Fred Sadaghiani CTO of Sift Science

(Joel Beasley at 00:00:01) Hello, my technology leader friends. A few updates before the show. First, I have a book recommendation for you today. It's called The E-Myth Revisited. We talk a lot about repeatable processes and teams and business on the show.

(Joel Beasley at 00:00:17) This book opened my eyes to an entirely new take on the process, and I recommend it to everyone, but I consider it required reading for any co-founder or anyone with an entrepreneurial desire. For those of you following the Leaderbits journey, thank you. We are growing every single day. The latest feature that we've been working on is our same-day help platform.

(Joel Beasley at 00:00:40) So a big frustration of mine when looking for leadership advice, scanning forums, being a part of a Slack, or reading blog posts is that it's information overload. You have no context to who's giving the advice, their qualifications, or their personality type, and your specific situation is always different. So to complement the weekly challenges where you practice the skills that grow you as a technology leader, we're opening up a feature to the members of Leaderbits where you can get same-day help and real advice from me and my team about your specific situation. This is available to all the leadership members, from the individual contributors all the way up to the CTOs and everything in between. At Leaderbits, we specialize in the transition from individual contributors to team leads, to leader of leaders, and then finally the C-level.

(Joel Beasley at 00:01:31) If you're at any of these stages, or you have teams or individuals struggling through these transitions, then visit leaderbits.io. I'm so excited about this episode. I got to go out after we recorded this and actually meet him in San Francisco. He is an absolutely amazing human being. Today we are talking to Fred, the CTO of Sift Science.

(Joel Beasley at 00:01:54) And we discuss scaling a business through the engineering lens, advice for companies going through an acquisition, and why embracing failure can be the key to success. All of this right here, right now on the Modern CTO podcast.

(Fred at 00:02:11) Here we go.

(Joel Beasley at 00:02:12) This is the Modern CTO podcast. Slack actually happens to be our most recurring special guest.

(Fred at 00:02:28) Yeah. Uninvited special guest, I'm guessing.

(Joel Beasley at 00:02:33) So where are you at today?

(Fred at 00:02:35) I am at our office in downtown San Francisco.

(Joel Beasley at 00:02:40) Downtown San Francisco. I've been out there quite a bit. I was just out there for RSA a few weeks ago.

(Fred at 00:02:45) Oh yeah, we went to RSA as well. The city is kind of overtaken by a lot of different people coming in and trying to learn about security.

(Joel Beasley at 00:02:53) Yeah. I was talking to the Uber driver and he said that you should be here when Salesforce comes, because there's so many humans. There's like apparently 120,000 instead of 60,000. So many humans that the network slows down.

(Fred at 00:03:07) Yeah. Dreamforce is kind of the conference to be reckoned with. The city literally transforms.

(Joel Beasley at 00:03:15) Really?

(Fred at 00:03:16) Boris takes over. Yeah. You see a whole bunch of new billboard ads go up. There are people in the streets just kind of organized around different areas of Moscone Center, and it's just—the city's overtaken. It's kind of crazy.

(Joel Beasley at 00:03:29) How long have you been in San Francisco?

(Fred at 00:03:31) I've been here about eight years. So I used to live in Seattle, worked at a number of companies and startups up there, and my last startup was acquired by Google. And so as part of the acquisition terms, we were made to relocate down to Mountain View. And I hung out at Google for a little over two years, splitting my time between Mountain View, San Francisco, and New York.

(Fred at 00:03:57) And then finally moved over to Sift Science, and I've been at Sift six and a half years.

(Joel Beasley at 00:04:03) I love the website, by the way. It's like some of the most beautiful design I've seen.

(Fred at 00:04:08) Yeah, thank you very much. I'll forward that along to our designers. We have this vision of kind of, you know, marrying design and technology together in a way that really makes it easy to understand what value we provide. I think a lot of people don't focus enough on design and the storytelling behind the technology and the products they build.

(Joel Beasley at 00:04:29) Yeah. Well, as humans, our memory is based on stories. And so if you're able to tell a good story, like Simon Sinek, right, you're able to get stuck in someone's memory. And so you've been around for six years, and a big part of your business is machine learning, right?

(Fred at 00:04:46) Absolutely. Yeah.

(Joel Beasley at 00:04:48) So you guys were before it was getting hot. So you guys started before everyone was on the AI/ML money train.

