. If you're five on a one to ten and you believe you're the best or close, how do you get to ten on a scale of one to ten? And who's going to do that and how?

[00:24:16] [SPEAKER_00] Yeah, the reason I'm saying we're better is just because of the benchmarks. There are a bunch of benchmarks, ours and other companies. We are number one on all of the existing benchmarks. That's the reason I'm saying we're best. But the vision for ten, you know, if you think of, if you remember the times when Google, when we used to use Google and you had to go through pages. You remember Google's search result pages? You would like normally to find what you need, you would go to page two, page three, page seven.

[00:24:46] [SPEAKER_00] And they're like, ah, there it is. There's a good result, right? Google was already the best search engine on the market. But compared to what they do today or what large language models do today, it's like it's nothing, right? And that's where people search market is right now. I think that focus is the way to do it because we have all these great companies like Google, all these search technologies built, but they're not enough for people search.

[00:25:14] [SPEAKER_00] No one has invested enough attention and great minds into solving this very problem. Probably they think it's a small problem. I personally think it's a lot bigger problem than finding the height of Eiffel Tower. So focus, I don't know who's going to do that. But I think that the way to do that is great people with a mission and narrow, narrow focus.

[00:25:38] [SPEAKER_02] So two years from now or five years from now, if you're at sort of a, we just hit my 10 out of 10. How do you know? Like what would cause you to say we're now at 10?

[00:25:51] [SPEAKER_00] I'm not saying that ever. I think 10 is not something you can achieve. It's like whatever we do, I'm going to motivate the team to do better. But I think that getting to something like 8 out of 10 in two to three years from now is possible. Maybe faster because AI is doing things a lot faster now. Maybe it's going to take less than I think.

[00:26:17] [SPEAKER_00] But yeah, 8 out of 10 is going to be a product where you just describe the person you would like to find in your own words. And you find someone that is as interested in you as you are in them. And they are a really, really, really great match. And you're going to get like 10 people like that, that are already interested in what you do because we've learned a lot of context. That's what is 8 out of 10.

[00:26:45] [SPEAKER_01] Well, Hrashik Ajamiyan of Perch, P-E-A-R-C-H dot A-I. Thank you for joining us. Steven, thank you for running this, managing it, tapping all the keys to make all the screens come up and all that good stuff. And I hope to see both of you at RecBuzz in Barcelona, April 13th and 14th. The best conference for job boards there is.

[00:27:15] [SPEAKER_02] And in one of the world's best cities. Ah, got it. And just as a reminder for the audience, we're now going monthly on the second Thursday of each month. Like, comment, repost. If you think that's a good idea. If you think it's a bad idea, blame Peter. Cheers, all.