Art Steinmetz, (formerly) @ Oppenheimer Funds | Communicating with Execs | Data Science Hangout

Happy Thursday, everybody, and welcome to the Data Science Hangout. Hope everybody is having a great week. If we haven't had a chance to meet yet, I'm Rachel, it's so nice to meet you. If it is your first time joining us at the Hangout, let us know in the chat so we can all welcome you in as well.

This is our open space to chat about data science leadership, questions you're facing, and getting to hear about what's going on in the world of data across different industries. So every week, we get to feature a different data science leader as my co-host here to help lead our discussion and answer questions from you all. So together, we're all dedicated to making this a welcoming environment for everybody. I love to hear from everyone, no matter your level of experience or area of work or background.

It's also totally okay to just listen in if you want. There's also three ways to ask questions and provide your own perspective on certain topics too. So you can jump in by raising your hand on Zoom. You can put questions into the Zoom chat and feel free to just put a little star next to it if you're in a coffee shop or something or your dog's barking and you want me to read it instead. We also have a Slido link which my colleagues will share here in the chat in just a second where you can also ask questions anonymously.

We do share the recordings of each session up to the Posit YouTube and the Hangout site so you can always go back and re-watch or share with a friend. But with all that, thank you all for joining us again this Thursday and thank you so much Art for joining us as the co-host. Art is a private investor and former chairman, CEO, and president of Oppenheimer Funds. And Art, I'd love to kick things off with having you introduce yourself, maybe sharing a bit about your journey and also something you like to do outside of work too.

Sure, sure. Well, thanks very much for inviting me. I'm speaking to you all from New York City where we have yet to see winter start. Supposedly we might get some snow this weekend but we haven't had any snow at all this winter so that's strange. I'm sitting here in my study and I'm drinking out of my Gigi plot mug which was designed by Selena Carter. She graciously let me put this design on what is one of those sites that lets you print out your own t-shirts and stuff like that. So I had myself a coffee mug made with that.

Anyway, so I'm technically retired. I used to work on Wall Street. I was a money manager for most of my time before I went into a leadership position at Oppenheimer Funds. Then I stepped down when we sold the company to Invesco back in 2019. Since then I've been involved in a lot of this and that board work, philanthropic work but what I love most is being a data science hobbyist.

I started working with R, the language, back in 2006, I think it was, and it just grew from there. I've just really enjoyed being able to use it first in my professional work and then now more as a hobbyist. As part of that, I really enjoyed becoming part of the R community and really appreciating the joys of open source software and its ability to put power into people's hands, to really get good work done, to do good analysis.

Like anything else, you can do lousy work and it's like, as we've seen with ChatGPT, which speaks very authoritatively and confidently but sometimes spews nonsense. It's certainly dangerous to put beautiful charts and graphs into people's hands that tell the wrong story and I think that's an important narrative around how open source propagates around in organizations. That's a cautionary lesson. But anyway, we can talk a little bit about that.

Anyway, that's what I do now. You can see my blog at artsteinmetz.com, where I share my data science experiments. I have to stress, I'm not a data science professional, right? I'm an investment professional, so I am very much an amateur, with all the caveats that implies. But again, I really enjoy hanging out with smart people like you folks who are in the main actual data scientists or actual practitioners professionally.

So, Rachel said, so what do you do for fun or something outside of work? Well, again, since I'm retired, I spend a lot of my time playing around with R. While I'm sitting at my desk, I like to ski. I like to hike. I was a home brewer. I kind of retired from that when I moved from New Jersey back into New York City. But I can talk at length about IPA recipes, if anybody wants to do that.

So, but again, one of the things that is nice about playing around with open source software is it's not an expensive hobby. You know, guys are famous for spending enormous amounts of money to buy gear for their hobbies, whether it's fishing or cars or whatever. Fortunately, open source software is whatever I can come up with, build in my head is what I'm spending money on. So anyway, that's enough about me.

I love it. I know the mountains are getting a lot of snow. Hopefully this weekend, are you going to head up skiing? Yeah, well, I was in Big Sky, Montana at the beginning of January. And, you know, at 64 years old, I probably shouldn't do the double black diamonds. But I went down one that had thin cover. And I hooked my armpit around a tree stump. And I tore my pec. So I had orthopedic surgery a couple weeks ago. So I've got to keep staying a sling for another few weeks. But hopefully I'm going to get a little late season skiing in this year when I go back out. Hopefully won't do it again. Tear myself up again.

