HRExaminer Radio – Executive Conversations: Episode #368: Michael Rosenbaum, Founder, Chairman, and CEO, Arena

HRExaminer Radio Executive Conversations Badge Podcast Logo

HRx Radio – Executive Conversations: On Friday mornings, John Sumser interviews key executives from around the industry. The conversation covers what makes the executive tick and what makes their company great.

HRx Radio – Executive Conversations
Guest: Michael Rosenbaum, Founder, Chairman, and CEO, Arena
Episode: 368
Air Date: June 5, 2020

 

Guest Bio

Michael Rosenbaum is the Founder, Chairman, and CEO of Arena, which applies predictive analytics and machine learning to rewire the labor market around outcomes. Arena is deployed into over 850 healthcare locations that together process approximately 2.6 million unique job applicants per year, or approximately 13% the US healthcare workforce, with clients including Mt Sinai Health System, Sunrise Senior Living, HCR Manorcare, the University of Maryland Medical System, and RWJ Barnabas. When using its technology to predict for retention, Arena has succeeded in reducing employee turnover at every single client, with a median impact of a turnover reduction by 21% at 12 months of deployment and 43% at 24 months. Other outcomes for which Arena’s technology predict include employee engagement, patient experience, time and attendance, and even unique predictions such as sales yield.

Mike is also Founder/Executive Chair of Catalyte, a company that uses predictive analytics and machine learning to build technology workforces for the world’s most progressive companies, with the goal of improving productivity and results of software engineering efforts by removing implicit biases related to race, class, gender, and other factors. Catalyte’s platform identifies the most exceptional individuals without relying on resumes and other poor signals of quality, and unlocks their potential to thrive in technical roles. The company serves Fortune 1000 clients who use Catalyte’s technology and approach to drive their strategies and to build more diverse, capable, and local workforces. Catalyte’s workforce, which today is over 750 people, has historically mirrored the metropolitan areas in which it operates.

Living in Baltimore with family, he has a JD from Harvard Law School, an MSc in Economics from London School of Economics and Political Science, and a BA from Harvard College.

 

Website | Twitter | Facebook | Linkedin

 

Audio MP3AudioPlayer.embed(“f-html5audio-0”, {soundFile: “https://www.hrexaminer.com/wp-content/uploads/2020/06/2020-06-05-HRExaminer-Radio-Executive-Conversations-Episode-368-Michael-Rosenbaum-Founder-Chairman-and-CEO-Arena-Final.mp3”});Audio MP3AudioPlayer.embed(“f-html5audio-0”, {soundFile: “https://www.hrexaminer.com/wp-content/uploads/2020/06/2020-06-05-HRExaminer-Radio-Executive-Conversations-Episode-368-Michael-Rosenbaum-Founder-Chairman-and-CEO-Arena-Final.mp3”});if (jQuery.browser.mozilla) {tempaud=document.getElementsByTagName(“audio”)[0]; jQuery(tempaud).remove(); jQuery(“div.audio_wrap div”).show()} else jQuery(“div.audio_wrap div *”).remove();

 

Subscribe to Podcast or Download

 

Subscribe in iTunes

Subscribe via RSS

Download MP3 File

 

Transcript

 

Important: Our transcripts at HRExaminer are AI-powered (and fairly accurate) but there are still instances where the robots get confused and make errors. Please expect some inaccuracies as you read through the text of this conversation. Thank you for your understanding.

Full Transcript with timecode
 

John Sumser: [00:00:00] Good morning and welcome to HR Examiner’s Executive Conversations. I’m your host, John Sumser. Today we’re going to be talking with Michael Rosenbaum, who is the CEO, chairman, founder, and all of those things for a company called Arena. It’s going to be an exciting conversation.

As far as I can tell Arena is the only company in the HR Tech market that has actually figured out how to use AI to directly generate profitability while pursuing a noble agenda. And so listen in, Michael’s a pretty amazing guy. Good morning, how are you?

Michael Rosenbaum: [00:00:51] Thank you so much. I’m great. Thanks John, for having me. I really appreciate it.

John Sumser: [00:00:56] Oh, this is going to be fun, this is going to be a fun. I so enjoy talking to you. Take a moment and introduce yourself.