(Fred at 00:04:56) Yeah, absolutely. There is a train. I mean, one of the funny things here is that machine learning has been around for many decades. The algorithms in particular that power machine learning. But it's only been about the last six or seven years where we've seen this really widespread usage of the algorithms across various different tasks. And so yeah, we were right there at the beginning where it started to take off and take shape. There were, you know, I'd say maybe a half dozen popular libraries like scikit-learn and the like that provided implementations of the algorithms that you could use, but nothing at the scale that we see today.

(Joel Beasley at 00:05:39) Yeah, it's amazing to see the different projects that emerge. When it started to become real popular and I started hearing about it a lot more, I did some research into it. And I was like, oh man, I saw this stuff, you know, back in the early 2000s when I was just searching around and tinkering. And I started reading some of the reports, and I'm like reading papers from like 1989 and stuff on machine learning white papers and just these really basic concepts and individuals getting involved. So what exactly does Sift do?

(Fred at 00:06:11) So we're in the business of protecting merchants and internet users from different kinds of bad actors, and that spans a kind of spectrum of fraud and abuse. So when we started out in the early days, the problem that we solved was to protect merchants from what we call payment abuse. This is chargebacks and people using other kinds of stolen credentials and cards to purchase things online. From there, we evolved to capture different types of fraud and abuse. We learned that, you know, bad actors, fraudsters—they're very, very motivated. Where there's money, they will go. And secondly, they're really advanced. You know, they're sophisticated. They're starting to deploy more and more sophisticated means of subverting systems. And wherever there's kind of an opportunity for them to get value, they will go.

(Fred at 00:07:05) And so we expanded past the payments abuse and started getting into content abuse, account takeover, account-type fraud like fake accounts, promotion abuse, and so on. And so we've grown over the years from solving that one task of payments abuse into building more of a platform approach to protecting merchants from different kinds of bad actors.

(Joel Beasley at 00:07:31) Yeah. And another area where you did a fantastic job is the content, the educational content. So I was consuming all of it and learning so much about the things you were mentioning, about account takeovers. And I, as an engineer, you know, for fifteen, seventeen years, I'm like, oh yeah, I could build that. Like I could build the attacking thing. And I was like, you would need defense against that. Like I could build that too. Like, oh, I could see how I could do that. Yeah. I was like, oh, you know what? The whole world is coming online.

(Joel Beasley at 00:08:01) Like, you know, we've lived through it here, so it's very commonplace to us. But to realize that there are cities in different parts of the world that are actually getting internet for the first time and learning all the things that you can possibly learn and knowing that, oh, if I go attack this cluster and perform this attack, I can actually make money somehow. And so they're seeing it as like, you know, maybe even in a Robin Hood-esque way of like, yeah, let's attack sources that are attackable. So that whole—I guess you call them bad actors—the whole concept was made aware to me recently. Like I just didn't have perspective. I've been so buried in business logic and enterprise systems that this is something that's really interesting me. So I'm glad—thank you for being like the superhero.

(Fred at 00:08:48) Oh yeah, definitely. I mean, we go to great lengths to try to educate the world about the different types of fraud. I think just to kind of touch on what you're saying, I think definitely there are examples of where there are the lone fraudster who's attacking, you know, some e-commerce site and trying to use stolen credentials that he or she bought on the dark web or whatever channel they got these stolen credentials from. But there's a whole different layer that people are completely oblivious to, and that's organized crime.

(Joel Beasley at 00:09:20) Yep.

(Fred at 00:09:20) So organized crime—I think when, you know, if I were to tell this story to my brother or to a friend, they would think of Italian mobsters, you know, trying to get their taxes, collect their taxes in the real estate, the property that they owned across, you know, the neighborhoods—like Godfather days. But the nature of organized crime is very, very different today. At least there exists a presence of organized crime on the internet where there are actual businesses that run profit and loss centers, where they hire people like you and me who can code, specifically with the task of stealing money, funds, value, reputation, identities. And then they have salespeople who carry quota and have to sell the stolen identities, the stolen cards, the stolen value to other fraudsters. This is real business.

