When you want insights to big industry moves you go to the insiders.
In this episode The Chad & Cheese are joined by phenom, Death Match Europe winner, and former CEO of newly acquired Opening.io, Andreea Wade. Andreea dragged along iCIMS' technical commander-in-chief Al "Witness Protection" Smith to discuss how Opening.io's team, vision, and tech will play into iCIMS' future.
Enjoy this NEXXT exclusive.
PODACST TRANSCRIPTION sponsored by: Disability Solutions provides training and development to help your workplace leaders and employees integrate with and value people with disabilities.
James Ellis: Hey, this is James Ellis from The Talent Cast Podcast, and you're listening to The Chad and Cheese Podcast. So perhaps treat this message like an intervention. Why are you doing this to yourself? You have so much to live for, why would you waste your time here of all places?
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.
Joel: Andreea. Yo, what's up everybody? This is The Chad and Cheese Podcast, I'm your co-host Joel Cheesman.
Chad: And I'm Chad, I love matching, Sowash.
Joel: It's a Monday, so we're just waking up from a beautiful weekend, here in the Midwest. We are happy to be joined today by newly acquired, Opening.io founder, Andreea Wade, but currently potfolio director of AI and machine learning, I hope I got that right it's a new title, and iCIMS's CTO, Al Smith. Guys, girls, welcome to the show.
Andreea: Thank you. Thank you. Thank you.
Chad: So, Al's a part of the witness ... Just so we get into this real quick, Al's a part of the witness protection program, hence his name Al Smith, very simple, hard to find, there are so many of them, right? Yeah.
Al: It's actually based on what I'm capable of remembering, and it's a name I can't remember.
Joel: Speaking of preppers.
Chad: Joel just mentioned Andreea, that you are the AI, ML portfolio manager. Okay. So, what the hell does that mean? What do you do? Is this new? Is this something entirely new? Or has iCIMS had this for a while?
Andreea: So the role of portfolio director, it's something that iCIMS had for a while for various different areas of their business, and I know that I will be able to give a better answer in relation to that. My role, you kind of know what we did. We're very much, we always were an R&D labs company. So we want to continue to do more of that. We want to grow and build iCIMS Talent Logic. And then we want to further imagine the future. So we'll just keep R&Ding here, and doing loads of cool stuff.
Al: Yeah. Look, I'll piggyback on that. Our portfolio directors kind of own the business strategy and the investment strategy for particular space for us. And so broadly, we participate with solutions in a variety of different parts of the market. We ask the portfolio directors to understand how big that is, who's the competition, how much could we address, how much to invest and what would we bring to market when, and set out a vision of what that kind of value is, and the solution is. And then product managers, paired with an engineering manager, bring those products to market. So in many ways, consistent with Andreea, was doing as the co-founder for Opening. So, I feel like it's a pretty good fit.
Joel: A lot of our listeners don't know Opening, don't know the genesis of it, it was founded in 2015. And Andreea, you're a marketing media person, how in the hell did you get into this business, in machine learning and AI back in 2015?
Andreea: Yeah. No background in the industry, but we felt it. We were candidates and hiring managers and this was to become my fourth company, I guess. And I used to be a journalist, and I used to be a product manager, and I had a branding company and all these kinds of things. But there was one, I guess it ... Everything came together by my passion of just, without a focus and an agenda, of bringing good people together. I used to just, because I get involved and I got involved in a lot of the tech scene here, I ended up being the person who nearly knew everyone and I would just match people together. I would go, "What are you trying to do? Talk to that person," and so on. And 2015 comes, and I get very, very interested at that stage, I had another company. I get very interested in data and AI and ML. And I started talking to my cofounder, Adrian, who just moved here, in Ireland. I've been here 18 years and he moved here then. And as product people, as data people, as a process that he just went through at that stage and something that I felt as a person who was hiring people, we started looking at the industry. And we started playing with classical machine learning, it was weekends and afternoons. And just for about eight, nine months, we looked around at this industry that we're completely clueless around. But the vision from day one was surely there has to be a better way for someone to understand my CV.
Chad: So it is interesting because your background is not in this industry. Why do you think you succeeded where all of these other industry insiders generally really just have to eject?
Joel: Yeah. And by the way, the matching space, a little bit crowded.
