This month, Jason Ezratty joins host Bruce Morton on Subject to Talent for a discussion on workforce intelligence – its beginnings, its evolution and how its powering better decision-making on how, when and where work gets done. As the cofounder, chairman and chief data scientist of leading augmented analytics company Brightfield, Jason digs into the smart technologies (NLP, ML and AI) and changing ideologies that are driving the new world of work.
Bruce Morton: Welcome to Subject to Talent, brought to you by Allegis Global Solutions. Similar to you, we're always trying to learn more. On this podcast, we speak to workforce and talent experts from around the world, that we market trends and technology and our ever-evolving dynamic industry.
Hi, I'm Bruce Morton, the host of Allegis Global Solutions Subject to Talent Podcast. Today I'm joined by a very good friend of mine, Jason Ezratty. Jason is the Co-Founder, Chairman and Chief Data Scientist of Brightfield, the leading augmented analytics company powering the economics of digital transformation in the new world of work. At Brightfield, Jason leads development efforts on workflow-impacting machine learning products with the Talent Data Exchange, some of you may know as TDX, the world's largest contingent workforce data consortium. Jason, welcome.
Jason Ezratty: Thank you, Bruce. Pleasure to be here.
Bruce: Great to speak to you today. Jason, we always ask our guests the same first question. So here we go. How did you get into the workforce industry and what was your journey to where you are today?
Jason: For sure, I got into the workforce industry by accident. Was not a plotted path. My career began in science, which led me to creating a biotech startup. That led me into a strategy consultancy that also had a lot of software development that was going on, on behalf of clients. And when that folded amidst the dot-com craze in 2001, about 11 of my colleagues went to a VMS company and they lured me over there. So I was brought over there because everyone was looking for a job after September 11 here in New York. And that's the one that found me.
Bruce: Oh wow. And have enjoyed ever since?
Jason: And I've enjoyed it ever since. It was about a year into that job that I realized that we were doing more than just saving 10% on someone's labor budget, but we were actually creating the architecture and the fabric for what will become the future of work. The idea that work can be portioned out in the same way that other things were being purchased on a catalog at the time seeming revolutionary. Just seemed like an obvious future.
Bruce: Great. In case our listeners don't know Allegis, we've had a long relationship with Brightfield and we recently announced an expanded partnership, where we were named as your Brightfield's first Platinum MSP Partner. We'll talk a little bit about that later on.
But first, can you introduce our listeners to the Brightfield organization, who you are, what you do and the evolution of the company?
Jason: Thank you. In 2006, Christopher Minnick and I started Brightfield Strategies, then a consultancy, and our job was to help companies not just save some money on the contingent labor, but try to figure out why did they have contingent workers in the first place, was it the same reason in one country versus another; and really getting down to not just the quantification of the contingent workforce, but also the strategic and underlying policy issues.
As we realized that this was more and more of an analytics need in this marketplace, that there were others that were able to satisfy a variety of the needs, but what it really needed was a best practice in how to calculate, quantify, and ultimately aggregate data; if we're going to create a marketplace where there's any degree of sameness, we can make claims about what is typical in a given phenomenon, whether it's bill rates, whether it's how long it takes to find someone, that's going to be what's most special; and that's when we dropped the "Strategies" from our name and just became Brightfield and decidedly more positioned as workforce intelligence, as a product company, as opposed to a services company.
Bruce: Right. And then, as you made that transition... Obviously, a couple of buzzwords, but I'll throw them in here and let you respond... To the AI and machine learning: how is that impacting the usefulness of data?
Jason: When it comes to aggregating data that was not meant to be aggregated in the first place, you have to help it to conform. If you don't have sameness amongst the parts, then any of the math that you do is going to be inherently filled with error. So the first reason that we need AI is to be able to declare which of these types of work transactions are same or similar and to what degree, how and why. That was really the first piece to be able to do that better than a group of people would do if they sat around looking at requisitions one at a time; to be able to automate that, not just to make it faster, but to make it better and more specific, all the details that a group of people would never otherwise be able to keep in their heads let alone across all of those roles.
Then beyond that, trying to figure out what are the patterns in the market. Not just what are the pieces in the market, but what are the patterns of the market, what is it that makes a job close faster, what is it that make rates go higher in certain situations than others. And then increasingly, what is it that makes a buyer behave the way that they do versus someplace different or someplace better. So AI is at the heart of any dataset that is this complex, this multifaceted, and frankly, this fuzzy.
Bruce: Right. I like that term. All this data was never meant to be aggregated. That's why it all started in different places.