Michael Rosenbaum: [00:01:03] Thank you again, my name’s Mike Rosenbaum. I amm the founder and CEO of Arena and my background is I’ve been working in this space for a while. I started my career, I was going to be an academic and I was a fellow at Harvard teaching and doing research and writing on economics and law in the late nineties. And did some work at the time for the Clinton White House originally around technology policy and then around labor issues.

The Clinton White House was building a cluster of policies based on an idea that low income communities were untapped retail markets and build policies that incentivize folks like Walmart and 7-11 to expand in lower income communities. And I disagreed with the policy. I said, that might be true, but more significantly a lower income community is an untapped talent market.

And it’s an untapped talent market. Cause the market for talent is based on resumes. Resumes correlate with socioeconomic background and race and gender, but they’re not necessarily great predictors of success in a job. And at the time the hedge fund, the high speed trading hedge fund industry was ramping up.

And I said, no high speed trading, hedge funds are using data to make the financial markets work better. Why couldn’t we do the same thing in the labor markets, you know, if we did that, folks who are best at a job could get the job and organizations could get better talent and therefore be happy. And the response was the guy who came up with the other ideas, more famous than you are, we’re going with them. So I lost that argument. I went back to my ivory tower and my academic advisor. Who’s going to help me get the assistant professor job said, I think you’re more interested in sinking your teeth into this problem. So, I moved to Baltimore because it was a post-industrial city that had had longterm macroeconomic shocks.

And I started a company that applied these ideas to the software engineering market. And I built out a company that it was built on a technology that was designed to predict the likelihood that someone after receiving training would be in the top 2% of all software engineers. Scaled that business up so, you know, that business today is now sort of nine figures in revenue and focused on the software engineering space. But in 2010 had socially met an executive at a hospital in the DC suburbs who asked me what I did. And I told her, and she said, you know, we have this really big problem, which is that we turn over something like a quarter of our staff every year.

And over a third of the people we hire don’t last a year in certain jobs. Could you help us? And I said, I don’t actually know anything about healthcare. I know a lot about tech and a lot about data, but I don’t know much about healthcare. And she said, what if we’re an early customer, you think he could figure it out?

So we made a copy of the technology in that first company in that first company called Catalyte and dropped it into a subsidiary and tuned it to predict for the likelihood that a given individual would achieve a particular outcome, specifically retention. In a particular role, in a particular location, in a particular department, under a particular manager, in certain situations, on a particular shift.

And then we made those predictions and then delivered them back to the organization. And we deployed that technology into the original hospital and four other hospitals and two longterm care faciliies and the results looked too good. You know, our worst result was a reduction in turnover by 45%. And I said, I don’t buy it.

We don’t even know what an RN does for a living. We certainly don’t know how a hospital works. So we spent three years working in hospitals and long term care facilities, figuring out what data we could get, what data we couldn’t get and what an RN does and how the organization’s put together and change the technology and tune the data.

And the numbers got better. So starting 2013, we started to scale it up. Spun the company out as a subsidiary of the original company at the end of 2014, spent a year and a half splitting up the businesses. So for the last four years, this business has been a standalone business and we’ve been scaling it up primarily in healthcare.

And now, now in some other restaurants, like then other verticals like restaurants,

John Sumser: [00:04:24] So I just want to double back and underline a couple of things that you just said for the benefit of the audience. I don’t know exactly how old you are, mid forties maybe. And you’ve already built a nine figure revenue company. A nine figure revenue company.

That’s not a kind of a normal accomplishment. I’ve talked to hundreds and hundreds of CEOs here and there. The club that you’re in is a small club. And then to go, oh, well, well, I, you know, I built this nine figure company, I’m more interested in this other thing I think I’m going to go do that. You know, you know, having built your nine figure company you’re supposed to, you’re supposed to launch into an era of complete hedonism and decay. You chose this other path. What’s wrong with you?

Michael Rosenbaum: [00:05:16] So, in all seriousness it took a very long time for that company to get to nine figures. And one of the ways it got to nine figures was, a number of years ago I realized that the technology and the product were, were really interesting to me and, you know, and I could be effective at, but as that company got bigger, There were certainly other people out there who’d be much better at being CEO of that company then than I would be.