Journey into data science and R

So I know when we were talking yesterday a little bit, Art, you were mentioning your journey into data science started when you became a little bit frustrated by the limits of Excel. And I'm just curious to hear about how that journey first started and any tips you have for people who are also trying to help others in their organization who would love to learn too.

Yeah. And the disclaimer here is partly to say it was a different time then. I mean, a lot of times when we talk about cultural issues, we talk about, well, it was different back then. And it was different back then for data science. Because back then, employees at my firm couldn't install any software they pleased on their computer. And that is certainly not true in any responsible organization today.

But it was a different time. So even though I could have put in the budget request for MATLAB, I'm just inherently cheap. And so when I learned about R, I was really interested in that. But I had become frustrated with the limitations of Excel. Playing around with like spreadsheet geeks will do, I had these giant spreadsheets. And they had a lot of VBA macros, visual basic macros involved with them. But they were increasingly difficult to maintain.

And I would go back and pull up a project I had worked on several months earlier and was not able to figure out what the heck I was doing or what I was trying to accomplish in that thing. Or if I said, well, oh, I could use my former work because now this new project I'm working on is kind of like that. So let me reuse some of my old work. Figuring out what I did back then, what former me had done, just was a huge pain in the neck. So I was frustrated by that probably. That was probably my biggest frustration.

And in addition, just the power of Excel when working on large data sets with lots of computations was also not, you know, I was running up against the limitations there. So this was, but again, this was just me. There was nothing I would say I was in production for other people to use. I had dashboards with graphs and stuff, and I would share those around in meetings. But it was really just me. So it wasn't really an enterprise application.

But anyway, so I discovered R and was hooked immediately and was able to do some really cool work. But I was an island in the organization. And I think many of us who use open source software may feel that way. And, but slowly but surely, other people started asking me, well, how did you do that? Or that doesn't look like an Excel chart or something like that.

And at the same time, my organization started becoming, in general, more quantitative. You know, we were, Oppenheimer was, and now is part of Invesco, is a, you know, a money management firm. We manage portfolios of stocks and bonds and whatever else. But it's, we were very much old school fundamental analysis, which is great. It just isn't, doesn't come out from the quantitative side. It's a much more touchy-feely type of thing.

And, you know, like the rest of the industry, we all started becoming more interested in applying, you know, computational power to huge amounts of information. Because the price of computation is going down, the accessibility of huge information was going up. So it just was part of a natural trend. And, you know, I was able to be, you know, I was more fluent than the average portfolio manager in my firm with the techniques needed to work with that data because of just my freelance work with R. So I think that helped me out.

That did not play a part in my career advancement. I don't want to suggest that at all. But I also happened to advance further in my career at Oppenheimer, ultimately becoming CEO for the last five years of my stay there. But, you know, I continued to play around with R, even though I have to admit, I was playing hooky from my real job when I was doing data science work. And, you know, I would do analysis of our funds, analysis of our competitors' funds on various types of performance metrics and so forth. But, you know, as CEO, that was not my job. There were people I could ask to do that work for me. But that's how I, you know, that was sort of a release for me and a little bit of an unwind for doing that.

But at the same time, people in my organization understood that, you know, I was very sympathetic or receptive to using these kinds of tools in the organization. And I was constantly evangelizing about or thinking about ways that our organization could apply more quantitative techniques to upping our game all across the business. So, you know, people, other, you know, there are other islands around our company of people who are using this kind of stuff, whether it was R or Python or other tools. And they knew that I was, you know, very much a champion of those sorts of things.

What that allowed us to do is from their standpoint, practically, it allowed them to fast track permission to use these tools or to get software updates or things like that. Because as I said, you know, we all got to the point where you couldn't just install whatever the heck you wanted on your computer. So at least they knew that the IT guys were going to be sympathetic to their requests. You know, and this is a dichotomy I'm sure many of you see in your organizations all the time. There's an inherent tension between the high priests in IT and the individuals on their desks who want to be free to use these open source tools. And, you know, having leadership in the organization understand and sympathizing with that is very important.