(Fred at 00:10:15) And it's kind of shocking to, you know, learn about this. But there's a lot of money stolen every single day from across the world, and it happens in the U.S. too. I mean, it's not just these new geographies that are kind of coming online where, in the example you gave, I can now leech funds or steal funds. It happens in the U.S. every single day. So it's kind of shocking, I think, when people really learn the true cost of fraud, the real size and scale of the problem.

(Joel Beasley at 00:10:48) Oh man. I have a couple friends that are in the Washington area, and they were telling me about how there's a big push right now for offensive systems. And so now this whole concept of like conflict is—we, our countries are now offensively—it's like war, but like digital, but like not so obvious. It's like a very slow, strange situation. Have you looked into that at all? About how we work as countries with that?

(Fred at 00:11:23) No, not at that level. But I could imagine that being the case. I mean, the nature of attacks and abuse are evolving and changing very, very quickly. And I think like the mental model that, you know, most of the world, I would say, has as far as what constitutes fraud, what constitutes abuse, is really changing. I mean, our recent kind of election and this notion of fake news and how, you know, Facebook played a part in allowing adversaries to use advertising to convince people of one opinion or not. I mean, I'm not trying to go on one side of this issue or another, but just make clear that we're now learning that there are many tools—and they're mostly technology-based, data and information-based—that are being used to attack us. And we're just not ready. There's a lot of lack of preparedness, not just in the tools and the technologies that exist there, but in our own mental models of how can we be attacked. You know, if I was standing in front of you and I was going to attack you, you would kind of see it. You know, you would understand it.

(Fred at 00:12:31) But if I was about to attack your computer system right now, you probably wouldn't have a good sense for how that could come about, what the pain and the downstream problems from that could be. And what we found with Sift is that most merchants, as their businesses grow, they don't actually understand—or they misunderstand—how the nature of fraud and abuse on their sites can really lead to the downfall of their businesses, be it through the chargebacks and the costs that they have to pay in penalties to their downstream providers, or through reputation and identity abuse that, you know, you can go and take a site and just abuse it with inappropriate or unsavory things that just destroy the reputation. And that trust is the very fabric, is the very fiber of why people are using many of these products, and that's at risk.

(Joel Beasley at 00:13:28) Yeah. And the—well, let's just talk about the unrealistic ability for a human to approve every piece of content. It just doesn't exist. It doesn't work with any business model, period. Right?

(Fred at 00:13:41) Doesn't scale.

(Joel Beasley at 00:13:41) Yeah, it doesn't scale at all. And so, yeah, the need—and you're right, the trust, man. I became a fan reading through your site, like, because it's so true. Like, we all have these concepts, and we don't speak about it much, but we all have these ceilings of brands and there's an unspoken trust associated with them. I was speaking about it first with humans. So I had this one talk I gave that was discussing you have different people in your life and you literally trust them and apply ratings to them based on the content that they bring into your life. So I know if I get information from one person and they're really involved with this one area, that there's like a high trust factor.

(Joel Beasley at 00:14:25) Historically, they have good information. They have no ulterior motive. They're a stable person. And then other people I know, if I hear something, the problem is like nope. But we also have that same relationship with brands. It's just not talked about a lot.

(Fred at 00:14:39) It's not talked about. It's not really, you know—I think to some extent, it's not really understood. There are elements of that where there are like external forces, peer pressure where you think it's a good thing. Like Facebook, I think people just accept it as being a good platform. I don't mean to pick on it, but now that we've learned how Facebook uses the data, now that we've learned what that might entail from a kind of experiential perspective—your experience might be totally different from mine—we're becoming more and more kind of aware and keen to this idea that this data is really, really powerful.

(Joel Beasley at 00:15:15) So you're the CTO, correct?

(Fred at 00:15:18) Correct.

(Joel Beasley at 00:15:19) And you guys have scaled. How many people do you have in engineering currently?

(Fred at 00:15:23) We have about fifty-ish engineers. Yeah.

(Joel Beasley at 00:15:27) And are you a co-founder of Sift?

(Fred at 00:15:29) I'm not formally a co-founder, no. But I was part of the founding team.

(Joel Beasley at 00:15:33) Okay. So you are part of the founding team. And so you've grown the engineering from the beginning to where it's at today?

(Fred at 00:15:39) Yeah.

(Joel Beasley at 00:15:40) That's amazing. Congratulations on that. I mean, 50 people. That's fantastic.