And so I’m still the controlling shareholder of that company, but I brought on a CEO to run it a number of years ago and he is about a thousand times as good at that job as I ever was. And so he’s really been able to do wonders with it and, you know, and it allowed me to focus on that business, on the things that I’m more effective at, you know, but equally, if not more significantly, allow me to really focus my time on Arena, where we’re, we’re able to have really dramatic, horizontal scale on issues of the labor market not working correctly.

And, you know, I’m sure your audience knows as much or more about this than I do, but, you know, but the core problem of the information like cemetery and the labor market. There’s one that has a whole bunch of incredibly malignant impacts and the fact that each of us aren’t great at hiring. And also each of us has potential job applicants or folks doing a particular job.
I’m don’t necessarily know where we’re going to thrive, but the problem of not knowing those things, not information like symmetry. Causes a host of problems, including things as momentous as frankly, a weakening middle-class. So the idea of being able to impact that at scale across many verticals was really interesting, I mean.

John Sumser: [00:06:50] So it’s worth taking some time to unpack some of the things that you said, because you’re so steeped in the language of, is this an academic economics? It’s not clear that what you’re talking about is a tool set that booze people from one industry into another industry by making prints addictions about how well they will do there.

And that you have figured out how to do this very profitably. And that you are at a, at a point in the process where you are figuring out how to really scale the operation to massive size. And so this, it couldn’t be more appropriate that we’re talking today because this is the kind of actual solution to the structural problems that we’re having, that people are hunting around for, right?
No. Yeah. We’ll talk a little bit about that. This is a profitable idealism. Is that fair?

Michael Rosenbaum: [00:07:51] So, you know, for profit versus nonprofit entities or beliefs or compact, that’s a condition fundamentally. If you can’t, whether you’re a for profit or nonprofit. If you can’t generate margins, if what you were, what, what your innovation does, does not generate economic value that can be captured in some way that generates a lot more incoming cash than it costs you.

And you can’t really scale something to be massive. Fundamentally the goal of arenas to rewire the entire labor market around outcomes. And our belief is that if we can rewire the entire labor market around outcomes, then frankly enterprises will be able to unlock massive amounts of productivity they otherwise couldn’t do.
And, you know, we can create pathways for upward economic mobility for every individual in society. But in order to generate the kind of scale you need cash flow and, you know, without cash flow, you can’t get that. Okay,

John Sumser: [00:08:42] so that’s interesting, but you also do this fascinating thing for your clients, which is particularly in this early, going time, your clients save enormous amounts of money when they use your service.
I wonder if you’ve ever thought about asking them to send the savings. Mmm, improve copy.
Um,

Michael Rosenbaum: [00:09:09] so we, we are a commercial enterprise, so telling people what to do with the money, we saved them. And isn’t always the best commercial strategy, but I will tell you that, you know, the core of arena, it’s a platform that sits behind the other systems of a large enterprise or an enterprise. You know, on the larger side of small, medium, or large, and typically folks initially use arena to improve retention.

That’s the metric that someone wants to move. Initially, when they first used the platform on the premise that increased turnover has a whole bunch of costs, but some of the most easily measurable are things like contract labor over time, and some sort of pure marginal cost that you can actually trace to a budget quarter.

And no, we have a hundred percent success rate across our client base at improving retention or average impact. 12 months after going live is a reduction of turnover by 21%, 24 months after going live. So reduction for number by 43% and turnover is important. I mean, it’s, you know, it’s mundane in some ways, but irrelevant in other ways.

And you know, because there’s cash, you can tie to it. It makes it easier for someone to adopt this. And so if you can say, we’re going to reduce your contract labor and overtime. Cost by five to 10 times what it costs you to use arena. Then it’s a really easy. Yes. And we don’t necessarily say what to do with them four to nine times, our fees that someone saving and easily measurable cash in a budget quarter.

But once folks see at work and trust us, it allows us to move beyond the initial champions who are typically the folks who bring us in. So typically, do you have a large enterprise? We have a handful of people that really get why this works. They’ll bring a Rena in a bunch of other folks will say, Oh, that’s the group that always likes the next, the new shiny object.

We’re just going to wait him out. But then they see retention improve and retention hurts everybody. They feel it particularly first-year turnover. So, um, so it gets a larger group of people involved and then people will say to us, what else can we do? And it’s not stuff that we’ve charge for, but, you know, but things like, what else can we do to solve some of the underlying issues that cause some of our challenges.