One of the great things about Posit, of course, is that, you know, Posit has evolved to really facilitate easing of that tension, if you will, by providing, you know, walled garden types of tools where there can be auditing and control of the versioning of all the various packages and languages and so forth that people have on their computers. But also letting people have access to those tools and not be constrained by a large bureaucracy for every little thing they want to use. So, you know, I think Posit's doing great work on, you know, reducing friction in the enterprise. So I like that.

Open source in sales and the redemption rate story

So again, we started using these tools more, but here's the other ironic thing. So I'm an investment person. And like I said, I would like to use quantitative tools to look at time series of interest rates and stock prices and correlations and capital asset pricing model metrics around investments and stuff like that. And while I was an evangelist for this stuff, the people that were getting more excited about this were not my fellow portfolio managers and investment analysts. It was the guys in sales, the people in sales that were getting more interested in this.

Because like the investment people, they were getting, you know, terabytes of data from their organizations, you know, stuff that would come through our CRM, we used Salesforce, or the advisors who would sell our funds. So, you know, we were the asset managers. And then we would have advisors or stockbrokers, financial advisors, who would sell it to retail individuals. And they, for a fee, would let us see their customer data, what people were buying, what people selling, you know, whether they were buying more growth stocks or value stocks or whatever it is. And our salespeople would find those tools very valuable. But it was a data dump. It was a deluge coming from different sources. A lot of places where it was unstructured data.

And even when it was structured, the formats from Morgan Stanley were different than the formats from Merrill Lynch and this kind of stuff. But our salespeople, you know, they got it sooner than our investment people that we can use open source data science tools to make some sense of this. So we did, and even today, they're the ones that are at Invesco, my former colleagues at Oppenheimer, now at Invesco, are the ones pushing the envelope on this, on the distribution organization.

One of the things that is a crazy thing about the mutual fund business is, you know, people buy mutual funds and then they decide to sell them. Maybe they just sell them because they need the money for other purposes. Someone dies and the money gets distributed in the estate, or the performance stinks and people stop liking the fund. And so they say they want out. All of those things trigger redemptions.

Now, we are, the industry overall has a background redemption rate of about 25% a year. That means just to stay even with how much money we manage, we have to go out and resell 25% of our total asset base every year. So that's billions and billions of dollars. And the other crazy thing is, well, it's not crazy at all, but we sell, we pay commissions to salespeople. So getting a dollar is much more expensive than keeping a dollar. So we could drop a lot of money right to the bottom line if we could cut our redemption rate. So I sat down with our salespeople and that became kind of a thing I was really pounding the table about.

So they went and we put together a data science team of about nine PhD data analytics people. And this was all coming out of the sales budget. They cheerfully paid for all this. And we started looking at how can we reduce our redemption rate? At that time, our redemption rate was running above the industry average of about 27. And the nice thing when you're a big firm, you can run A, B tests. We tried different kinds of interventions, different kind of marketing materials, changed who we visited and how much time we spent with people, tried so many different things. And we found out what worked and what didn't work.

But the bottom line is that we were able to cut our redemption rate in a couple of years from 27 to 23%. So again, that was tens of millions of profit dollars that dropped to the bottom line. And again, we were using R and RStudio to do these redemption control tests. So it was just tremendous.

But the bottom line is that we were able to cut our redemption rate in a couple of years from 27 to 23%. So again, that was tens of millions of profit dollars that dropped to the bottom line.

Communicating with executives

That's amazing. I remember seeing the blog post where you mentioned that. So I want to make sure to go grab that and put it into the chat for everybody too. But something else I wanted to ask you is on the Hangouts, we talk a lot about best practices for sharing our work with executives. And it's rare that we actually have the executive here on the call. So as a former CEO, I was wondering if you have any tips for all of us on communicating with executives and sharing our data work with them?

Yeah. The two key points. One is, you know, data visualization is the most powerful way to tell, get a narrative across quickly. So spend time learning good visualization techniques. Even if you're a math whiz and a coding whiz, understand good visualization. It's an aesthetic issue. It's an art. And, you know, there's lots of blogs. There's lots of literature that talk about good visualization and bad visualization. And, you know, don't use a linear scale when you should be using a log scale. You know, simple stuff like that.