Andreea: Yes. Yes. But we were naive and clueless and we had a vision. Right? And you don't ... If you want to start something and go, "Ah, these are all ..." You can very quickly find all the reasons why you shouldn't. Right? So we had questions. And I think the key, like when you were asking me this question, the key was that we had the questions and we asked as many people as possible from all sides of industry, talked to job boards, talked to ATS's, talked to CRM's, talked to recruitment agencies, talked to RPM's, talked to everyone who would give me half an hour, or be it Adrian, or whatever. We literally ... Because we felt we were so self conscious that we did not know, and maybe understand this industry, that we just continuously learn. And we asked everyone the same question 50 times. And I think that was really key because we try to understand that. And then what we did have was, we were pretty good product people. And we imagine things in ways that we saw good product being built, but not in this industry. Like literally good product being built.
Chad: Yeah. And I'd like to say, just what I've seen from you guys and watching you through this time frame, you had amazing focus as well. It didn't seem like you guys were looking to pivot five or six different times to try to "re-invent". So, your discipline seemed to be there where in most cases, most of those companies coming in and startups, they just don't have that discipline.
Joel: Yes. The company that they were during Death Match a year ago, is the same company that they are today. And I'll put a plug in there that she was a Death Match winner, because that's always awesome.
Chad: Oh yeah.
Joel: Andreea, you've spoken it a lot ... You have a tight relationship with Microsoft, I believe you're using their AI, their translation products and others, want to ask you about that relationship and what it's meant? And how that might segue with iCIMS's recent integration with Microsoft? And how maybe all of these worlds are colliding for a future acquisition?
Andreea: For us?
Joel: Yeah. I got that in there.
Andreea: I'm just going to talk from my perspective and then I'll let Al add to it. Look, for us it was really important, and I guess from day one ... We're in Ireland, and Ireland is small, and Microsoft had a really cool team on the ground that was talking to startups. And these two guys, I'm not even in my Opening role, I'm kind of mentoring at a hackathon and go away, and I get approached by these guys and they go, "We're here to support this hackathon and we're from Microsoft and we want to do all these things. And I know that you're here as a mentor, but I'm curious about what you're building." And long story short, we meet at different events and they go, "We really, really, really want to support you." And I have to say that the way that they have ... And at the same time, we were talking to various different other large corporates as well, that had these startup initiatives. But Microsoft felt real for many reasons. And it wasn't just, use this technology to do A, B and C, it was more, let us talk about you, what can we do for you? Let us put together a few videos. Let us support you with marketing. Let us sell you. All this, and it just felt very real.
Joel: Wow. Okay.
Andreea: And it just felt like they knew what they were doing and their approach was legit. We got to meet Peggy Johnson. Peggy Johnson who bought LinkedIn for Microsoft, about a couple of years ago, came to Ireland, because Microsoft has in Ireland, their first building that they own outside of the US. And Peggy Johnson came over and, our Taoiseach, the prime minister here in Ireland, our Taoiseach was there and they were talking. And then the MD of Microsoft here picked four AI startups to meet Peggy. And we had an hour with her. So we had various different supports that went outside of Ireland [Inaudible 00:09:52.13] and it was real and legit. And so coming together with iCIMS, I've learned, when I first met iCIMS in 2017, I think it was a week after I got to meet them in the States, was when they announced that Microsoft is a client. The announcements in the last month or so and our talks pre-acquisition, it just seemed that there was a common vision there as well with what we're trying to do. But I'll let Al explain more what iCIMS and Microsoft are doing. And we still have our network and we're working really hard to do more with them.
Al: Look, I'll take a swag at this. So I think you guys know Microsoft's been an important customer of ours for years, and we've been always trying to look for opportunities and when our customers have parts of their business that overlap or channel to try to go to market together. And that's why the dynamics, acquisition or integration made a lot of sense, and working with them. And you guys also know that we've been on a hunt to make sure that our talent solutions integrate with all the best HCM's, and also to make sure that we're a platform of choice as the HCM to actually address customers that have serious talent needs. So that that's kind of enough to think about it. But as Andreea said, when I first met Opening and learned about how they've approached the problem set ... I think both of you, we've probably had a couple of meetings over the last couple of years, talking about AI and ML, and I've been a pretty big fan of saying, "Look, it's an early market, let's not just go rush in." I want to get in to make sure that as we build things they're explainable, it's transparent, it's human led.
