How Data and Distributed Databases Drive Fintech Innovation | Catalysts of Change Series | DSS 2021

That's great. First of all, congratulations to Karthik, to you. You've built a great team here and a fabulous product. I think having Bill join you is a great decision. Bill and I have a relationship for many years, and I know the combined leadership of both of you really given to the company that you guys are leading today. I know will produce a great result. So that's terrific. And looking at my own life, you know, we talked about Morgan Stanley. I was working. I came out of college with what I thought was going to be a degree in computer science, and right before I graduated, the semester before, I changed it to a business degree because I wanted to make sure that everything I learned I could turn to be something commercial. I took a job working for the State Treasury of New Jersey. They taught me how to enhance my software skills with some things that I did there. But after about 18 months, the pace for me just wasn't the way I wanted to live my life. My wife was literally, I think she was eight months pregnant, and I went home and I said, 'I'm gonna leave my job.' She said, 'What are you talking about?' I said, 'A friend of mine called me and I went up for an interview. I didn't say anything because I didn't want you to get nervous. I went up for an interview with Morgan Stanley. An hour interview turned into five hours. They grilled me non-stop. I was able to answer every technical question and build some things for them while we were talking. And I'm gonna go do this job.' She said, 'Great, what are you getting paid?' I said, 'It's a 30-day consulting agreement.' She said, 'What do you mean? What happens after 30 days?' 'After 30 days, if they don't hire me, I don't have a job.' She said, 'Oh my God, what are we going to do?' I said, 'Don't worry.' And that's the way I honestly, if you take me back to Morgan Stanley, that's really where my head has been my whole life: very confident but humble, and in many respects, always wanted to be the absolute best at the things that I did. I decided I didn't know whether I was going to be a software engineer my whole life, but I decided I was going to get as many experiences in this shorter period of time as possible so that I could figure out how to put those together in a unique way that would be differentiating in the marketplace. I started at Morgan Stanley as a DBA. At that point in time, I was learning older technologies, database and Natural. Then we moved into the distributed systems world with Sybase and Oracle, and then we moved into a whole series of advanced open source type database technologies. I did a lot of other things after that, including being the CTO, and that was really the first CIO of Morgan Stanley that didn't come from a business. That's why I was prefacing some of the comments earlier about staying commercial. I think my whole view on innovation and my life has been: find the diamonds in the rough, mentor and shape them to really be 10x performers, but focus them on real-life commercial results. Otherwise, these beautiful ideas will hang on the vine and never get mature. So what Morgan Stanley taught me was to do everything from being under a desk playing with wires, to being in a data center playing with wires, to writing software, to learning databases, to building training systems, to running large groups of people and learning how to get a company positioned. The favorite thing I did, Bill, you may not remember, but I was basically an extension at Morgan Stanley to one of the best tech investment banking groups in the world. In fact, Morgan Stanley is still one of the best tech investment banking groups. So I've been on the trading floor for many IPOs through the internet era, Google. The one that I remember most fondly because we wrote the Dutch auction software. I've been in most places in the world building trading technology. That's what shapes you. These experiences early shape people, and those are some of the things that really shaped me in the early part of my career.