(Fred at 00:15:46) Yeah, it's been a really great experience. So we're about seven years old. When we started the company, it was four of us sitting around a conference room in a poorly ventilated old—not abandoned building—it actually wasn't older abandoned, but it kind of felt that way. And yeah, where we are today is we have two offices. So we do some R&D out of our Seattle office. We're about 25 people there. And the other half is about here in San Francisco.

(Fred at 00:16:17) And we've scaled the company from, you know, just a couple folks to now, I think, about 150 people across sales, across marketing, and biz ops and product and engineering and so on. And so, yeah, the company has grown tremendously. We're a real company now. We're really solving a real problem, so it's kind of exciting.

(Joel Beasley at 00:16:43) So I found that one of the best things in life is to have a meaningful contribution. Like after you get over the financial hurdle, right, and you meet the Maslow hierarchy of needs to where you can live comfortably, now it's like the reward comes from not climbing a ladder, but being able to bring more value to more people.

(Fred at 00:17:01) Yeah, I totally agree with that. I mean, my kind of view on this is it fits into this framework of hard, valuable, and fun. This is something that one of our early investors, Max Levchin, really taught us. You should really spend your time working on something that's very, very hard, something that's very, very valuable, and you have to have fun doing it. And the way I describe this is, look, we're working in San Francisco and Seattle.

(Fred at 00:17:33) These are two huge tech markets. For any engineer, it'd be super easy to walk out the door. You could literally hold a piece of paper on which you used your Sharpie to say, like, hey, I can code, give me a job.

(Fred at 00:17:45) And somebody would come and talk to you, and you would get a job working at any of the big five companies, assuming you could pass the interview. So there's no shortage of opportunities. What there is a shortage of is really hard problems that are fun and valuable. And so really, I think that what I have prioritized my career on is working on things that are very, very hard.

(Fred at 00:18:10) That's where I've learned the most and forced myself out of my comfort area where, oh, I know how to solve that problem, and that path is clear, to the place of, hey, I have no idea how we're gonna build this business. I have no idea how we're gonna get all the data and deploy all these models and mix them in ensembles and maximize AUCRC for these curves. And then the valuable side is, like, look, we could easily build simple apps that solve a small problem, but we want to change the world. We want to help elevate the conversation and the understanding in the world around how fraud and abuse are really disrupting us in negative ways and provide the technology and the product to really solve that in a meaningful way.

(Fred at 00:19:01) And then the last element, the fun, is, look, you're gonna be doing this for a long time. If you can't go to the office and smile and have fun with your coworkers and be proud of the accomplishments, then you're totally doing the wrong thing, and you need to reevaluate where you are. So it's really those three things that have been the cornerstones of how I think about it and what it takes to really be at a place for an enduring amount of time.

(Joel Beasley at 00:19:32) So I'm gonna ask some value questions for the CTOs. We have a large CTO audience, lead engineer audience. And so you've been working on this fun, hard, valuable problem, right? And you've scaled the team up to 50.

(Joel Beasley at 00:19:46) What comes to mind is, like, your areas of growth while you've been growing the team?

(Fred at 00:19:52) So I think from the CTO lens looking at that, I'll just give you a very reductionist kind of quick overview.

(Joel Beasley at 00:20:00) Yeah.

(Fred at 00:20:00) You know, in the early days, I spent the lion's share of my time coding. I was really focused with the rest of the team on building the product and just finding the value, finding the product-market fit. And that involved a lot of hackery, a lot of code that we're not proud of, and code that doesn't live anymore, but it got us to the point where we could get close to that zero-to-one threshold, you know, like Peter Thiel.

(Joel Beasley at 00:20:26) Yeah, I would click right there. Yeah.

(Fred at 00:20:28) Yeah. Yeah. And so that was where the lion's share of my time was spent. It's just trying to understand what is the product, what is the problem, how can we fit these things together. The next kind of coarse-grain step was making those systems operate and run.

(Fred at 00:20:46) Right? So our very initial service was Mongo. It was based on Ruby and Rails, and that didn't really scale. I mean, we could get into the reasons why, but you know, as we started taking in the traffic of tens of thousands of websites, our database and our API server was not able to keep up with the scale. And we moved to a more kind of robust infrastructure.

