Build, Buy, or Partner
Build, buy, or partner? Fewer questions have dogged mankind as much as this one, especially when it comes to HR tech. That’s why Chad & Cheese added Textkernel’s VP of sales in North America, Chris Conrad, to get to the bottom of this pickle, live from the big stage at UNLEASH America in Las Vegas. Along the way, they break down the issue, visit the historical implications and shine a light for current and future generations to answer this puzzling question, once and for all.
PODCAST TRANSCRIPTION sponsored by:
Disability Solutions helps support and educate your workforce through disability awareness and inclusion training.
Intro: Hide your kids. Lock the doors. You're listening to HR's most dangerous podcast. Chad Sowash and Joel Cheesman are here to punch the recruiting industry right where it hurts. Complete with breaking news, brash opinion and loads of snark. Buckle up boys and girls. It's time for the Chad and Cheese Podcast.
Intro: In this episode of the Chad and Cheese we take to the main stage of UNLEASH America in Las Vegas with our friend Chris Conrad from Textkernel. Talking about should companies build, buy or partner? Check it out.
Jess: Chris Conrad, VP of sales for North America from Textkernel. He's a global... Textkernel is a global leader in AI-powered semantic technology for understanding, connecting and analyzing people and jobs better. We've talked about that on the consultant panel a little bit ago. Over 15 years of experience in recruitment, and recruitment technology. Chris brings a ton of knowledge and experience to this panel and to this conversation. He's gonna sit alongside Chad Sowash and Joel Cheesman, of course the co-founders and co-hosts of the Chad and Cheese Podcast. I learned a whole bunch of things when I read this intro earlier. I didn't know all of these things about Chad, so I'm gonna share them with you. Former infantry, army infantry drill sergeant. He cut his teeth in online recruitment back in '98 with an outfit called Online Career Center, before the next year it became Monster. Thank you.
Jess: And then of course he went on to build direct employers from the ground up. He steered recruit military toward revenue as their Chief Experience officer, and built Randstad's first military veteran recruitment program. I didn't even know all of this about Chad. That's an amazing bio.
Jess: Joel Cheeseman has over 20 years of experience in online recruitment as well. Worked with both international and local job boards in the late '90s and early 2000s. In 2005, Joel founded HR SEO, a search engine marketing company for HR and launched an award-winning industry blog called Cheesehead. Let's bring them to stage.
Chad: My turn?
Jess: Amazing. Yes, absolutely.
Chad: Hello, hello, hello. Is this thing on? Is this thing on?
Jess: Good to see you.
Jess: Have fun. Yes.
Chad: Wake up, Vegas.
Joel: I know that dude in the back. They said can you and Chad get on the couch together? I don't know what they're expecting at Unleash to happen with that.
Joel: And those lights are bright.
Joel: It's very interesting vibe right now.
Chris Conrad: So what are we we're talking big back better. Is that, is that...
Chad: Build Back better?
Joel: Build back Better, big back Better, love it. That was at my house.
Joel: Immediately moving over. Build...
Joel: Buy or partner.
Chad: Yes, exactly.
Joel: I think that's the topic, right? The hot topic, yeah.
Chad: Which is... Yeah. Which is what the other group tried to stay away from the entire time, and everybody kept asking. So we might as well go ahead and and knock it out.
Joel: So, I feel like we should really simplify it, what those terms mean. So, build, do you have your tech team build this thing from the ground up?
Joel: Do you buy? Do you acquire somebody that's already doing it and plug it into your system?
Chad: Really questions for the ages, to be quite frank.
Joel: Or partner.
Joel: Yes, partner. Someone does it really well. You know that you are not... I don't know. You're not named Google or AWS or something and you just go buy that and partner with them to do it. So, I think, historically if we take it back there, there's been some mistakes that companies in our space have made. One that comes to my mind is Jobbing Video. I don't know if you remember Jobbing Video.
Chad: Who was around long enough to remember Jobbing Video?
Joel: Who remembers Jobbing? Oh, you're able. We got one.