Because I've seen so many bad charts. And at worst, they, you know, you're going to lose your audience very quickly. Well, not necessarily even at worst. At best, maybe you're going to lose your audience. At worst, your audience will draw the wrong conclusions and make bad decisions. You know, Richard Feynman, the physicist, was on the Challenger disaster panel. That was a space shuttle that blew up on reentry. And, you know, his, one of the big takeaways in that inquiry was that PowerPoint was to blame.

So the other point I would make is don't use, and I learned this from reading about the history of that inquest, is don't use a PowerPoint presentation when you should be using a memo that people can read and a narrative. Because all of you are familiar with the problem of seeing a PowerPoint slide. So you've got it, landscape orientation, and you've got this incredible amount of text on, you know, 12-point font text filling up a slide, you know, as if it's a slide deck. And sometimes you're subjected to sitting in a conference room having that thing with two minutes worth of text on a slide that you're supposed to process. Of course, nobody looks at that at all. And, you know, they're listening to what, hopefully, what the speaker is saying.

But it's a complete waste of everyone's time to create and display this giant PowerPoint, this massive quantity of text on a PowerPoint slide. So don't use PowerPoint when a pre-read memo would do the job. And, yes, some people aren't going to read the memo. But treat your audience with respect anyway and tell them you're going to assume people have read the materials. And then have your, if you're going to have a PowerPoint or a slide deck, you know, have, you know, three, four bullet points most on a slide, and then have compelling visualizations that tell your story. That's going to earn you far more respect than showing that you've crammed your PowerPoint with a million words.

Data visualization is the most powerful way to tell, get a narrative across quickly. So don't use PowerPoint when a pre-read memo would do the job.

Model risk and regulation in finance

I'm curious, Art, if in finance or in portfolio management, if you have to be concerned about things like validated systems, software quality, the sorts of things we talk about here a lot with regards to finance, not finance, with pharma and med device and stuff, are there places in the business where you've got lots of landscape to experiment with tools? And then other places where there are folks doing work that have to be much more diligent about, you know, being validated or being controlled? And just because I don't know anything about it, are there like, you know, enterprise applications that are made for purpose for that specific kind of stuff? Or does every shop have their own tools that, you know, somebody made and everybody just keeps using them?

Yeah, good question. Like health care, finance is a highly regulated industry, you know, particularly as you're, you know, consumer facing. So what we have enormous amounts of regulations over what we can say to people and what our responsibilities and potential liabilities are, are when we misrepresent what we are saying to people. Like if you say you can't lose money, this is a guaranteed win or something like that, to state the obvious. But to your point, yes, there is a huge problem in our industry around model risk.

And if we use any model, and that's very liberally defined, by the way, to help us with our investment decision making, and we come back at some point or it's disclosed somehow that that model had a coding error, a bug, and gave an incorrect answer because of a bug in the code, strictly speaking, we are liable for making investors whole for any losses on that. And there have been many instances of lawsuits around exactly that. So we have to be very, very careful.

And, you know, as you all know, there's no such thing as bug-free software. There will be errors and problems. And it's incumbent on us to self-report whether or not that caused losses. And the SEC has leveled hundreds of millions of dollars of fines over the last decade to firms for model errors. And invariably, you know, whenever people lose a lot of money in an investment that someone else is managing, there will be lawsuits.

So in 2008, there was the great financial crisis and stocks went down. Everything went down. There were lots of problems. And there were lots of lawsuits. And the heart of many of those were, I lost money and I'm mad about it. I want somebody to make me whole. So now if I'm hiring a lawyer, how do I craft a case that says it was somebody else's fault other than just the market and I therefore should be made whole? One of the first things they go looking for, whether or not there were models involved in making the trades, whether any robots were doing the work, and then were there bugs in that.

So there are, every model is different. Every process is different. There's no automated way of saying, is this bug free? Obviously. But we had, there are outside all the big accounting firms will audit your models if you pay them. So that's what we had to do. And between you and me, I don't think if there was some subtle coding error that switched a positive sign to a negative sign somewhere, they're not going to discover that. But nonetheless, that becomes a legal defense that you say, well, I had an auditor review the code. But it is a huge issue in the industry.

Handling buzzwords and communicating upward

But one of the first anonymous questions was, how do we help leadership, say VP level, really understand what it is so that they don't just succumb to the buzzwords like AI, machine learning, big data, for example?