Look, I think, Karthik, you probably subscribe to this as well. I look for highly inquisitive people. Most people I hire, I don't hire for their job; I hire for a career. Literally before I came here today, we were talking about popping up an innovation center, and I was talking to one of the top recruiting firms in the country about what I want to hire, what I'm looking for. We spent time, and they brought me a whole series of people and thoughts and resumes, just like somebody might do if you were building a house: here's 40 different options. Honestly, I took them and I didn't rip them up; I put them on the side of the table and I said, 'You got it wrong. This isn't the way you build a team.' What you have to focus on are people. They were a lot of big bank performers, and everybody's expensive, just like software companies are expensive these days. But I look for people who are sort of in the midpoint of their career. I measure how many experiences they've been able to attain, how many things can they talk to me about, and then how many things can I talk to them about? Meaning the inquisitiveness. What are they interested in, not only what they're doing today, but what are they interested in when they leave the office? That doesn't have to be software and technology; that could be anything. But are they inquisitive? One of the first things I often say to people when I walk into a room when I have a problem is, 'Your opinion matters. Sometimes I say, if you don't have an opinion, you can leave now because I don't want to embarrass you. But if you're in this room helping me solve this problem, I don't need problems; I need solutions. All your ideas are equal in this meeting. Let's see what we can do to advance.' So for me, I think important things for some of the people in your company and listening to this would be: absolutely number one, if you're talking to me in a conversation, I care about who you are, what you've learned, and how you want to conduct yourself going forward. The inquisitiveness piece. The second thing I look for is, have you been able to collect a number of experiences? Some people just want to be narrow and deep. I'm okay with that, but what kind of experiences in that narrow and deep landscape? Other people want to be the other part of the T. I call this the T. Other people want broad. Whatever it is, you really have to think about your career over a timeline and make sure that everything you do leads to something. Then after that, I'm looking for diamonds in the rough. I believe, I mean, you guys know me and a lot of people at Yugabyte, but I'm a very humble person. The one thing that I do, I think people who are the rough diamonds, I turn them into the most beautiful, exquisite diamonds in the world. That's really, when I left Morgan Stanley to go to JP Morgan, I brought very few people. When I left JP Morgan to go to First Data, I brought very few people. I say that only because I didn't want it to be like the political system: the Democrats come out, the Republicans come in, everybody shifts. That's the lazy person's way. The goal is how do you find raw talent, how do you cultivate it and groom it over time, and how do you make it leadership, technical track leadership, managerial leadership in all aspects? To me, that's the most fun. That's why, what I was joking earlier about my wife, I say to my wife all the time, 'Forget what people say about me. Morgan Stanley, JP Morgan, First Data, now Fiserv. You can't do four and be successful unless you have a formula. You can't depend on luck, and you can't depend on the same five people to do it. You have to have a formula.' I feel confident that that formula I just described works for me, and I think that gives good input for people listening on how they can fit into it.

Well, I don't know whether I can give you a combination of answers because the ones that I think are important are also the ones I love. The ones that excite me are the ones I like to talk about. I am like on AI in general. I believe, the reason I left JP Morgan to go to First Data were numerous, but one of them was that First Data, because it was a leader in its industry, 40% market share in merchant acquiring and payments compared to the next competitor which was at maybe high single digits at the time, meant that it had data and information that nobody else in the world had. Visa, Amex, Mastercard, Discover, debit, credit, gift, EBT, you go. It had a wealth of information that nobody else in the world had. With that in mind, I always felt like I always said to myself, 'What Google did was turn the landfill into the gold mine. Could I take First Data's landfill and turn it into a gold mine?' Some would say maybe we didn't, but we sold the company for 20 or 40 billion dollars depending on how you want to count it. So I would say it's reasonably successful because when I got there, it was bankrupt and losing nearly four billion dollars a year. So the bottom line is, if you can create, if you can find assets that can't be rivaled, you have a starting point that nobody else has. Then the second thing is, if you can figure out how to turn that data into truly not only information that you can plot history for, but you could use it to predict the future, then you really have a significant advantage. Here we're now 12,000 transactions a second. We move money in some type of way 12,000 times a second. I know more than any economist in this country if I do my job right. I know more than maybe anybody else in the country as it relates to money movement, financing, and not just payments, but the way trillions of ACH move into bank accounts, the way people walk into businesses and spend money, the way they move money P2P. We got it all. So I believe AI, with some ML alongside of it, but I think AI is critically important. That goes back to my days of being a database person. It just stuck in my mind as most valuable. Look, I think digital is the new normal. For us, if you think about the fact that we serve 4,000 banks and millions of merchants, businesses are all digital now from the way they go to market. Banks, the only way to differentiate themselves really is through omni-commerce around the channel. So digital has to be core. If you're a small bank, it's pretty tough to deal against JP Morgan or Wells Fargo or Citibank because you can't have all those branches. But if you have the right digital solutions, you can compete with anybody on an equal footing in the world regardless of size and scale. You still need capitalization, but those are the key points. So for me, I think mobile and 5G are big movements because they can change the user experience while you could also bring incredible leverage to big and small. I always say make small big and big small. Big small means if you work in a big company, package your groups up into smaller neighborhoods and let them innovate at a rapid pace. Small big means once you do that, use this massive distribution channel like Fiserv has to get that solution to millions and billions of people around the globe. So I think those two areas are critical. And then the pace, the thing that you're working on, 12,000 times a second, I have to be perfect. I have to be trusted 12,000 times a second from a security point of view. I have to be reliable from a responsiveness perspective. I have to be global in nature. I have to have state at all points in time. I have to have complete resiliency and continuous operations. So the reason I'm doing this with you today, to be honest, is not only because I like you guys, which I do, but I believe that you're on the frontier of something that is very important. The fact that you're cloud native and have all these other capabilities means that the processing that you do, the information I keep there, the integration of my clients which are all over the world, just become infinitely easier. So to me, those are really some of the things that matter.