Chad: Jesus Christ.
Joel: Well, Jobbing was a job board, headquartered out in the southwest. And in the mid 2000s, YouTube was a thing, google Video came out. And Jobbing created Jobbing Video where they thought, "We can do a video player as good as YouTube.
Chris Conrad: Yeah.
Joel: Turned out they couldn't.
Chad: For about five minutes. About five minutes.
Joel: They couldn't. It cost too much. Hosting it was too much. They couldn't do HD, they couldn't do the the things that that YouTube did. So they went back to YouTube. So that's one example of, they should have just partnered with someone like YouTube to start with.
Chad: But sometimes you're afraid to do those things because there might be an API, I don't know. Twitter might come out with an API and you use the API and you start to build revenue off of the API. And then they pull the plug on the API.
Joel: Yeah. Yeah. You were a career builder. What did Monster have on Facebook back in the day? You remember that? Remember, Be Known.
Chris Conrad: Oh, my God.
Joel: Be Known, anybody remember Be Known?
Chris Conrad: I try to forget about that...
Joel: I know. Yeah. I heard you, so...
Chad: Everybody was doing social media, Facebook.
Joel: Be Known, and they had Branch Out. Branch Out was a startup that was on Facebook and they claimed to have a billion users. Which is...
Chris Conrad: How do you remember all of this stuff?
Joel: This is what I do.
Chad: It's our job.
Joel: I have no life. So, we have a billion users. And then Facebook said, "You know what? We're gonna switch this up." And they're dead. And so is Monster's Be Known.
Chris Conrad: I'm pretty sure that they'll be known in my office somewhere.
Joel: So, there are significant threats. A guy named Elon just bought Twitter. Now you have to pay for that API. So all these companies that were using Twitter's API are now proper, you know what, in using that.
Joel: So, in our space we have to ask the question "Build, buy or partner?" My man here at Textkernel, a lot of you know Sovereign, they acquired Sovereign a couple years ago. You have tons of partners leveraging your program. And I assume you see this question with your partners. How do you convince them not to be scared that you're not gonna go away, that they shouldn't build it themselves? What's that conversation like with companies?
Chris Conrad: Surprisingly easy sometimes.
Chris Conrad: Especially when you start talking about R and D right? Yeah. I think you know a lot of the audience here are actually customers of our Sovereign product line. Within HR Tech we have a high level usage. I think probably maybe over a thousand different partners in North America alone.
Joel: Yeah. And if they're not using you directly, they're probably using someone that's using you guys and saying that they built it themselves. Right?
Chris Conrad: That is a phenomenon.
Joel: That happens. Yeah.
Chris Conrad: We've ran into that before. But, yeah, so the thing is, for maybe a lot of you guys, what we do is fundamentally resume parsing. So, I think something like two and a half billion resumes get parsed through Sovereign and Textkernel.
Chad: Two a half bi-bi-billion?
Chris Conrad: Yeah.
Chad: Is that a year?
Chris Conrad: Yeah, per year.
Chad: Holy shit.
Chris Conrad: It's insane. Yeah. All unique people, obviously. All unique. Yeah.
Chris Conrad: No, but it's crazy stats, right. And so, yeah, we do a lot of this and we've built up a lot of energy and expertise, but the challenge is who's building a new parser? Is it anyone's core business model to build that? Even if you're a job board, do you even want to have that proprietary tech?
Joel: And we're talking how many years between Sovereign and you guys, have you guys been doing this?
Chris Conrad: They both started 20 years ago.
Chris Conrad: And that's crazy. I'm not aware of another parser that's been came up in the last five, really. But, really, I think any more it's less about just the parsing, it's more about the enrichment of skills, professions, it's about all the other stuff. And what's kinda crazy... And this, I feel like I'm getting in a parsing tangent, this is weird. It's like all the creativity that people have on resumes, the advice that your mom would give you.
Chris Conrad: Like on your first game job. Like, "Go stand out."
Chad: Yeah. Yeah.
Chris Conrad: And so then people started making these really creative resumes.