Yeah, well, the buzzwords aren't useless. They, to the extent that a senior leader in the organization is going to say, if a senior leader in the organization gets all excited about AI and is worried about being left behind, then that means they're going to give you more support for your work. They're going to listen to you. And then you can, you know, since maybe I'm sitting in my nice corner office and I'm just hearing the buzzwords, I'm like the boss at Dilbert or something like that. And you can, but if I'm willing to listen, then you can give me a deeper understanding of what's involved. So I think the buzzwords are your friend.

The important thing, I mean, if it creates any sort of urgency or, you know, creates more opportunity for, you know, data scientists to elevate the visibility of their work. The most important thing though is you have to be honest. So when I, as the boss, call you into my office and say, I heard about this chat GPT thing. Let's fire our entire, you know, our phone customer support people and just turn it over to chat GPT. You have to be honest about what these things can and can't do.

And one of the things that, this will never change. This has been true since the times of Hammurabi that, you know, subordinates will come into the boss's office and they'll blow sunshine up your skirt. And, you know, avoiding that, getting an environment, and this is important. This is what the bosses have to do. What the leaders have to do is create an environment of trust where people can deliver bad news about what something can and cannot accomplish to the boss. And they can't be afraid of saying, you know, sorry boss, you know, chat GPT isn't the be all and end all.

And, you know, I mean, again, I experienced this firsthand, but, you know, people, if I would say, oh, this sounds like fun, let's do it. Or this looks like this will be interesting, let's do it. People will start scrambling around and try and do it because the boss likes it. And well, on one sense I'll say good because I'm the boss. I'd like to have some influence over what people do. But you have to have the courage to say, no, this is not going to work if it's not going to work.

A Shiny app in production and getting off the island

The description you have of your experience at Oppenheimer, I know it's almost backwards for a lot of financial firms and definitely for other big organizations. This is data science excellence and opportunity pushed from the top down. A couple of questions around that. Have you had any success in discussing with your CEO peers this opportunity and have you won them over to that? And then on the other side for folks working up the chain, what do you think is a successful way to convey that opportunity upward?

Yeah, well, I think, like I said, it was a different time then, but the world has come around to understanding the power of big data. Again, it's a buzzword, right? Big data, data science. And my alma mater, Denison University, has launched a whole school around data science, data analytics, and it's a liberal arts college. I was a Russian major in college. But I think CEOs today very much, even if they don't necessarily understand it, going back to the buzzwords, they think it's important and they need to pay attention to it.

So I don't see much resistance to that. You know, it's funny though, like I said, the investment people in my organization to some extent still are very old school. They are very wary of quantitative techniques because there are many quant funds out there and some of them do great work. We are fundamental investors. So it's reading annual reports and talking to management, visiting the factories or whatever it happens to be. That's fundamental management. We try to beat the market doing better research in those lines. And there was a lot of resistance to quantitative techniques because we're fundamental. We're not gonna let the robots take over.

Now, the argument I would make, I did make to, you know, when I became chief investment officer and then CEO, the argument I would make to the investment people, and I was not going to push anything down their throats. I was trying to evangelize to make them want to do this. But I said to them, look, this isn't about the robots taking your jobs or we're gonna, you know, I mean, literally, that's what they were worried about, that we were just gonna make these quant funds because we had some. We launched because they would, nobody would come around on the fundamental side. We launched some quant funds.

The argument I made with limited success was, look, I'm not talking about getting rid of you. I'm saying you have been digging ditches with a shovel your whole career. Now, I come and give you a backhoe. You're not gonna use it. You know, it's a force multiplier if you can apply these quantitative tools to suddenly process far more data than you used to be able to. It's making your analysts far more efficient.

But, you know, like I said, in terms of the rest of other CEOs, other people in asset management, everyone's come around, I think, to that. And, you know, how do you get, so I mean, I would be interested in hearing what other people say about this because you folks are more mid-career, early career than I certainly am as retired. You would probably have a better sense of the receptivity in your organization to using more powerful quantitative tools.

I noticed a few people resonated with when you said living on your own island. So a lot of people were feeling that too. You know, what I hope, again, this worked to some extent in my organization, and it's happening right now at Invesco too, I should add. People look over your shoulder and say, wow, that is cool. How do you do that? Or I see a way I could apply what you're doing there.