And there are things that, you know, that arena by applying, got it, the sound sort of lingo ish, but you know, machine learning to building out a workforce by doing that, you know, there are things that arena can help with. Is there a larger number of things that you can’t directly help with? But part of what you’re doing is because you’re demonstrating that machine learning use effectively can solve some of the greatest pain points of a wide range of people across an organization. It gets people bought in to the idea of innovating more.

John Sumser: [00:11:43] The next question is right out of what you were just talking about. So around the 1st of February, the 15th of February of this year, much of our historic data, that’s getting used around the industry. It came kind of questionable and you don’t have to look as far as the fine, the great pontificators.

Maybe me included, um, declaring if that’s the death, the induced coma for machine learning, because it’s not so good at dealing with emergent phenomenon. And, and there’s a problem with the data in a lot of places, but you’re not saying that, right. You’re saying that your data is your date is good. The answer to that is that you have the capacity to artists clear.

Concrete feedback to validate whether or not your models are working. Um, and that’s not always the case in the kinds of recommendation answers that are coming to market. So, so talk a little bit about the role of hard feedback as performance objectives in making your machine learning. Implementation works.

Michael Rosenbaum: [00:12:53] Yeah, no, absolutely. I mean the most important part of arena, I mean, lots of important parts, but the core to arena is the ability to get a real time or. Almost real time. Speed of outcome data from every client. So for retention, that’s raw payroll data for whether or not someone’s going to show up to work on time.

That’s from a time and attendance or a scheduling system for employee engagement. It’s from an employee engagement platform for name your outcome. But the core to arena is that every client with every client, we get a feed. Of that data that comes with some frequency no less frequently than once every 30 days, depending on the nature of the integration.

And that’s the data Lake that lets us see what’s going on. And, you know, that has always been important. Um, you know, every enterprise is not just a snapshot in time. It’s, it’s evolving, it’s evolving because of external forces. It’s evolving because of internal forces, you know, in the early days of what we were doing, some of the more interesting stuff was what happens to the predictions in a particular department when the manager changes over.

And, you know, being able to pick up on that and see those shifts and adjust to those shifts chef’s was, was fundamental to the ability of our generating the outcomes we needed to generate. You know, and at the same time, things like using demographic data feeds from data, we would get in an ATS system that would fit to the side, obviously not in the models, but so decide to be able to evaluate whether or not, you know, there was implicit bias or explicit bias seeping into certain decision making was fundamental to our ability to achieve what we achieved.

And interestingly, that same functionality became exponentially more important. With the rapidity of the shocks to the labor markets that have happened over the last 90 to 120 days. And what it’s let us do is let it do a couple of things. One is it allows us because, you know, so we now have say 90 or 120 days of data on impacts of the kinds of shocks that we’ve seen.

It allows us to, to look at that and see potentially what that means and get some early thoughts on it. It also allows us to look over time at things like what happens when they’re, I mean, there’s a very different scale of issue, but, you know, in healthcare, obviously very impacted, very significantly impacted by public health issues.

What happens when there’s a particularly acute flu epidemic in a particular geographic area? What does that do to various outcome metrics in the way the models shift in that area. And we can use some of those learnings to apply together with the data that we see real time or not actually real time, but close to real time.

That is outcome data that allows us to generate these predictions. So can we tell you for certainty what our impacts are going to be in the next year? No, of course not. But do we have the kinds of feedback loops that allow us to evolve those predictions as the macro and. Micro-environments change.
Absolutely.

John Sumser: [00:15:49] So one of the things I get from listening to you talk about this is that the depth of the models that you so is extraordinary, but in order to deliver this to the market, you have to understate the depth of your expertise. How did, how do you deal with that? And if you have a customer who’s curious about sort of the deep elements of your models, How do you, how do you know that there must be quite a range of interest in terms of the customer base?

That’s a, okay. The questions that boils down to, Oh, the simplification get in your way. Mmm. If somebody, if somebody wants to dig deeper, is that a good thing or a bad thing? Well,

Michael Rosenbaum: [00:16:33] as you can probably tell, we love it when people want to dig deeper, love it. When people wanted to keep up, you know, some of the conversations we love to have are around implicit bias.