Chad: And you guys strip the creativeness out.
Chris Conrad: And it's horrible for...
Chris Conrad: So you have to do a lot of work, be like, alright well... Because people are putting addresses all over the place or they have headers where their footers should be, and it's all over the...
Joel: And doesn't LinkedIn change things up pretty regularly to keep you guys on your toes. Just keeping up with LinkedIn is a total pain in the ass, right?
Chris Conrad: Yeah. Yeah. They changed style, like what? I feel quarterly now.
Joel: Yeah. That's a whack-a-mole that no one wants to play.
Chris Conrad: Yeah. No, but... And so that's a thing, right? So when we're talking about all the different technologies that are coming out, It's hard to keep up with the Joneses. So I think when we're all talking about different vendors in this segment, it's what can we differentiate? What can we special in? What can we bring a unique solution to the market? But really almost all of us have to interact with either job data or candidate data. And so you start having the question, "What needs to be proprietary technology for your business to be unique in the marketplace?" And surprisingly, and especially I feel like the conversation's changing daily now, it's limiting. Like, what makes you unique is sometimes a very small set of skills, very small set of capabilities that your tool might have or problems that you're solving with your solution.
Joel: So, let's... We don't have a lot of time on stage. Do we open up the chatGPT Pandora's box yet, or no?
Chad: Open up the can of worms. Because we...
Chris Conrad: We haven't talked about that, right?
Joel: No one has talked about this yet. This is brand new for everybody, I know. ChatGPT, in case you don't know it, go ahead and Google it. Is this a conversation? 'Cause I fear a little bit like we're moving really fast to adopt this, to integrate it, without sort of thinking about it. And some of you may know...
Chad: You feel that way?
Joel: So, Samsung.
Chris Conrad: Yeah.
Joel: Not job specific, but...
Chad: Well you've seen how many products pop out that are GPT in the last 2 weeks for God's sakes.
Joel: Oh, yeah.
Chad: It is ridiculous.
Joel: Wait till you go to the expo hall.
Chad: It is ridiculous.
Joel: Yeah, it's the new AI in big data over there. But at Samsung, some developers said, "Let's throw all our code into chat GPT and see what happens."
Chris Conrad: Why not, right? Why not.
Joel: Well, to their dismay now open AI has that code, and it was sort of an oopsie moment. But we have not just code and IP, but we're talking about candidates, human beings.
Chris Conrad: Yeah.
Joel: Databases of profiles. Should we be afraid about unleashing the chatGPT, the OpenAI animal on all this data? Should vendors be asking, should we, instead of why not?
Chris Conrad: We talking about people, right? What's the risk?
Chad: Yeah. I guess, yeah, the big question is cannibalizing your own current business and future business. So do we actually utilize the secret source that... We're talking about Samsung, right?
Chris Conrad: Yeah.
Chad: The secret source that nobody else has. It's our code, it's our data, what have you. Do we go ahead and allow a company like OpenAI or Google, or maybe soon, Amazon, to actually gobble in this?
Joel: Elon's got one too, x.ai. Everyone's gonna be flooding that with data.
Chris Conrad: Yeah.
Jess: So, pros and cons behind that. Because you guys, you've powered pretty much the entire industry between Sovereign and Textkernel, powered the entire industry and you know what white labeling and partnering looks like, but now this is changing the game entirely.
Joel: Yeah. Yeah.
Chad: So tell us about that.
Chris Conrad: Yeah. A lot of our company is PhDs, so. I'm a sales guy, but we're a small sales, and we don't need to be. We have huge R&D efforts comparatively for our company size. And so I'm talking to our head of R&D, and getting our position on large language models, which I feel right now should be a drinking game. Like every single time you say generative AI, you take drink.
Joel: Why Not?
Chad: Somebody get the Makers...
Chris Conrad: I got a bottle of Makers.
Chad: Yeah. Get a Maker's Mark out of my bag.
Chris Conrad: Let's get after it. No on will...
Joel: Backstage is crazy, everybody.