I do, I distinctly remember. Actually, one of the things that you waited to tell at the towards the middle or the end, that you actually had a background in databases, because I was surprised. Oh wow, man, this guy's just keeping up with everything, what's going on here. But yeah, I totally remember that. And for all your employees that are watching, I think Bill either was there or I told them afterwards. I said when we were done, I said, 'Not only do I understand everything you just said, but if you could do two-thirds of what you said, you're going to be an incredibly successful company.' I really, that's why I say that up front because I don't want people listening to think that I'm just being nice to you guys because we're on video. I said that to Karthik after a good 30 minutes of conversation where I let him talk probably for the first 20, and I said, 'Just talk as fast as you can. I want to digest and listen.' Then I gave him a bunch of opinions. But I was very, very impressed. I felt very optimistic about your company the minute I left, and I said to Bill afterwards, 'You made a really good decision.'

Yeah, I've used them all: Cassandra, Cockroach. I mean, I've been around all of them. Why another house? Why another boat? Why another car? I love my Tesla. It's driving as a service. It's software, it's updated, it's always, somebody's doing something every minute to make it state of the art, versus the traditional car you buy which stands still for most times. I think that if you want to continue to progress in a world, payments is the fastest moving sector. If you look at capital flows and investment, I think payments and fintech probably outpace most others, maybe with the exception of some of the eco and environmental activities, by miles. That means that with all that innovation coming, there are not only competitors coming right around the corner, but it means change, it means things are going to evolve very rapidly. So from where I sit, when I was at JP Morgan, my budget was $10 billion a year for tech, and I had 36,000 technology professionals. That's not the norm. That's not the world. That'll change too. Most companies have to figure out what's core and what's not. The thing that I believe that you're doing in reinventing is the new normal for transactional processing. There are some other competitors there, but you said, 'This is the thing I'm going to pick, I'm going to be the best at it, and I'm going to give you the things I wrote off earlier.' It's going to be global, it's going to be multi, it's going to be state, but maybe even multi-state. It's going to be continuous. It's going to allow me to have fragments of information searched and clicked. You've done all the things that every other great database company has done but unfortunately has not pulled together. You have pulled this all together, and as a result, it simplifies the world for me. It means that everything I build, because I'm in a cycle of building the new and the next generation, the next digital frontier, it totally makes sense to do it with a next frontier solution set. Look, in my whole life, I've been good and lucky. I think you have to have both. You're good and lucky. Why are you lucky? You're lucky because of your timing. Either your timing was picked by you or your timing was fortunate. But you have arrived on a scene with very good capabilities at a time when most people are thinking like I am right now about their future. That's why it's time for another one. If you were 10 years ago or five years ago or maybe even three years ago, I think your ramp pattern might have been a little bit less of a slope, and people would have to think about it in a more judicious way. But here, the cloud native component, your capabilities, and I think the problem set that businesses are facing right now, you are in the perfect storm in a positive way. That's why I think so. Again, there always has to be luck. You have to get lucky with timing or lucky by playing in traffic in certain ways. But you've really combined the two and created an advantage that I believe will be very differentiating, as long as you execute. Execution is key. Without execution, all this stuff doesn't matter. You have to execute, and you have to scale. Most people like me are watching: can you scale? We're ready to date, but are you really going to be able to handle us plus all your other new friends? Can this thing scale in a way? Bill faced that at Pivotal, and Pivotal was a great company as well. These are the things that will differentiate you, that make people say it was worth another database rather than 'Why did I do that?'