And talking about predictability. And I'm not sure that any of us would have predicted the way the economy has bounced back in our world that we live in, a matter of organizations at a hiring and there surely some talent and so on. But as companies are bouncing back or needing to bounce back so, so quickly, the demand for talent is obviously front of mind for everybody, and far greater demand than anybody possibly anticipated. So in that landscape of contingent labor, that’s been dramatically reshaped in a way by this massive uptick. Can you talk about how workforce intelligence and all the work that you're doing is actually playing a role in helping companies move on from that rabbit in the headlights panic moment to actually start to adjust and remain competitive and, you know, staffed in this marketplace?
Jason: Absolutely. I mean, importantly, the reasons why workforce intelligence is important now are the very same reasons as it was important the months prior to COVID, it's just, COVID made it that much more relevant, that much more urgent, that much more critical. And so it's the ability to understand what is the nature of the environment that you're trying to acquire labor within. If you underprice your labor, it's going to be hard to find talent, let alone good talent. If you overprice your labor, you're wasting resources that could otherwise be spent elsewhere, whether it's more talent, better talent, whatever that may be. Many of our customers have been dealing with both problems simultaneously, where there's a need to find cash, savings, but at the same time need to spend more, to find better talent. So finding that balance of how do you execute on that? Well that's with workforce intelligence, because that's what enables you to put things under the microscope, so you can see those types of opportunities distinctly. It's not a one size fits all type of situation.
Bruce: And there's a lot of talk now of this talent anywhere concept. And I guess what I know, but I'll just bring the points up that, pre COVID it was okay. I need to know what a developer is going to cost me in Palo Alto’s zip code. And now of course it's so if you can hire anywhere that must have really exploded the requests that you're getting from your clients.
Jason: Yeah. As well as the need to understand what does that do to the economics, right. So on one hand, it just opens up that you have a greater pipeline, you have a fatter pipeline that your talent can be accessed through. That's very important in and of itself. And to your point, what's driving it, but it also enables you to rethink, what are appropriate economics. If you're not paying a mortgage in Palo Alto, it's likely that you can find top talent at a rate that's not necessarily as stratospheric and importantly, those types of companies, your Palo Alto type of companies, they're not just looking to pay average, their question is what's top quartile.
And so being able to have that kind of resolution of the market to say, not just 50th percentile, but a 75th percentile and why that matters. And then how does that compare to rest of country 75th percentile so that you are still an attractant, that's all about workforce intelligence.
Bruce: Yeah. So as you said, it's not just about cost savings. It's where is the best talent? How can I get to them? What's it going to cost me to get to them? And how are organizations thinking about that? What type of talent they should be bringing in, whether they should be bringing them under a typical contract staff or type of arrangement where they're paying them on an hourly or daily versus bundling up piece of work and putting a price label on them, how does the workforce intelligence help navigate that maze?
Jason: Our role is to first and foremost, quantify the trade-offs. So if you're saying I'm going to go an SOW channel, because I can have better predictability of the level of quality, I have better predictability that I have an account manager that I have a relationship with. And if I have issues, I know who I'm complaining to and what I'm complaining to them about. They'll take on more of those headaches. And as long as you know, what that premium costs you, and you're prepared to pay that, then it's fine. So our job in that kind of a case would be to present, what should you be expecting and what ought that cost as, as one example.
So at the primary basis, just what are the tradeoffs? And are you, are you in line with what those trade-offs are in your buying environment? From there, it's also making sure that people are actually thinking about all of their alternatives, especially a lot of the new alternatives. Once they realize that there's better alternatives than their traditional conventional way of thinking, if you can actually get them to an optimal path and present them more alternatives, then I think you're in the right sourcing game. Otherwise, you're just policing activity.
Bruce: I'm putting your scientific brain on the data that you're seeing the impact of COVID on certain skillsets and certain markets, I guess, as well as geographies, what are some of those trends that you're seeing that are more than just a blip and we can start using the data, not just to respond to market conditions, but to actually start predicting the market conditions in all.
Jason: Yeah. So I guess some of the obvious things, there's twice as many contingent nurses in that status in the United States than there were a year ago, there's those types of things that make sense, align with expectations, what's happening to healthcare rates in general, what we saw in contingent worker demand at the factory floors at each of those predictable stages around the pandemic, all of those things were true. Some of the things that were a bit more surprising is actually the pay rate impacts to those contract workers were not necessarily the same in line with, with all of those increases in demand. That was not necessarily seen the same way as it was seen in IT and in healthcare. The other thing is the actual market data does not necessarily always tell the exact same story as what you see in the headlines.