Chris Conrad: Soo, one of the anomalies, I guess, the smart people are talking about is the concept of hallucinations. Where the generative AI, which is naturally creative, will just make up stuff, which is great when you're asking it to write up some marketing content, maybe draft a prospecting email or make a job description. But when you're talking about something like, let's just say a resume, we're testing it out and we're like, "This is sometimes just coming up with just fake history."
Chad: So you're saying a hallucination for AI is pretty much them, the AI itself, not having enough data to fill a gap so they come up with their own, and it could just be total bullshit.
Chris Conrad: Could be. And I think how it's been explained to me by the Smarty pants is that it's like when you're having a conversation with someone, but you anticipate what someone is gonna say next, how often do you get that wrong a little bit? I know I kind of screw it up a bit. Well, it's like generative AI, it's kind of doing that. So it's making that assumption, it's drawing that conclusion, which sometimes... Most of the time it's right, sometimes it's not. And so when you're talking about doing this hundreds of times, thousands of times, billions of times, that level of inconsistency is horribly dangerous. And so you're talking about, like that's a challenge. So, what do you do with this? So there's things that it's really good at. And I personally believe we're at this moment where because of how accessible this technology is, I think it's gonna change a lot of things. But for the short term, we need to be able to augment what it's able to do, it's creativity with hard coded data sets, things that are industry specific.
Chris Conrad: So for example, we have extensive taxonomy and ontology around skills and professions, great. Use that to guide the AI, because that's similar to curation, right? So you're pointing in direction saying, "Hey, you build out this job description, but here are the skill that it should have. Don't guess, don't jump to some conclusions." Like, here's what we want on here.
Chad: So you're talking about, really... And we've talked to a couple of AI companies that focused on large language models and then having a link to domain specific. And that's smaller, right? But it is really organization's secret sauce.
Chris Conrad: Yeah.
Chad: And again, your secret sauce being 2 billion resumes that have been parsed, the ontologies, the taxonomies, all those things. That large language model does not currently have that information. How long do you think it's gonna take for them to actually pull that data in? Or do you think that it's gonna take a company to point them in that direction to be able to start gleaning that data?
Chris Conrad: That's a good question. I suspect not long. So we're already looking really heavily into how we incorporate, we're incorporating it in certain areas. And our main focus is around compliance, obviously, privacy, and what not. So there's a lot of concern, like what kind of data are you pumping into? Is this gonna be compliant with European legislation? Definitely not. But we know American legislation's gonna change. New York City has that weird law in the books that no one can seem to figure out what to do with it. But you know what, California's not far. Colorado's pretty close. This is all gonna change. Again, I kind of... We were talking about this over lunch, I see this as like the Model T, where traffic lights didn't exist before. The cars were around before.
Chris Conrad: But it's just the prevalent... The amount of 'em that started hitting the roadways. You had to put rules in place. You had to have, which side of the road are you gonna be on, what are stop signs? How are they gonna interact with horses? I think this is how we're gonna have in the future to be like, "Hey, How do we use this? How do we regulate it? How does it manage our environment?" But, that's beyond my scope too.
Joel: For a long time we talked about automation, AI replacing recruiters. And I feel like that has come full circle to it's gonna augment what a recruiter does.
Chris Conrad: Yeah.
Joel: Does Chat GPT have the ability to write job descriptions, to write rejection letters and cover letters, and does this replace the recruiter or does it augment them to do their jobs better, or focus on more important things?
Chris Conrad: Ooh, that's a dangerous question.
Joel: It' what I do.
Chris Conrad: We've talking about this for years. In terms of, obviously not with generative... Of this type of technology, but with different workflow automations, conversational AI, right? I don't think we're at that point, anywhere close to it, because at the end of the day, it's human connection. People wanna talk to people, you can read the room if you're gonna give someone some bad news. You need to have that conversation. But also it's kind of like if you... So my wife's a recruiter and she was singing an offer to someone, so she had the phone call with that individual, of course they wanted more money. So, how do you have that conversation with someone? I think where you're gonna use like a AI generated, voice of Bruce Willis, to communicate, like this stuff, is that where we're going?
