And the winner is ... The competition for Death Match dominance was hard-fought, bloody and mind-blowing, but at the end of the day, only one company was left standing. Enter Seekout, whose founder, Anoop Gupta, brought his A Game to Austin and took home the championship chain. Enjoy this Alexander Mann exclusive.
PODCAST TRANSCRIPTION sponsored by:
Disability Solutions is your sourcing and recruiting partner for people with disabilities.
Chad: Welcome to Death Match North America 2019 part four of four the grand champion edition. This Chad and Cheese Death Match episode features Anoop Gupta founder and CEO of Seekout. Death Match took place at TA tech on September 26th, 2019 in Austin, Texas with a room full of TA tech practitioners, Chad and Cheese double fisted Micheladas while the judges took aim at their last target. Enjoy, right after a word from our sponsor.
Joel: Death Match is brought to you by Alexander Mann Solutions. Hiring great people is no easy feat. There are new obstacles around every corner and your competition for talent is intense. Together we need to be bold in our approach to tackling these challenges. Alexander Mann Solutions can be your bold next step with a team of nearly 5,000 professionals around the globe delivering market, leading recruitment, outsourcing, and talent consulting services. And in early 2020 AMS will unveil an exciting new digital solution that will disrupt how you connect with job seekers and hire the best fit candidates.
Joel: Now is the time to create purpose built solutions focused on solving your unique challenges when it comes to engaging and hiring the people your business depends on. To learn more about how Alexander Mann Solutions is working with talent acquisition professionals around the world. Visit alexandermannsolutions.com today.
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 a breaking news, brash opinion and loads of snark. Buckle up boys and girls, it's time for the Chad and Cheese Podcast.
Peter Weddle: All right. If you've not been to a TA tech event before, you are in for a memorable treat. You're not here to see me so without further ado, let me introduce Podcaster's extraordinary Chad and Cheese.
Chad: Raise that drink everybody. So thanks first and foremost, thanks Peter and TA tech or making sure that we had alcohol in the morning because a lot of us got in late. We need that. Yeah, keep the party going. So Death Match, who's seen a Death Match before? This is our third Death Match. There we go. So you know, that's why you're here, right? Death Match really a variation of firing squad. This is more the onstage live version. So obviously for of the podcast, check out firing squad every single one of these pitches today. Without further ado, today we have four startups. They will have two minutes to pitch. After those two minutes. We have a stunning panel of judges who includes Cindy Songne from Talroo, Quincy Valencia from Alexander Mann Solutions, woo-hoo give it up, and Robert Ruff Sovren Technologies everybody. All amazing sponsors and we love them.
Chad: All right, you can hear the music. Get ready. We have a new Anoop Gupta from SeekOut. This is the anticipation. He's building anticipation. There it is. Oh, there we go. Sucking up to the judges. Very good. Buttering up the judges. Looking good. Are you ready? All right Joel, are you ready?
Anoop: Ready to be here. I am Anoop Gupta, co-founder and CEO of SeekOut. We are a startup in the Seattle area. Every company is going through digitalization. They need to hire developers, engineers, scientists, first to simply just survive. If they're not going to go to the digital age, they're going to fail and not to thrive. Now the people that you're hiring here have some of the worst LinkedIn profiles. They don't list a lot of information, many of the profiles are simply missing, they don't even have the profiles, and also diversity is very important for this particular audience.
Anoop: And so LinkedIn, the product that everybody else uses simply doesn't suffice for this very critical piece of talent. So the solution that SeekOut is building. The first thing we do is we build aggregated whole-person profiles. So we not only take the LinkedIn information if it exists, we combine it with what they've done on GitHub, we take it for what papers, patents, all of the information, bring it together, and then we infer properties from that. It might be gender, you know ethnicity, it might be the coder score, it might be programming languages, frameworks they are familiar with.
Anoop: The second aspect is an intelligent search engine
because how you write your job description, what are equal in things, how do you find the same person on a get hub site? Okay, what is listed. So there's a very smart ML AI engine there to do the matching. The third big component is hyper-personalized messaging. You tell a developer you rock in Java and you think they're going to respond to you? Zero chance. Right, so we can hyper-personalize based on these rich profiles, what they do ... we are used by some of the largest companies in tech industry and pharma-
Anoop: And banks. Okay. Everywhere else. So ... thank you.
Joel: First, I think it's worth mentioning that you were once a direct hire to Bill Gates, so to see you come in in that fashion was very entertaining for me with that knowledge. There's some real questions about how you get LinkedIn data. There's some legal cases. How do you get access to LinkedIn? Are you paying LinkedIn to get it? Are you scraping? How are you getting that information from them?