What you're seeing in the headlines is what's then very exciting and compelling of that moment as the market then comes back down to normal. There's no follow-up headline to say, and now it's back down to normal. That's just not as exciting. And so when we see you, oh, that was really a Q1 phenomenon. That was something about March and now things are really coming back to normal. And so really what we're seeing is as of today, we're seeing pre-pandemic kind of levels of demand having returned and sort of coming back to a normalcy in terms of demand. It's not to say that we haven't seen sustained impacts to prices to bill rates, but a lot of the demand factors across the board have come back to normal where it's staying the same as healthcare and IT.
Bruce: Okay. And just to bring all of this alive, this is great information. Can you give one, two or three, a few examples of some of those early adopters and what value they've seen over that time period and how organizations are truly understanding to get to everything that possibly can out of.
Jason: Yeah. My favorite stories happen in each of the stages of maturation. So it's first and foremost, here's three roles that we were really calling one role and being able to use our product, to see those patterns and help them split out and that they each have their own separate levels of metrics that present different challenges and opportunities to be addressed. There's certainly often very differently priced and have different supplier opportunities. So that's all at phase. One of just where we, calling the workers, something too broad that made it so that we were unable to really understand what their differences are at a metric level and how they compare to market to find those opportunities. So that's step one, beyond that, getting more interesting answering some of the workforce decision in questions that you raised before. Well, here's someone that started on a SOW, was that the most correct path or was that violating policy for this to be a statement of work worker?
So then taking customers through the automatic identification of which sow were misclassified and presented opportunities for bringing in staff augmentation worker or, or otherwise finding other forms of better contracting and likely savings. And then ultimately, as they go through the full spectrum, we're now also showing how things fill better. So it's not just finding the best rate, but finding the best rate at the ideal level of fill. Cause that's really the question that people are asking and always will be asking, I need to find talent. It's not just about what will I be paying for it, but set all of my expectations around it. So I can, I can solve that equation.
Bruce: And is that looking at the labor sort of market demand versus supply by locations at the way you're getting to that information? Is it watching trends of hotspots to avoid or because there's a wage hike or an hourly rate hike?
Jason: It can be. And the flip side of that, especially, to your earlier point of work from anywhere to see that you're not constrained to these markets that have lower supply, high demand, higher rates that you can, you can take advantage of parts of the market in a given country that don't have all of those constraints.
Bruce: Right. The one thing I know that you're very passionate about is the concept of a bell curve. As we think about market rate versus one point of medium, can you just share to the audience why that's important and just explain what I just meant, but what you mean by this.
Jason: Yeah, it's the old plus minus. So, when we talk in averages or, or even medians, when we think about these types of statistics, we forget that even within a given business unit of one, given buying customer, they have radically different bill rates. Sometimes that create dots along a distribution of different rates. And so we might say the average is $67.15. And then we think it has that kind of laser beam precision, but really what's happening is to your point, and many times that shape will look like a bell curve. And the better we understand what are the things that push points to one side of that bell curve or another, the better that we can bring intelligence to workforce decisioning.
Some of those things just describe things like skills, which skills are north or south of a given median. And therefore it should adjust how we think about the cost of a given role, but other things that are in our ability to change things like source type, should I be thinking about one source type versus another, should I be thinking about one supplier versus another, one location versus another, all things that you know you have in your control?
Bruce: Yeah, that's great. And I think that's why we're so excited about our platinum partnership. You know, having our team certified on the platform, understanding how to actually read the data, into the data, then make advice based on the data, but taking other things into account, right? And the impact of, if you drop the rate by $5 an hour, what does that really do to attrition and to quality? I love the fact that even though AI machine learning data has come so far, there is still a bit of an art around having that human. We still have a place, a part to play here of that human advice that we're giving. So is that the way you see that?
Jason: Absolutely. No matter how good the quantified answer is, there is always going to be degrees of change management and something that is so essentially human.
This is regarding not just people and buying behaviors of people, but also the work, the working behaviors and hiring behaviors of people. And what does compensation mean to people? So there's no question that there is a human element of that and the change management element of that, the types of things that we will look at and test, and say, okay, instead of, instead of X, do Y, don't have that policy anymore, lift that restriction, someone has to bring the change about them. That it's not just about calculating prospectively, what that impact might be. There's a massive amount of execution and change management.