Chad: They can definitely do that. Yeah.
Chris Conrad: There's also the rights. Why not?
Joel: There's also the flip side, which is less controllable, which is the job seeker.
Chris Conrad: Yeah.
Joel: And we're starting to see stories about smart developers that are applying to jobs that aren't them, and they're passing pre-screening questions, and they're passing conversational AI discussions, and they're getting job interviews.
Chris Conrad: Yeah.
Joel: So we're building a beast on both sides.
Chris Conrad: Yeah.
Joel: I always joke about eventually robots are just gonna interview robots until that first day of work, no one's gonna see anybody. But I think the job seeker side of it is worth...
Chris Conrad: Yeah. Wasn't, bill Gates given an interview where he's talking about where his test for whether open AI was ready was whether it could pass AP Bio exam.
Chris Conrad: And he thought it would take like two years from the do. And I guess they call him back in a couple months being like...
Chris Conrad: We're ready to do this now.
Chris Conrad: Yeah. I think that we're there, there. But, again, it's I think the human interaction, we always, I think, discredit that a little bit. There's things you can do, obviously, in terms of matching, understanding different data points, different pieces of the puzzle. But at the end of the day, still being able to understand whether someone's the right fit for your company culture, are we going to rely on that to generative AI?
Chad: I think that's the answer. For the most part is because how long have we given job seekers a shitty experience anyway, right?
Chris Conrad: I mean.
Chad: They've gone into a black hole. And they just want some type of touch, some type of love, some type of, communication. They don't care where it comes from, just as long as they know whether they got the job, they're going through to the next... To the interview or what have you. I think when we're talking about recruiters, they're gonna be able to provide... When they utilize this tech in companies who smartly build this tech or partner to build this tech, they will provide recruiters with the opportunity to have more time that they don't have today 'cause they're doing so much administrative shit that takes up all their time and they don't get a chance to actually be human and talk to humans. So AI could, in this sense, help recruiters be more human because it gives more time back to them to be able to have that conversation about pre-interview, post-interview, offer, et cetera, et cetera.
Chris Conrad: Yeah. And kind of switch gears a little bit, think about like... Also if you're starting up a new tech company. Right now we have all the different building blocks, we're starting to get to that point for a company to really build unique solutions, right? So they could... You're leveraging the public cloud to store all your data. You got processing power from, let's just say using Generative AI, you're buying that processing power. You can leverage companies like us to process the initial data, building out your indexes, build out your data set. So you have all the different components. So you can actually deliver solutions, I think, in the future, a lot faster to market... And with less headcounts, but you have to still have that unique solution, that unique idea, that differentiation. You can't just churn out like a copy of a copy.
Chad: Is it just the idea?
Joel: Is that a good thing though?
Chad: I mean, no. Well, first is it just the idea that now you can take pieces and parts and APIs and large language models and start to build your own, not even with your own tech. I mean, you don't even have to have a layer of your own tech in there other than bleeding all those together and having UI, right?
Chris Conrad: Yeah. Yeah.
Chad: So how close are we to that today?
Chris Conrad: I think we're close. So I got a good buddy of mine and he runs a sales org, or a small startup, that does a little bit of conversational AI, a niche player. And so I was talking to him, I was like, "All right, so tell me more about the tech." And none of it's proprietary, none of it. So, which is... This is interesting, and they're doing great. So what's proprietary obviously is their marketing campaign, their ability to sell and position within like the niche segmentation that they do. And they're really good at the staffing industry. So they have the local knowledge of how to apply the technology to that, into solving those problems. So they're doing great, but none of it truly is their unique IP.
Chad: So therefore, from an acquisition standpoint, that's more than likely not gonna happen. But they've got a great lifestyle company.
Chris Conrad: Yeah. Well, they're making money.