Anoop: Basically what for use on the LinkedIn side is public profiles and there are many providers off the public profile data and what the legal precedent that's building is public profile is public profile data. That information is owned by the candidates. And then there is a lot of other data we get from Google, Bing, GitHub, you know, we have a deal with Microsoft and papers patterns so there's a lot of stuff we get.
Joel: Okay. So knowing that legally, you know, it seems like there'll be a precedent to say these are public profiles and you can get information, but there's also some reason to believe that LinkedIn does not want people to get this data and playing a game of Whac-A-Mole with how do I keep the spiders, you know, on their toes. How do I change, you know, URL structures, information structures. And I think knowing that Microsoft owns LinkedIn as well as GitHub, which you get GitHub data as well, does that impact your business and do you think that LinkedIn will win that game of Whac-A-Mole and keep the scrapers off their site?
Anoop: I personally don't think LinkedIn will really follow the Whac-A-Mole. They got into pretty big trouble with Haiku and from monopolistic ... so you know, they are just a lot of issues in following that strategy. LinkedIn is hugely growing their business on advertising, marketing, everything else. That's the biggest growth opportunity and I think it helps them to have innovation happening on top of their platform.
Quincy: Good morning. Good morning question. So I heard what you said and I hear where you get your data from. And I think it's a really interesting concept, but I don't really understand your business model. Are you creating a database of these candidates?
Anoop: Yes so we-
Quincy: How are you marketing it? How do you get the candidates? How are you getting people to go to SeekOut to get that information?
Anoop: So you know we have 450 million public profile, 16 plus million GitHub data, 87 million people that papers patents. So this is our database that we have where we have looked at that, analyzed it, indexed it's the best search engine on top of that, the business model is, we sell licenses today like LinkedIn. So you know, we have licensed prices of three thousand five thousand and $10,000 per
seat license. And then we also getting into enterprise wide licenses.
Quincy: Got it. And how do you verify accuracy of the profiles that you're creating? So how do you know if you say Anoop Gupta has 10 patents or has written these papers? There's more than one Anoop Gupta out there.
Anoop: Yes. Yes. So there is actually DPI and I'm involved in that too. In general there's a trade off called precision versus recall. You can match me totally, exactly a hundred percent confidence or 90% confidence. So we understand those confidence levels and we picked thresholds of, you know, very high confidence that they are the right match.
Quincy: And one more thing for me cause he keeps saying things that prompt new questions. I understand tech and engineering is your focus, but I also saw on your site you're also focused on diversity. Can you-
Quincy: Talk about that a bit please?
Anoop: Yes. So we infer diversity for women, for African Americans, Latinos, veterans and things like that. But the way we think about diversity first is insights. You need to understand that you know ordinary company, you need to understand competition. You need to understand for roles. So we have very rich talent pool insights in the platform so you can understand all of that.
Anoop: Secondly, we give you filters so you can find women candidates, African American candidates. Totally we have a blind hiring mode that reduces unconscious bias. Everybody becomes a cat, the names gets removed, the email to get redacted. So you can always do all of that. When you share this information with the hiring managers, you can turn blind hiring mode on. So you are nudging them towards unconscious bias. And finally we are also a messaging platform so you can message in, you know, ethnic gender appropriate ways when you're reaching out. So it's a pretty comprehensive solution.
Chad: So being pretty comprehensive solution, especially when you're dealing with all of that data. How are you actually going after companies? What's your market strategy in being able to really sell this product? Is it straight to recruiters? Is it talent acquisition, RPO? Where's your focus on the sale side?
Anoop: So we are not focused on RPOs today. Our focus is directly to enterprises, it's on the largest enterprises in retail and tech and banking are our customers. It is a both bottom up and top down strategy. So I was at SourceCon right before this and you know, lots of people are fans. So in their keynotes et cetera they talk about how SeekOut is an amazing solution. They go to their managers. We also go directly to directors of TA of HR and sell from that.
Chad: So RPO and staffing really understand tech and tech stacks and this data inefficiencies much better than talent acquisition. Right? Because it's their business. TA it's their job. So why did you make that decision not the focus on the actual quote unquote professionals versus-?
Anoop: So one is we are only a two year old company and you've got to start somewhere and you know, that is where we have started. That's where we saw early success, that's where we had connections. We've had from the limited conversations with RPOs because they're so large. So actually there are small RPOs who are customers that is there. The decision process, the cost sensitivity. So there's just a bunch we don't understand our fault and many of your here. I would love to chat with you afterwards show you the solutions and see where we can be a fit. So great question. I love to have connections afterwards.