For example. So we’re headquartered in Baltimore. And one of the advantages of being headquartered in Baltimore is that we’re essentially next to the national security agency, which provides a rich pool of data science talent. And, you know, one of the mechanisms we use to deal with implicit bias is basically a repurposed methodology that was originally developed by the intelligence community to deal with deep stakes.

And it’s called generative adversarial networks. And we will occasionally have a client who has a deep technical understanding. He wants to go deep on that, and nothing is more fun than that for us on that. But at the other end of the spectrum, Um, you know, I would say in an earlier iteration of this product, we would deliver predictions to end users in an enterprise that would be yep.
Prediction, a zero to a hundred prediction of the likelihood of something happening. And what we realized was that at the other end of the sophistication spectrum, we would had end users. So, you know, a manager in a hospital who was fabulous and had great bedside manner and incredibly motivated interpersonally, but very uncomfortable with quantitative concepts.

And, you know, and that delivering a number what’s a little off putting. And so we actually transitioned the way that we deliver our predictions to certain end users, to green, yellow, red, which is about as simplistic as you can get with green, meaning someone’s likely to improve that outcome metric yellow, meaning they’re likely to have no impact on that outcome metrics and red being meaning they’re likely to make that outcome metric worse, but that we realized that there were folks who didn’t really, they wanted to know if it was real.

But didn’t really love the idea of digging too deep on the technology and the data and other folks who were skeptical and wanted to go really deep. And we needed to develop a frankly, a sales and client engagement model that allowed us to meet a client where they are. So from a sophistication level, you know, which was really just a parallel to frankly, how you have to build the technology generally, which is again, to meet clients where they are.

I mean, one of the reasons we sit behind the other systems of an enterprise is that change is difficult and. You know, most people don’t want to learn a new system. And so we realized that meeting folks where they were meant using the tools they already had in place and just adding additional functionality that enabled them to make productions.

So can I, I hope that answered your questions.

John Sumser: [00:18:53] Yeah. So it sounds like rather than aspiring to become some sort of fight for what you’re looking to do is to become a that’s sort of, and add in that allows. Better control of outcome for customers by delivering the predictive capacity inside of the tools are already using. Is that fair?

Michael Rosenbaum: [00:19:17] That’s exactly it. And that’s the way that we can drive adoption most significantly. So, you know, an example, hostel systems use, tend to use ATSs. With very high adoption rates, assisted living operators that are geographically dispersed, but generally have the executive director of a community not making those.

That’s typically the hiring manager. That’s the person who’s making the hiring decisions generally have very low adoption rates, but high usage of email. And so. You know, in an assisted living operator with low ACS adoption, typical workflow is executive director shows up, spends five minutes, checking email, does it stand up with their team?

And then all bets are off. So we deliver an email in the five minutes, they’re already checking their emails. It says certain people who applied since your last, since he last checked and Kira their predictions of the outcomes they’re going to achieve. And if you want to do something, click here. So again, just meeting folks where they are.

John Sumser: [00:20:07] Cool. So I’m a big fan of your much larger vision for the company, which will take a couple of minutes and talk about going from a healthy, profitable sort of out of the pack, basic startup, too, moderating the national labor market. That’s such an interesting thing that you’re trying to do. So tell me about that.
As we wrap this up. Absolutely.

Michael Rosenbaum: [00:20:32] So last year we did a couple of things that became incredibly important. 90 days ago, one of which is we dropped a product. It was in beta onto our platform that was designed to point the technology outside the applicant pools in the workforces of an enterprise. So instead of making predictions about folks in the applicant pool, in the workforce, we’re going to have folks who come from social media or through a job board, or who are rejected applicants from other clients or the other client agreed to it and predict the likelihood that, that individual.

Would be likely to achieve an outcome in some job they didn’t know about before. And the other thing we did is we expanded into restaurants. And so a few months ago, when the labor shocked labor market shops started, we said, we’re going to offer this product for free for a number of months to our client base.