So for instance, at Invesco right now, and this started after I had left, one of their folks, again, in sales has created a Shiny app that if one of our customers, which is to say our brokers, shares their customers, which is the end investor, their portfolios with us, what other funds they're invested in, what other stocks are invested in, whatever, we can run an analysis on that to help them optimize their tax situation, when they should take capital gains, when they should take capital losses, what they should sell to realize a loss or a gain, and then if they want to keep the portfolio complexion the same, what should they then buy on the other side to keep the portfolio complexion, the overall risk or aggressiveness, whatever it happens to be, the overall portfolio complexion the same. So he built a Shiny app that does that.

So all of our salespeople, all of Invesco salespeople can look at all of the advisors that they serve and look at that, and that generates discussions. And salespeople love to have opportunities to call their customers up and have an interesting discussion and say, look, here's an issue you have that I think I can help you solve. That's salespeople love getting that kind of a lead. So it generates that sort of thing. And then if the discussion goes further with one click, they generate a detailed PDF report on exactly what the gains and losses are across the portfolio. And the salespeople just love that tool, but it's one individual. And my fear is that it dies with him. If he leaves the firm, who's gonna maintain that? It's an island.

And yet it's a Shiny app in production that has become an extremely popular tool among 200 salespeople. So it's an interesting opportunity. And this person is not naturally a real outgoing, gregarious individual. So I've been trying to encourage him to think about ways that he can let other people know this is a tool that he built and you can have this kind of a cool tool also in your area. So getting them off the island, I've made it a... I have no vested interest in this. I've made it my mission.

It's really interesting to think about maybe some of us could be targeting our work to different groups of people within the company like sales, which we may not have thought of before. And I see somebody had asked an anonymous question that sales was so interested in the data was really interesting to them. And they asked, were there any groups that you were surprised weren't interested? And why do you think that was?

Yeah, well, again, I've already told the story about the investment people. And they were pretty skeptical. The other group as well, is it worth the cost? What's the IRR on this investment? And again, in the anecdote I shared about redemption control, the IRR was strongly positive, but it's not always gonna be that way. But I think it is important to have a clear eyed understanding of what the IRR is going to be on a project like this.

What I would not be concerned about is whether there are enough projects for a data science team. And most organizations, you might have a particular use case when you spin up some kind of a data science organization. But like I said, very quickly, other people will see the cool stuff that is happening and want in on it.

Book recommendations

So, oh, one of the questions I saw going back into the chat was, what books you might recommend that help you throughout your career to develop your mindset as a data science CEO? Well, let's see here. I already talked about Tufte, right? So, he's got a whole bunch of books, but this is the one that started it all. The visual, I got this, the visual state of information. And this guy is kind of a lovable curmudgeon, kind of a crank, but he's passionate about making sure that your visualizations convey the information that you want them to convey. So, his whole series of books, they're all classics in the field.

Another one to pull off the show that I think is a ripping yarn is Against the Gods by Peter Bernstein, the remarkable story of risk. For anyone who wants to understand how we went from thinking everything is just in the hands of the gods and it's all fate to being able to price derivatives using normal distributions and understanding probability. This is a great book and this is not dry in the least, but I think it's very useful for understanding how the history of risk evolved. The ancient Romans were inveterate gamblers but they had no sense of probability. So, if you wanted to go play cards, play poker with the Romans, you could really fleece them. But it was Blaise Pascal who was a philosopher and mathematician who really developed probability and he did it because he was a gambler.

So, you know, there's another book that got me thinking about it. It's out of reach because it's over in my business books section. But there's another book called The Outsiders by Richard? His last name is Thorndike. I'm not sure his first name is Richard, but he has a great book for thinking about how CEOs should think about maximizing the value of their businesses. And that had a profound influence on me.

One of the things that is true, if you look at most mergers of publicly traded companies, the vast majority, three plus three equals five, which is to say that the mergers are usually destructive of shareholder value. And I won't name names, but I've experienced this so often in my career. And the problem is the CEO of the combined entity now gets to say, I grew the company. It's much bigger now. And he, of course, he or she, obviously, usually a white man, is running a much bigger organization. And that could be used to justify the board paying them a lot more generous compensation. And he gets the reputation as a deal maker.