Cindy: That was the same question that I would have and it was more that the RPOs are proactive. They're not waiting for the candidates to come to them. And it seems like your system would be better for those that are actively recruiting people rather than the talent acquisition folks no insults needed. But for those folks who tend to wait for the candidates come to them.
Anoop: So no we are not active so we are much more passive. So you know for most of the TAC is totally passive. So you go and find candidates on our system and then you reach out and hyper personalize it.
Cindy: Right. And that's what I'm saying an RPO recruiter is more proactive. They will use the system to go to it. Yes. My other question is how do you keep it fresh and do you, are you I guess two questions. How do you keep it fresh and also are you marketing to a job seeker candidates also?
Anoop: No. So we are not marketing to job seeker candidates while on the freshness pipe we try and keep it roughly every three months data within three months that is there. And that is a lot of effort, energy, money that is spent on making sure that's happening.
Cindy: Okay. Thank you.
Robert: I want to ask again about the integrity of the data that you're pulling on. So you're augmenting data that ... starting with let's say a profile from LinkedIn and then you're going out and you're finding other sources.
Robert: I'm going to challenge you on that and I want to see if you'll give me a little more specific answer.
Anoop: Yes. No of course.
Robert: We had an employee that was Ken Smith for years. And every time we heard this story about how we can actually aggregate all this stuff, we're like, okay, go find Ken and bring in all his other data. It was never Ken, it was always Ken, but not that Ken.
Anoop: So one is as I will say that is the issue of recall and precision. Right? So it means if you're getting the some Ken that's recall you're finding a Ken. Precision is the wrong one. So you get negatives. So out of the 16, 17 million profiles we have on GitHub, we have matched only three to 4 million for the rest we have made the trade off. The information of GitHub is still there. You can filter by all the filters, you can filter by diversity, you can filter by a lot of things, but we have not found their LinkedIn profiles. So we don't just go and say we're going to aggregate those candidate pools out there. They work as they are. You can find them, but we don't match unless we feel confident enough.
Robert: How do you match people with patents? I mean a patent database is not going to have your contact information.
Anoop: Okay, so what pattern databases have, so if you look at me, my background, so you'll find me with the Carnegie Mellon, Stanford, Microsoft, SeekOut. So if you understand my LinkedIn profile in terms of those associations that I've had, if you understand the name, if you understand some other things, you can do a good job of doing that. So these are again very hard. So I'm totally with you. These are very hard challenges and this is what we are trying to solve. So that we can and enable and empower recruiting organizations to succeed.
Robert: Good answer. One last question.
Robert: What is it that is the value of bringing in this augmented data from the standpoint of, isn't it true that if you ... I looked at your profile on LinkedIn, if you have patents, you're probably reasonably proud of that and aren't going to put it out there. What are you learning in addition from those patents that you go out and you'd link somebody to? Are you doing something else with that data?
Anoop: So there are two kinds of things. So you know, self-driving cars are pretty hot.
Anoop: Okay. Look at the LinkedIn profiles and say who's worked on LIDAR sensor integration fusion. You'll know very little about them. But if in fact a part of our target is not just recruiters, it's also how do hiring managers get involved, right? Means they know they can look at that and they can look at that information and say, this is the right person I want. We have a single block. You say everybody who's published in CVPR, this is the computer vision main conference with a single click, you can get the 12,000 people who have published there and you can look at them in a variety of ways. So that is not for everyone. So we do a lot of horizontal things that are relevant for all organizations, all recruiting, including diversity. But then we are also very specialized where we can find you candidates that are going to transform your company.
Joel: Outside of, you know, LinkedIn sort of banning you in some form, form or fashion. The second greatest threat I see to your business is the commoditization of profiles. It seems like there's never a week that goes by on Hacker News where somebody you know, has gotten access to phone numbers and Facebook or other information that's really specialized. And it seems to me like anyone who can put together a Chrome extension can create a competitor to your business. So prove me wrong that this isn't a commodity business that you're in.
Anoop: Okay. So just one is I'll tell you, we spent close to on the Azure elastic search $1 million dollars a year. Okay. Building a search engine that is quick, fast, we'll give you the insights, we'll do everything else is nontrivial. It's not kid's play. My co-founder was one of the main people behind the Bing search engine and how you build a good search engine versus kind of search engine. There is a big difference between those things. Okay. The kinds of insights we give you and how quickly we give you those insights.
Anoop: LinkedIn can't do it. They can't do it with the same amount of flexibility. They can do it with the same amount of speed. So at a surface level, and there is a lot of noise. So actually the part that I will totally take from you is there's a lot of noise. A lot of people raising their hand. They say, I do AI. People don't understand what the hell AI is. Okay. There's lots of different kinds of AI there's data science in AI ... so all I'm saying is, so our hope of differentiation is that given our backgrounds, given our deep techs, you know, we will be able to deliver.