Yeah, look, we started on the digital front. In our fintech environment, we serve banks with not only the core bank processing, the core platforms that handle banking, but their teller systems, their ATM systems, their card issuance, their mobile and online banking, their business banking, the commercial banking. Every one of those areas, our clients have said to us, if they're in the lower end of the scale, they want a completely packaged solution. If they're at the higher end of the asset stack, they want a best-of-breed solution to integrate with our core. So the way you're relevant is they want simple APIs to be able to do their processing. In this case, transaction processing with another database, it could be more AI or intelligence or trending or forecasting or things like that. They want high speed, always on, and not quite global but certainly regional solutions to problems. But they want to be completely transparent. They want 24x7, not just always up, always available, but they want 24-hour processing. Banking is moving to real time just as trading did. So they need solutions that really allow for truly real-time processing. That means I process a transaction, and it's now reflected in multiple states. It's an active-active design, and I can always know the state of either my bank, my settlement, my clearing, my lending. So for banking, it's really around the movement to digital and all the surrounds that we're building. It's about more and more real-time processing and continuous availability and continuous knowledge of state. And then in payments, it's really just like, I mean, it's every country. Cash is in a downward trend, non-cash is an upward trend. E-commerce and mobile commerce are off the charts. For those types of client experiences and journeys and the types of transactions that we're doing, they just require not only a lot more sophistication but they require us to be in the cloud because we have to be wherever the clients want to meet us. They want to meet us at my private club, that's great. If they want to meet us in the public cloud, that's great. You can do both. But the most important thing is they want to meet us in these public domains to do all kinds of activity and processing, and I don't want to have to move the data to so many different places that it becomes unwieldy. We want to create this sandbox, this container that really expresses the client value and real-time information to their clients. They want to do real-time marketing in line with the payment that's going to happen in 20 milliseconds. That can't take three seconds or the next day. It's got to be real. We want to draw people into a variety of different experiences that are very different than today. That's what people want: they want what they want, when they want it, wherever they are, in the simplest form possible. All of that plays to what you've created here. I just hope that we can figure out how to use it. But I also, and you don't have to answer it today, but the question again for your listeners, this is what I would do if we were in a room. I'd say, 'Bill and Karthik, pretend you're running Fiserv, and you have at your disposal, how would you use it? Where would you put it? How would you maximize value?' That's the way that I would love your employees to think, because I'll take those ideas every day of the week. I bet every one of your employees interacts with a bank. I bet every single one of them interacts either in a branch, at an ATM, at some other form of kiosk, has a credit card, has a debit card, uses a mobile platform. How would you use Yugabyte as a consumer interacting with one of your banks? That's us, except for the top six or seven banks. We got most of them in the country, or at least half of them. And in payments, we're doing the world's payments. How would you change the interactions that you have? We've seen things go to contactless. We've seen different types of payment methodologies. We've seen fraud go off the chart again because as you get into these more digital type purchasing patterns, knowing who the person is behind it, even if it's not a person, is critically important. So how do you either stop fraud or increase authorization optimization? Those are the things we have to deal with. We have to render probably not terabytes, but we certainly have to render gigabytes of information about a person making a purchasing decision in less than 100 milliseconds and be perfect every time. Because if we decline it and it's wrong, the client doesn't like us. If we approve it and it's wrong, the client doesn't like us because there's fraud that goes with it. So we get paid to actually be perfect: no false positives, all good transactions. The closer we get to 100% accuracy is key. How am I going to do that honestly without you? How am I going to do that? I could if I built my own database. I'd love to do it again. I did it once or twice already. That's not the answer. I'd rather really smart people like you do that instead of me. It's hard to do. Does that make sense? I'm assuming it's relevant.

I think very well said. It took me back because you were talking a bit. I remember, you won't remember these names, but Bob Epstein, Mark Christensen, 2001. I was building with them the first Sybase replication server. I had written papers on it, had built it, understood these problems early. These have been, and when you were building trading systems, that synchronous piece was important, but asynchronous was the most likely way to keep state in multiple geographies: New York, London, Tokyo, Hong Kong. So I think these are things that you have perfected. Bill asked the question earlier, I gave you some examples about where I would use it. But when I think about you, it's a cliche that I've used before: cloud, social, mobile, data, security. When I think about where could I use you, multiple places. In cloud type applications, mobile for sure. Social, anything that is social in nature lends itself very naturally to here. Security, we talked about fraud, but there's also security aspects here that go at anything from authentication, authorization capabilities that have to be dynamic and nested. On the data front, it's where you are. So when I think about you, I think cloud, social, mobile, data, security are really the opportunities. What fits underneath that, that's how I visualize when I close my eyes what you could do. We all have a little bit of evangelizing to do. You have a lot more evangelizing to do as a company, which is important. But so do we as users of the technology, because I succeed if you succeed. If I'm your only customer, you're not going to survive. So everybody has to grow and scale and go from there. I like the Tesla model. I like the idea that somebody is in a continuous upgrade cycle, as long as again you execute well. It's a very good model, and it stops all this bad legacy. There's a lot of good legacy, but it stops all this bad legacy from growing out of the ground and never going away. By the way, your comment about Sybase, we've had a few recent requests to help move some of that stuff. This time it's 20 years now, so it's time.