Al: The idea that it should help you with context to a set of decisions you're doing, yes, it can set up some automations for you, but it shouldn't be this black box just being here as the answer. One, there's all kinds of reasons why you may not have the right answer, number one. Number two, when we get into, and I think this is something you all know, iCIMS has always been really careful around data privacy and security and compliance, and the black box model just doesn't work. And when we met Andreea and Adrian, we just saw a company that had focused on building, not just a bunch of algorithms but a platform, and a platform that actually strove to deliver the explainability and the transparency around the decisions and going past just match. And that's some of the stuff that really got us excited thinking about the future. Match is great, don't get me wrong, but there's so much more past that.
Chad: We did notice the Freudian slip, you said acquisition there, Al. We're talking about Microsoft
SFX: Hell yeah.
Chad: Microsoft and iCIMS, just so that you knew that we caught that. Al, I'm a big fan of the prospect and we have talked about AI and ML over the years and matching. And I remember in Arizona at iCIMS Influence, late last year, I spoke with a few of iCIMS leaders, and you guys talked about a "Jibe matching engine", what happened to that? This was interesting, and it was really great to see that you guys obviously going through the acquisition process with Opening, but were there a lot of, kind of skunkworks things happening at iCIMS? And then you just thought, hell, we just need to go out and buy this.
Al: Actually, it's a continuity, right? So Opening had actually landed Jibe as a customer, and part of a Jibe matching engine used some of Opening's technology. And that was actually how we got the introduction. So the good news is the continuity goes all the way through. Now, there's other technology that sits around that we do believe that right now, the way these different models get developed and different algorithms get trained, that just relying on a single solution, doesn't always give you the best outcome. And so we've taken this kind of ensemble strategy on search/match problems, where we feed a set of data into three different models simultaneously, one of those obviously is Opening, and then look at how well the model gives you a result set and serve up the best fit for the situation.
Chad: It's like a bake-off, right?
Al: Almost. The one thing I think that got us really excited about Opening's technology is, they do a fantastic job when there's sparse data. I have just a little bit of data, maybe it's from an application, they're only bits of data, I don't have a full CV or something else. Some of the other technology models that we use can do a good job when they're doing their lower level parsing, when there is a lot of data. But, so many of the industries iCIMS serves, that data isn't rich and deep about the candidates that you're trying to find. And we love the balance that that brings to a full solution. Always gets us excited. And then by the way, you guys, I think this is very consistent in what we discussed last fall.
Announcer: We'll get back to the interview in a minute, but first we have a question for Andy Katz, COO of Nexxt.
Andy Katz: Every Fortune 1000 company have to, anybody with extreme volume of jobs. You're recruiting for 20 positions a year, you don't need programmatic. You can go to a recruitment marketing agency or a job board and do a direct email with your company only you're not in with another 20 companies in a job alert, or you're not just on a career site or a job board. You can do banner advertising, buy premium placements. So where programmatic again, is one piece of the puzzle, it's not going to ever be the end or be all. And I do believe all the programmatic platforms out there have ancillary services to support that, knowing that you can't just survive on a one trick pony.
Announcer: For more information, go to hiring.nexxt.com. Remember, that's Nexxt with the double X, not the triple X. hiring.nexxt.com
Joel: Curious question for both of you, I guess, brand architecture, when you guys typically acquire a company, whether it be TextRecruit, Jibe, et cetera, they tend to be sort of standalone brands for a while, and then they eventually get sort of sucked in, I guess, to the iCIMS brand. I noticed that Andreea is no longer associated with Opening in terms of her title. What's going to happen to the Opening brand? Are there integrations out there with other services that will be impacted by a brand change if there is going to be one? Talk about the future of the brand and how it's going to integrate with iCIMS, if at all.
Al: Great question. So a couple of things, one is, those different acquisitions that we've done have had different contexts. When we did TextRecruit, the actual plan was to leave them independent for a very long period of time. For a variety of reasons. One was, we were trying to explore how we bring customers into the portfolio in a standalone sale, and then cross sell all the products to them as a solution. With Jibe, it was a much quicker, the Jibe employees became iCIMS employees day one, and we moved pretty quickly to incorporate it into the portfolio as our recruitment marketing suite. I think with Opening, we're moving even faster. It's a different sale and what Andreea and Adrian were doing themselves to other people who were embedding their technology, we'll continue to support that and offer that. But right away, what we thought would add the most velocity was giving brand to the solution. And so we chose iCIMS Talent Logic as the brand name from day one, and literally announced the acquisition with the brand name of the product suite that the technology is delivering on. And we thought, honestly, it's part of us just learning how to do that better, and it kind of speeds the time for everybody to just be able to understand where it fits and how it brings value.