But the problem is if the company, if the two companies were worth six apart and they're now worth five, you lost money. You didn't do a good job. This book, The Outsiders, really cracks that open and does a lot of, as business books like to do, there's a lot of anecdotes and case studies, shows a few CEOs who understood that and bucked that trend and really thought about how to grow the wealth of their shareholders, which is what they're accountable to.

One of the things that whenever companies get mashed together, there's talk of cost savings, right? You have synergies, economies of scale, et cetera. And that's all well and good, but there's hardly ever any talk about revenue synergies because there seldom are any, even though there will be some genuflecting in that direction. Anyway, that book, I thought was very formative in my thinking as someone who was running a company, someone who bought other companies and who ultimately sold my company to another company.

I love this, just grabbing the books off the back. It's like book club with art here. One of the things, this is, I'm a book, I have a fetish book. I fetishize them as objects and I'm a book fetishist. And so I really like sitting in my library with a Manhattan in the evening and reading a book.

Data literacy and Rock the Street, Wall Street

So I know something that you are also really involved with is enhancing data literacy and mathematical literacy. And you've been involved with Rock the Street, Wall Street's mission is to equip girls for the skills to succeed. And many other organizations. And I wanted to give some time to talk a little bit about that too and see if there's any that you'd recommend we get involved in or ways to get more involved.

Yeah, thanks for asking that because it's obviously a worthwhile topic. I don't have to sell anybody here on the value of being data literate or understanding basic math. But there's a weird cultural problem I think we have certainly in this country and maybe in a lot of other countries is that it's okay to be bad at math. People are perfectly happy, maybe even proud to say, oh, I'm bad at math. Oh, you know, it's like, no one says, oh, I'm bad at reading. Oh, I'm proud to be illiterate. Numeracy should be something that we all are very proud of having.

And the consequences of not being numerate are horrible. And you see, our country has become so polarized politically and the rhetoric that goes back and forth, so much of it is grounded on, so much of the persuasive power of that is grounded on people not understanding numbers, I find. And not being willing to and being very skeptical of science. Being very skeptical of experts. Like that's, those are elites or something like that. When, you know, all of you probably went to college and you learn a lot of stuff and you went on to graduate school or in your careers and you're very proud of being an expert in your field and all the knowledge you've created. And the idea that that just should be devalued somehow because people need to be, feel good about being ignorant is an unfortunate state of affairs.

I wish we could get back to a point where knowledge is respected in whatever field that happens to be. But certainly one of the ways to do that and make all of us more successful in our own lives is to make sure that kids understand, you know, know that it's cool is too strong a word, but cool to know math and be proud of having math skills. So, you know, I'm, I was involved with the National Museum of Mathematics, which their mission is to make math cool. And it's not a kid's museum, by the way. They have tons of kids programs and probably most of the visitors to the museum are children. But, you know, it really shares the wonder of mathematical concepts at an adult level as well. So it's, if you're ever in Manhattan, it's absolutely worth a visit because it's a very cool exploratorium kind of museum.

And, you know, one of the problems is in, obviously in poor communities, underprivileged communities, whatever word you would like to describe them, minority communities, they're, you know, the schools are generally poor and Rock the Street, Wall Street brings a finance curriculum into underserved schools and using volunteers. So the business model, if you will, of Rock the Street is we get a corporate sponsor and a corporate volunteer from that sponsor and they teach a, call it a finance course to kids. And one of the things, purposes of that is to get more girls because it is focused on girls. And I use the word girls because this is high school. Get girls interested in careers and financial services.

And while we know less than half of the girls are actually going to financial services, but every single one of them is going to get out of this program with basic financial literacy. They are going to understand that the credit limit on your MasterCard isn't an asset. And, you know, paying 25% a year in compound interest on your credit card balance is a really bad problem to get into. So, you know, those kinds of basic issues around life, financial life skills is an important part of what they do.

And of course, they go into these lesser communities. And one of the things that just like, it breaks my heart and why I'm passionate about this is, you know, you've all heard this term, the cycle of poverty. And there's this cycle of ignorance exists in these communities as well. And in many cases, you know, obviously the family situation is not a traditional two-parent family. And even when it is, you have issues where there isn't respect for learning in some of these families. They say, why do you want, that's not for you. You don't have a chance. You know, the system is rigged against you. Why bother?