It's probably the number one thing on my mind when we talk about this topic. I think about it a little bit. I think about it the same but different. First of all, remember we have four million small businesses around the world that we do business with, and then we have three to four thousand banks. And in between, pick your enterprise brand: Home Depot, Best Buy, Target, Walmart. Those are all our customers. So we have this tiering that we have to think about. But at the same time, it's all similar consumers. Consumers work with Walmart, consumers go to the deli, consumers go to the bank. So we are thinking about our product stack on a regular basis now around how to segment. And then this issue of single tenancy versus multi-tenancy is probably the biggest struggle at the moment. Because we need leverage in the things that we do, and that often comes from multi-tenancy. But then you also want segmentation and independence and the ability to turn dials around performance and costs and things like that, which drives you to less multi-tenancy, if not single tenancy, for large clients. I haven't had enough experience with Yugabyte yet to know what all the capabilities look like in real practice, how to turn that dial effectively. But I would say that's a concern of mine, not about you, but about our future architecture and how we deal with it. Obviously, from every customer's perspective, we want the perception of single tenancy. We want everybody to think they got their own sandbox, their own environment perfectly for them. So I think that's probably a key concern. And then segmentation, I don't know if I have any great advice yet because I don't really know how the first answer plays out to be able to give you the second piece of it. But I do believe that there are ways to optimize workloads, and workloads attach themselves to more customer journeys than customer types. Buying a house and lending is very different than buying a soda. I could go to trading examples as well. Trading a basket of stocks is very different than trading a single name. So a lot of this is really you teaching us the capabilities and the characteristics, us playing back to you what works, what doesn't work, and fine-tuning. But self-healing is important, self-scaling is important, auto-configuration is important. Because the goal here, I think, is also, I tell my infrastructure teams, 'You're never going to have more people than you have today. It's only going to be less.' So customization better be really thoughtful, or the places where we do customization, we better be very thoughtful. And automation also be very thoughtful on how many different buttons you choose to have. You ever sit in the cockpit of a plane? A lot of buttons there. You don't want to touch 95% of them. I think it's the same here. Knowing that you could, if you have a problem at your disposal, is great, but you don't really want to touch 95% of those. So the auto-configuration with a few more dials for me to turn, depending on whether small business, large business, high transaction, low transaction, structured, unstructured data, things like that, are really the way I think about the answer so far. But my bet is I'll have a better answer for you six months from today.

There's probably two categories at least. One is maybe one is EQ and IQ. So the first is how easy is the company to work with. I think to some degree you want it to be somewhat automated, but on the other side, you're going to have to teach a generation of people how to use your product effectively, interact with it, and advance it. So the first piece of advice I would give you, and one of the key measurements I think will come out of it, is I believe half of your success will be dictated by how well you listen. Most smart people don't listen. So your company at best right now is good, at best, because you haven't had enough experiences to make it great yet. You will. The more you listen, the more you'll make it great. So I think the first metric, which I know you would have never expected me to say, is how well you listen and how well you partner. Partner means teach and bring along. The biggest problem I have right now is people in my own company say, 'Why do I need another database?' They do. The smart ones I don't have to give the answer to; they're just teasing me. The ones that haven't seen the coffee yet, those we have to bring along and teach and help them understand. You'll do that, and I'll do that, and we'll do that, and we'll get there. But I think listening is really key, and how you partner. Other than that, I would say the three main metrics to me are really going to be around performance. Can we, our workloads are only increasing, so can we get performance? Performance might not be the traditional TPC benchmarks. Performance is really around the single tenancy and multi. How much multi-tenancy can I get and still get the output that I desire? Aside from performance, availability is key. The third one is true state. State means can I really have real-time state, or certainly near real-time state, in multiple locations so that I could truly provide either the multi-tenancy to a broader scale that I'm talking about, or true independence for regions and things like that to get past some of the privacy, onshore processing, and things like that. So I think those are key. And then the last one is really just around automatic. I'll call it automatic: watching the auto features, auto-scaling, auto, and self-healing things like that. Making sure I don't want to touch this thing. I just really want it to be like the best toy I've ever owned in my life, and it just does what it needs to do every single day, low maintenance and high value. Those are some of the things, and they obviously all turn into a certain set of metrics. But at a high level, I call it EQ and IQ. Some of it is about the company and the personality of the company and how you bring the product to customers in the market. The others are true metrics to be sure that we can get the workloads done. These APIs are important too. Karthik, your points around what you've done here, how you've structured them, that'll be the test of time. Can they really hold true to getting everything we need in and out of this environment? So those are special metrics that are on maybe the third category, but those first two are pretty important.

I don't think so. I'm just proud of you guys and the company. I'm proud to be up in this partnership. I'm excited about what the opportunities are for both of us. I couldn't thank you more for your flexibility and patience to date. Hopefully nothing but good comes out of it for both of us. Appreciate it guys. Thank you so much for your time.

Okay, see you later.