Chad: I want into jump into something that you said a little bit earlier, Al. Opening.io, or now iCIMS Talent Logic can do better matching with less information. And from the website, it says, "Accurately match candidates with relevant roles." I would assume that you both agree that job descriptions still suck. Dude, can I get a yes or amen from you?
Al: Yeah.
Andreea: Yes.
Chad: Okay.
SFX: Hell yeah.
Chad: So if you're working on garbage data, and it's garbage in, garbage out, how do you make something good out of that? Because we've been trying to get the industry to move toward better data, better job descriptions, and they still suck. And you have to churn off of that data, how do you make that work?
Al: I'll take a stab at it, then Andrea, maybe I'll hand off to you. But a couple of things that really appealed to us, and not to get too far out there, but I think you guys have been following our industry for a very long time, much longer than myself, it is very obvious. I joined the iCIMS five years ago and I was new to the industry, it largely looks like an automation of a paper process that relies on an awful lot of subjective decision making with very little context. And Andrea made a reference to the point that ... And Adrian's personal experience that kind of spurred this whole thing was, he applied it to a company, didn't get anywhere, months later went to a head hunter who placed him very highly in the company, and they hire him and they're delighted about it at their own cost, right? And it's the ultimate, hey, what's wrong with this process? Kind of thing.
Al: So we're going to try to use the technology that they've developed and we're building and continuing to invest in, around making a shift as opposed to a lousy job description to a very poorly fit and exact key words matches, and then the magic of somebody years of experience to have a feel about this candidate. Look, that can work at small scale and in certain industries, but on large scale, it falls apart. What we're really looking to do is help with better job descriptions. Help with understanding what does a good performer look like in this job today? What are the kinds of successes and derived experiences that they have? Use some of that when you're looking for people that, "Hey, I really like Chad. I really do. But I'd like four more Chads." I'm sorry, Joel. Yeah, I had to go there. If I was looking for four more Chads, what is it that I'm looking for when somebody comes to apply for something? So we see the ability to use the technology that Andreea and Adrian put together here as a start that we start going from candidate search to role fit, kind of moving, these are subtle statements, but moving from just a search, enterprise search to talent discovery. You may have the people you need right in your company. So a lot of exciting things that we're looking at what comes next.
Andreea: Yeah, absolutely. Now, if we can add here two things. We already have tech like skill extraction and skill recommendation technology. So run this job through the engine, we'll pick out three skills and we will tell you, "Why don't you add in another 10 or 15 or 20 skills," but also ... And we have this, right? And it's in production, and it can be used to enrich both resumes or poor job descriptions. But to go back to Al's point around our tech working on very small bits, pieces of data, this is because we don't just look at what you give us. We look at it in the context of everything else. So if you tell us, "I'm hiring for this," we will immediately link that with the universe of skills that we have. And we go, "Hey, this is what we think you're asking us." Right? And it's really, the example that I can give, as asking me, telling me that you have a headache and me telling you to take a pill or taking an aspirin. Or asking a doctor, telling a doctor that you have a headache, and she will immediately probably tell you 50 things that you might have because she studied for many, many years. And our engine has studied and understands when you say, "Blah," it goes, "I think it can mean all these things."
Andreea: But absolutely, really good question. And that's something that we learned as we were discovering this process. You can't search with this data, it means nothing. And we even went ... And I'll just say this, and then I'll stop talking about this. Actually, we put poorly put together job descriptions or very vague job ads or job descriptions back in front of recruiters that they were hiring for those roles, and we asked them, we hid the title, but we left everything else, and we said, "Tell us, what is this for?" And they couldn't tell us.
Chad: Big surprise. Right? So, Al, yes, obviously everybody wants four Chads. But the problem there is that humans are bias, and four Chads means I just hired four white dudes. Right? We talk about AI, we talk about bias, how can Talent Logic, or how can the new tech today, broad scope, actually focus on being unbiased when you're using all this biased human data? It's all historical, it's all things that we've done before, but it's all laden with bias. How do you get through that?