And, you know, fighting against a familial culture in some cases to give these girls the belief that yes, you can achieve. That doesn't mean the road is easy, but this is possible. And it just takes hard work and learning and you can achieve that. And so Rock the Street does great work, you know, swimming against that tide of ignorance. So it's been heartbreaking to see some of this stuff, but also so inspiring to see what the change that it can make in some kids' lives.

But, you know, girls in particular are underrepresented or women in particular are very underrepresented in asset management. So that was one of the things that got me interested in Rock the Street to begin with. We need lots more representation among people of color as well, but women also are, again, very underrepresented. So that's just trying to get a little more inclusivity in finance.

But also what Rock the Street in particular recognize is that girls self-select out of STEM fields very early. The cultural background is such that women are told explicitly or implicitly, girls are told explicitly or implicitly that they're not good at math. They're not good at science. So that's not for them. And again, it doesn't have to be explicit, but we find that they self-select out of STEM very early in high school. And that's why it's important to, you know, there's lots of groups that help give women early career support or in college. There's Girls Who Code, for instance. That's more college level. Getting down into the high school level is critical because of this problem.

Open source ecosystems and career advice

And one anonymous question was, what's the number one reason you think, whether it's financial services or any company, should move to open source languages like R and Python?

Yeah, because of the ecosystem, you know, with R or Python, it's the packages that will do anything you want. And these packages, of course, are constantly peer reviewed. There was a notion 10, 15 years ago that commercial software was better because you had people who were paid to mind it, whereas open source, anybody could do anything and they could, who knows whether the code is any good, which the truth was the opposite of that, right? When you've got a million eyes on something, errors are much, are spotted much more quickly. Now there were commercial software firms that had a vested interest in perpetuating that notion.

There was a blog post, it was taken down at, I guess it was SaaS, where they were ripping open source code and they made a very kind of disgusting dig at Hadley Wickham. They said, you know, do you want trained professionals auditing your code or do you want some drunk guy in New Zealand? Which I thought could only mean Hadley. That Hadley's from New Zealand, I have no idea whether he enjoys liquor at all, but why would they say that? It's ridiculous. But, you know, it just shows the nastiness that was out there. But it's clear that the huge ecosystem of packages maintained by passionate individuals acting as a community is what gives open sources power.

One other anonymous question was, as a data person, what are the best ways to develop domain and business knowledge? They said, it benefits me to talk with the business people, but not systematically. Yeah, that's a, you know, kind of a tough problem because it requires you being an advocate for yourself in the organization, and not all of us are equipped to do that. I wasn't. My path to the top of the organization was very serendipitous, and I'll spare you the details on that as well. But, you know, you need to be able to get out there and advocate for yourself. So you have to show up and stick your nose in when there are discussions about the business.

And, you know, hopefully you can present yourself as someone who is, you know, eager to help solve problems. If, you know, if you say, look, I'm here because I want to understand the problem so I can do better helping you, you know, however you want to say that. I mean, that's the kind of approach that always worked with me when people say, you know, because I certainly could have no objection to someone who's, say, in IT coming around saying, I want to understand how you run money better so I can understand your needs better. No one's going to say, oh, get out of my face. But it does require you getting out of your comfort zone, perhaps, and being an advocate for yourself.

So one question I did want to ask you before you go is, what is the best career advice that you've ever received? You know, it was, you know, just, you know, don't look to your career advancement. Yes, you might be an ambitious person, but don't think about everything in terms of, how is this going to advance my career? Think in terms of how this will help others in their career. It's kind of enlightened altruism, if you will, because if you gain a reputation in your organization for someone who is boosting others, that is going to grease the skids for you in a very powerful way. Someone told me that once, and it was absolutely true, you know, because no one likes that guy who, you know, he's already looking for his next job.

Think in terms of how this will help others in their career. It's kind of enlightened altruism, if you will, because if you gain a reputation in your organization for someone who is boosting others, that is going to grease the skids for you in a very powerful way.

Well, thank you. Thank you so much, everybody, for all the great questions. Thank you so much, Art, for joining us here. This has been a blast. It was a lot of fun. I appreciate it. I hope the rest of you enjoyed it as well. Thank you so much. I'm going to share your LinkedIn here in the chat so people know where to find you. Anything else you want to leave us with? ArtsteinMetz.com. Awesome. I'm going to re-share that again in the chat too. There we go. Thank you so much, Art. Have a great rest of the day, everybody.