And we’re going to ask the restaurant clients, if we can route folks who otherwise would have worked in restaurants into health care. And the answer was yes. That then took off on the premise that if you can basically create a website and say, we’re filling today, it’s like we’re filling 150,000 jobs a year in our client base and, you know, spend a few minutes getting us some information and we’ll predict the likelihood that you’ll achieve some outcome in jobs within 10 miles of you.

And then we deliver it, that list of jobs that are jobs that are already on our platform to the individuals who have come to the site and the hiring managers on those jobs are using the same platform to make hiring decisions. And so if someone says they’re interested, they’re very likely to be hired.

And what it lets us do is it lets us do things like identify when someone’s working in a hotel and isn’t working in that hotel anymore. But it turns out would be great in a hospital somewhere. And that allows us to really get at the problems we’re trying to solve. Right. So problem, one being that we all have biases when we hire folks know, we generally like to hire people like ourselves, some of that is malicious.

So class, race, gender. Others are not Melissa malicious. I think someone with a four year college degree is going to be better at this than someone with a community college degree, but both are biases that lead us to make core decisions. And so you’ve solved that issue, but also as individuals, why do I work at, in this hotel?

Because I had a friend who worked there, it’s down the street from me. I knew about it. You know, that hotel had a big recruiting, marketing budget. And I may not have realized that I’d be much happier working in this other industry, in this completely different role. And not only would I, it might be, I’d be happier.

I might actually make a lot more money. And by using a hub like this, to address that question, How does many people on that hub as possible? And given the footprint we have, I mean, today we’re deployed into healthcare operators, my healthcare providers, those as a group process three and a half million unique job applicants per year.

So 17% of the us healthcare workforce, the ability to sort of route folks across verticals. Allows us to help folks find places where they will thrive, where they’ll find big and be where they’ll have economic growth. And it helps us. It helps us help an enterprise retail, I think how to source talent and who is going to be best in the job in a way that allows us to put a much more effective and productive person in a job, then they can find otherwise.
So the parts

John Sumser: [00:23:27] of, I liked so much that you about you, this going well beyond health care and moving people from hospitality to healthcare, to a large slug of the overall labor market. How bout a little bit more about how you give out of the, the niche that you’re in into the broader game.

Michael Rosenbaum: [00:23:45] Definitely. So what we are doing now is we are, we’re working with early adopters in different verticals.
So we’re doing that already in restaurants. And we’re looking for a few more verticals where we can have sort of two or three progressive companies that are interested in working with us to make whatever tweaks needed to meet, need, to be made for this to work seamlessly. In that particular vertical. We do think this verticalized because of the verticalization of outcome metrics.

And so our strategy is to work with a handful of our early adopter customers and a handful of verticals as the product gets tweaked in those, with those early adopters, then to scale aggressively into those verticals, with the idea that over the next several years, you know, we’re across numerous verticals, you know, in the U S and potentially beyond the U S in a way that allows us to route folks across verticals and really to move the needle on different outcome metrics in each of these verticals, and to be across a very substantial portion of the entire labor market.

John Sumser: [00:24:38] Thanks for that. This has been such a great conversation. I wish we had a couple of hours.

Michael Rosenbaum: [00:24:42] Definitely.

John Sumser: [00:24:43] Would you take a moment to reintroduce yourself and tell people how they might get a hold of you?

Michael Rosenbaum: [00:24:49] Absolutely, my name’s Mike Rosenbaum and the company is Arena. The website is arena.io, and you could feel free to reach out to me directly. at M Rosenbaum. So M R O S E N B A U M@arena.io.

John Sumser: [00:25:05] Great, and I just want to take a moment to notice that the work that you’re doing is actually solving problems in the underlying sector that the protests are all about right now. And what I appreciate so much is that we didn’t hear any chest beating about how relevant you are at all of that, over the course of the conversation, that stuff is getting pretty obnoxious.
Thank you. Thank you. Thank you for just doing the job and not rubbing your chest while you’re doing the job. You’ve been listening to,

Michael Rosenbaum: [00:25:36] Thank you, I always love talking with you John.

John Sumser: [00:25:40] Yeah. You’ve been listening to HR Examiner Executive Conversations and we’ve been talking with Mike Rosenbaum, who is founder CEO, and all the rest of those things for a company called Arena.io.

Look them up on the web. They’re pretty amazing. Thanks for tuning in and we will talk to you same time next week. Bye bye now.