Joel: So am I hearing you say that the data is out there, but the way that you guys aggregate it, serve it up is the differentiator.
Anoop: Yeah, you know, so if you look at a GitHub profile, how do you say what this person is good at? It's nontrivial. Right? And your being able to infer that. So you know, I don't know how much time they have as you see on my favorite AI example is you're a mid-Western firm making ball-bearings, right? Your board says hire machine learning people instead of people examining it, you know, do it automatically. What kind of, you know, personally are you going to hire?
Anoop: Do they be contributing to TensorFlow careers? Do they know how to use it? So JDS are terrible the way they are in today. The matching is terrible. How you engage with people is terrible. So there's good work to be done across all of those things. And where we are headed now is not just passive sourcing. There are lots of old candidates, you know ATSs. How do you surface them up? How do you update them? How do you search on the latest things? They are referrals, they are inside internal mobility. So there's a whole top of the funnel with a unified view that you can optimize in serving people and creating value.
Quincy: Have you done any testing on the accuracy of what you're pulling in? So validation of the profiles that you're pulling in is the first part of the question. And the second is how does your system handle contradictory information? So if there's one thing on LinkedIn, for example, and something else on GitHub, how is it managing through that?
Anoop: So my answer in some senses, these are fundamental trade offs in any system you build in terms of accuracy, UpToDate, contradictory. In general there is, sorry, I'm getting a little geeky but you know things called recall precision and those kinds of trade offs. You decide and you make bets and trade offs. You can show both kinds of information in the end profile. You say you know the here's the LinkedIn profile, you can go there, you can go as LinkedIn. We have combined them. You can give us feedback. If we did a bad job, we can fix it up later. So through a combination of social processes and choices that we make eventually we hope we are creating a lot more value for you than the trouble we are causing you giving us feedback.
Robert: You're an ex-Microsoft people.
Robert: And as it's been pointed out, you're pretty much trolling down the Microsoft stack. So if I'm a customer of yours, my concern is that on one hand Microsoft decides to cut you off in some way, either directly or indirectly. Or two they decide that, "You know what? We don't need these guys out there. We're going to buy them". How do you assuage those fears of how this is going to be worth my investment over a longer period of time?
Anoop: I think as a customer, what you have to worry about is if they cut off our data sources in a way which can legally take a long time, et cetera, that are you going to be ... so you know, whether we refund you whatever way we your return, your investment. If they acquire us, they acquire us because they are getting unique value that they're not finding. So you know, I have acquired lots of companies while I was at Microsoft. It's all visibility by decision that a large company makes. And you know there's focus issues that a large company has. There's nothing that Facebook or Google couldn't do. Startups exist because we take risks, we make investments, you know, we create things. Then they say, you know, instead of doing it ourselves, these are the right people.
Robert: But isn't there also a build buy or kill decision sometimes?
Anoop: There is a build, buy, kill decision but it depends on ... so one is, so I'll just mention it and I know this. We know people at the highest levels at ... you know Satya and I used to sit on the same leadership team. So killing will be in a very interesting for-
Robert: You have assurances is what you're saying.
Anoop: No, we don't have assurances. There's no assurance in life. There is no assurance in life. You're not looking for that. But you know, we will have a dialogue and there are lots of other players who can also benefit from us, not just Microsoft.
Chad: Okay. So GDPR is obviously a big, a big issue for some platforms because they didn't plan for it. GDPR is not in the U.S. yet but certain versions are in California now that will pretty much we believe, become the standard throughout the world. How are you going to attack really that that piece of GDPR? Because that data is, once again, it's public data, but it's also my data because I'm the candidate, right?
Chad: So how do you deal with that?
Anoop: Yeah, so GDPR is an important policy on privacy. We totally respect that. Now there's still a lot of ambiguity. So you know, you can say, Hey, do you have your profile? I need to explicitly ask you for permission, right? There is legitimate interest. What's called in the GDPR which says, you know, you don't have to this thing for jobs, you can do that. There's something called unreasonable burden where if you're going after 400 million people, you don't have to get permissions individually. You have to be sensitive to do you have sensitive, you know, personal information? Or you have more generic factual information that is there? There is when I'm contacting them, so there's a lot of color to it. We have it on our policy and we have it working with people. Thank you so much.
Chad: Anoop Gupta everybody!
Ema: Hi, I'm Ema. Thanks for listening to my dad, the Chad and his buddy cheese. This has been The Chad and Cheese Podcast. Be sure to subscribe on iTunes, Google play or wherever you get your podcasts so you don't miss a single show. Be sure to check out our sponsors because their money goes to my college fund. For more visit chadcheese.com.