Al: Thanks for bringing this one up. This is so important to us. Probably never more so than what's going on in the world today. I think a couple things when we stand back and look at the problem, I personally think we're approaching it from the wrong direction as an industry. We keep talking, you just used the words, how do we avoid the bias? How do we avoid the bias? Look, there are some things we should be very careful with when we train algorithms with data, around what we think the likely outcome is going to be. Because these algorithms, at the end of the day are all designed for effectiveness and efficiency and do more of what you ask it to do. And so I think some of the big, very visible, very big blow ups around having things trying to match and having an unexpected bias shouldn't be a shocking because, whether it's in your data, whether it's in your process steps, whether it's in other things that were unintentionally included, I think this is what algorithms will always kind of bring to the surface.
Al: I want to kind of flip the thought on its head. I think when I go out and talk to CHRO's and the heads of talent, the first conversation is, if you're trying to build a winning workforce made up of diversity and inclusion, and every business study did says the more diverse, the more inclusive your workforce is, the more likely you're going to be a successful company. The first starting point is, how diverse is your workforce? And do you have the tool to help you figure that out? We think we have the basis of some of that capability of what the Opening team has brought us, to help you, first of all, assess that. We have a fair number of our customers who keep all their existing employees in the ATS as a function of, the best next employee might be in existing employee. And so helping understand how diverse your current workforce is, and then looking where you may be out of balance of what your goals might be. Can I use this technology to then go address the gaps, address the weaknesses and have it help me do that in a more automated fashion, so that the outcome I'm trying to get to is actually one of the model that I think fits the business, the company I'm trying to run? I think there, the technology can be excellent. And it's because you're using the converse, which is, go find me folks that match this.
Al: And again, with [Inaudible 00:26:13.19] data, we can do a really thorough, these folks might be those people, let's consider them. While in the past, you might not have, because it was a knockout. I know many of our customers talk about knockout questions. How about you think about the other way and what are the inclusive questions? And then how do you get a representative pool of what it is you think the outcome looks to be? So those are some of the ways we're looking to do in a differently and how we go forward.
Joel: This one's a for Andreea, I'll let you guys, let you out on this one. I'm going to assume that you had more than one company in the running to buy or acquire Opening. Just the fact that you won Death Match meant that you had a flood of suitors trying to buy the company.
Chad: Oh yeah. Easily.
Joel: What was it about iCIMS that really drew you to the organization? The people, the opportunity. Why iCIMS out of the mini suitors that I'm sure you had?
Andreea: Why iCIMS? Period. Well
Joel: And you're taking too long to answer.
Al: I can put my fingers in my ears.
Andreea: For many reasons. Right? And some of them are stories that I've told internally to our new iCIMS colleagues, in the first couple of weeks of us being part of the company. And it's a story that started, as we kind of touched on a few minutes ago, with us integrating with Jibe as well, with really liking Jibe and the people in Jibe. Just feeling that we've made new friends in the industry and good people and allies and partners that we can count on. When they, were acquired, we literally celebrated. And I actually don't think I've said this to Al or anyone, we were so happy for Jibe, and I've actually said to Adrian that, "Why are we so happy? Feels like we got acquired." So that was kind of ... And it was a road of, we saw a company that saw us. Because there was a lot of noise in the industry, and as you guys said as well, there were so many players with a matching. But we knew why we were good and we knew what we had, and we were waiting in a way for that home, for everything that we built. So we found a company that saw us. We found people that we really, really like. We saw that ambition and that really exciting, where is this goal and a goal next, which I am really excited about. And then we already had people in there that we knew and that we liked, and that liked us. And I guess, the vision and where the company is going next, then how, that was very, very exciting for both me and Adrian and also for the team.
Chad: Well, Andreea, I have to say that when you do visit iCIMS or they visit you, any of those friends, you'd better be wearing that chain of champions. Okay? Because that's ...
Andreea: Yes Chad.
Chad: Al and Andrea, thank you so much for giving us your time. If people want to find out more about iCIMS Talent, Logic, although AKA Opening.io, now iCIMS Talent Logic, where would you send them?
Al: Yeah. Real easy. It's www.icims.com, and you'll find it right on page.
Chad: Excellent guys.
Joel: And with that, we out.
Chad: We out.
Andreea: We out.
Outro: This has been The Chad and Cheese Podcast. Subscribe on iTunes, Google Play, or wherever you get your podcasts, so you don't miss a single show. And be sure to check out our sponsors because they make it all possible. For more visit chadcheese.com. Oh yeah, you're welcome.