Bruce: Right. And a slight tangent. But I just like you to share with the audience, that one thing we're very proud of – I think both organizations through our partnership – the technology that we've created to give us the ability to lead statements of work at speed, to better analyze them, to answer some of those points you were making earlier about this data was never meant to be aggregated. That's why it's come from a different starting point. Can you just share from your lens how that tool is, how both organizations, but to get to that data quicker and to make sense of it?
Jason: Yeah. And it's an exciting time to be in the world of natural language processing. It’s one of those areas where advancements happen very quickly still. And so what that means for us is this is not only a type of data that wasn't originally thought of as being aggregated, but wasn't even originally thought of as being structured, they're just words on a page. So to first be able to say what words are on this page and with what visual patterns do they have in order to then say, okay, when we see these words this way, we know it means something we know that's the start date of a given contract. We know that they're talking about a supplier name at that particular point or whatever.
And so we are able to extract at this point over a hundred different fields that are various contracts, and then you can look at them in a database and start to understand what does that... Over time, how does it separate by group? What are the supplier impacts when looking at how different groups buying similar things, et cetera, even just at a first step of basic visibility, it's incredibly informative. And then you finally have a base to ask business questions which of these contracts could have been done differently. And what would that have meant?
Bruce: And I guess historically that work is obviously either done by humans or I experience a lot of cases. It just isn't done. Nobody's actually going back and reading those things right?
Jason: That's exactly right. Everyone knows it's a good idea. Everyone knows it's something that should be done. And it was a relatively few companies that had enough urgency on it to apply the manual labor to get it done. So the bringing automation to this world helps the 20 as well as the 80 in, in that equation.
Bruce: Okay. So as we start to wrap it up here, but I think my favorite quote for today was it's an exciting time to be in natural language processing. So thank you for that. I will try and use that late in the day. You can actually add, do yourself here by another quote by, so it's the crystal ball time. So if we would give you a crystal ball and pick a moment in the future, you make that decision, how far you want to go. But if you were looking at that crystal ball, what is the future of our industry look like with all of this phenomenal AI machine learning that we're doing, that when all this starts to really bear fruit, how do you see the future?
Jason: Well, I don't know. I think the harder part for me is going to say the when, than the what. So I don't know how far forward I want to look, but as a student of the history of technology and seeing which ideas that were possible actually happened and which things stayed the same. I think the part that drives the sense of the future for me is, is the fact that the stigma of contingent workforce that, that label, the non-employee label is no longer as stigmatized as it was. In fact, I don't really hear it talked about in that way any longer. That for me is more important than any piece of technology to be able to then say, I see the great normalization of contingent workers, and we won't use that word. And it'll become about the Hollywood model that the best actors and actresses and directors don't want to be in any contract that binds them away from being able to do their life's greatest work.
Jason: I forget who said this first, but it's the way Michelangelo got all of his work. He's a contingent worker. He did contract work. So I think that, that's what happens. And so how AI plays into that is being able to find ever greater aspects of fit between the parties, helping those who raise their requisitions as hiring managers and project managers do that ever better to say the right things that actually impact the work they're trying to get done, not just trying to get around policy. And then on the flip side, the same thing, helping those candidates and teams, and suppliers organize themselves around those opportunities in ever smarter ways. Right now, I think we can be polite and say that a lot of the negotiation feels like blind liar's poker. So being able to bring more intelligence to that, where they're actually talking about the solution, and then it's the best solution wins in light of best price. Not just a lot of negotiating and then trying to solution to budget after the fact.
Bruce: I love that. And if we can stop calling those workers contingent, perhaps we'll stop calling employees permanent.
Jason: Yeah... And then I think we'll see those numbers come down. I think that there is a greater chance for contingent. I also think that we'll see a concept of a longer-term contingency, of course, what would get in the way of these types of things might be politics. But I guess my bottom line is I think we are now at the inflection point of the great normalization of contingency.
Bruce: Right? Great. Well, that's a great point to end on, a very positive note to thank you for that. Thank you so much for joining us today. Where should our listeners go if they want to hear more about Brightfield?
Jason: Brightfield.com, there's a number of great market research reports, of course, a lot of great marketing as well, but there's also a slice of a blog resources and other market report and resources that will give more details to some of the things I've been saying.
Bruce: Great. Thank you, Jason. It's been an absolute pleasure. Thanks again.
Jason: Thank you, Bruce. And thank you for your partnership, it's been an honor.
Bruce: To learn more about AGS, please check us out at allegisglobalsolutions.com. You can also send questions for me or our guests, just tweet us here @AllegisGlobal, with a #SubjectToTalent or email us at email@example.com. And if you enjoyed our podcast today, please subscribe, rate us and leave a review, until next time. Cheers.