Joel: Let's explore this for a second. So when we got started in this, we'd go to Sherm, there'd be like 10 huge booths and a few 20 by 20s. And now we have like the startup alley where people have kiosks and these startups. And there're fewer big booths. Are we going to a place where just it's one big kiosk mall and every company is like 10 people or less, and 80 companies do the same thing? And what's gonna differentiate companies and if no one's getting a $100 million, everyone's getting 2 million in funding, and that they can make a business. I just see this going into a real confusing, convoluted space. Am I wrong there?
Chad: More than what it is now?
Chris Conrad: Who wants to be in that conference room?
Joel: Way worse. Can you imagine walking into HR tech and it's all kiosks. Like think of the worst mall ever.
Chad: I can't think anything worse than walking into HR tech expo.
Chris Conrad: It'll be the eightfold booth, 200 by 200, and then a bunch of kiosks. That's what HR Tech will be like...
Joel: Oh, and just like a whole bunch of singular sales reps with hungry eyes, asking you to come on by.
Chris Conrad: I don't think that's good for the industry. Is it?
Joel: I don't think so at all, right?
Chris Conrad: Is that where it's going?
Joel: I'm not against startups, obviously. But I think you still need to differentiate yourself. And the scenario you're talking about is like cookie cutter. Where you're not differentiating yourself. I don't think that's where the market's going. I think we're just talking about the ability to bring new products to market faster, and differentiate.
Chad: Well, in today's environment, there are still companies that utilise, as you'd said, tech not theirs, building companies for... I mean, it's happening today. The question is how much faster can this happen when you have these types of capabilities with organisations like Microsoft, Google, Amazon, Elon's dumbass...
Joel: The ability to differentiate technologically, to me is going to a place that's almost impossible. The only differentiator is brand and maybe I have money to spend on trade shows and commercials. But if everyone's using open AI or Barred or whatever, and if those all become commodities, it's really hard to see a world where you really have something unique.
Chris Conrad: Yeah. I think you're describing...
Joel: Unless you have 20 years of experience and data.
Chris Conrad: Well, obviously something to be said about having industry experience or niche industry experience, but I think you can't discredit human creativity. Whether we're talking about any of this stuff large language models are disrupting. This has all happened before. Human ingenuity. We're gonna always solve problems and this is gonna create a problem that we're probably gonna solve in another way. So I think we're just... The tool set is changing rapidly, but not the application or the creativity.
Chad: So in the future. Build, buy, partner? Is this even...
Joel: Are we doing buy? Buy or sell?
Chad: Yeah. I mean, is this even going to be...
Joel: Yes. I think build is a big sell.
Chad: Yeah. I mean, is build even gonna be a part of it when it comes to technology?
Joel: Unless you wanna be a two person company.
Chris Conrad: Yeah.
Joel: I guess.
Chris Conrad: I think build is... I think it's important if it's gonna be unique and it's gonna be a differentiator. But so much if it's not, if it's not part of your core business model, if it's not something that will set you apart in the marketplace, you've gotta partner every single time. Because you can be faster, more efficient, more economical, and then focus on what you can deliver, what you're in the business to do. Are we selling, buy?
Chad: Buy, definitely.
Chris Conrad: If buy becomes there's a dozen big companies with a lot of money and then a bunch of kiosks, at some point, all those kiosks get bought up by the big companies. So I think in that perspective, there will be a lot of consolidation from this. But ultimately, partnering, if you can be that trusted source of data or technology, and you're not gonna go anywhere, that seems like that is the best answer today.
Joel: I mean, that's the bet we're making.
Chad: At least from this stage. And unfortunately, the big screen says time is up.
Chris Conrad: Take a question?
Joel: Does anyone have a question?
Jess: Ooh, he's got a question. Louder.
Outro: All the big providers have massive marketplaces, and different partners, obviously. And they seem to be growing and growing every single day. Don't you feel a lot of the startups that are in those marketplaces are getting washed out, so it's hard for the actual customers of those companies...
Joel: So the question is around marketplaces. So most of you know, your ATS or big platform, you can build on top of that. And then their customers can access those products and services. So the question is, do those just get washed out?