(Fred at 00:21:11) We wrote our stack in Java. We now run on HBase. We now have literally many thousands of machines running in our infrastructure. And so I spent a lot of my time moving from defining the product to working on, okay, how can we operate and scale this thing so that it keeps up with new traffic, both as we acquire new customers and our existing customers themselves grow. We see more traffic, and so we need to invest there.

(Fred at 00:21:39) And then the next coarse-grain kind of evolution was, you know, now rather than work on that infrastructure or working on that product in the sense of coding every day, now get to the stage of figuring out how do we invest, what people do we need to bring on, what specialists do we need, and you start thinking about the people more. And then the next coarse-grain kind of evolution is like, okay, now you're not thinking about people so much, but you're thinking about teams because you have to have a team that's dedicated to operating our infrastructure. Another team is responsible for offline evals and experimentation and the data science, and another team is responsible for the front end and the usability, and another team is responsible for the design. So you just go up a level higher.

(Fred at 00:22:30) And then, again, being reductionist and very coarse, you know, the next level of this is, okay, now how can we take these investments and deploy them against bigger market opportunities? How can we further grow and solve meatier and bigger problems? And I'm not sure if that's answering your question.

(Joel Beasley at 00:22:49) Oh, yeah, that's like it's going in my book, man. No, that's good.

(Fred at 00:22:54) Give me credit in the footnotes.

(Joel Beasley at 00:22:56) 110%. No, that's good. I'm sitting over here, like, cheering, like, yeah. I know where he's going. He's going to people. I'm like, I know he's going to teams because you got like the turducken, the series of repeatable processes. They start stacking. You gotta figure out teams of teams. That's like the next stage.

(Joel Beasley at 00:23:11) I'm like, I've been there. And then I'm like, where, but then you just came out of nowhere with, like, investments to deploy against bigger problems. I'm like, yes, haven't heard that. I'm a fan. I'm a fan of Fred.

(Fred at 00:23:23) Likewise. Thank you. Yeah. I mean, it's the role of the CTO, I think, is one of the more dynamic in an organization because you really change what you're working on. I would say quarter by quarter, you just have to be very, very mindful and aware of how your time and your investments need to change over the quarters.

(Fred at 00:23:42) I mean, it really depends on what the growth of the company is and, you know, how much traction and how much progress the business is making. But if you're growing at a healthy clip, then you really need to be reevaluating. I mean, just to be clear with you, the engineers are very unhappy when I write code. Even though they double scrutinize what I do, and I totally love them for it, right? Because they're the ones that are out at the forefront of what is the code that's operating our production systems, and they internalize it way better than I do at this stage.

(Joel Beasley at 00:24:17) So here's the interesting reinforcement for you. So in the last year, I've spoken to 100 CTOs. They're probably like 110 now. And everywhere from two people to 100,000 people, right? And the biggest difference is what you just said. The difference between the companies that grow or that kind of stagnate or, you know, stay at a similar level is really hung up on, at many times, it's just the CTO's inability to let go of something they really enjoy doing in order to put their focus on something that they need to be doing in order to get to that next stage.

(Fred at 00:24:55) Yeah. I would underscore that point 100%. You know, this comes back to the way I like to think about problems, the hard, valuable, fun. It is really, really hard to do something that I'm not good at. So I think I'm a really good coder.

(Fred at 00:25:09) I think I'm really good at going from zero to one and building systems from nothing. I am not the best. I am not the most experienced when it comes to scaling organizations. But I know that I'm in a unique position given my experience, given my level, given kind of the broad view of the company, the customers, the market, to be the one to scale us in that way. And so while it is safer and more comfortable to be at home working on, you know, that feature we're trying to launch this quarter that does whatever, it's not the right thing. What the engineering team needs, what the product, and the rest of the business needs is me focusing on how to scale the business through the engineering lens, and that's where I spend my time.

(Joel Beasley at 00:25:56) Have you been getting into the speaking at all?

(Fred at 00:25:59) Yeah. I've been doing more and more of that. I was just in Dublin talking on a panel at MRC. This is Merchant Risk Council. It's kind of a major conference around the space of fraud and abuse and payments.

(Joel Beasley at 00:26:12) Yeah. I've actually saw like, I was just looking, I don't know, a year or two ago online, and this ad caught my eye. It's like, do you like to geek out over, like, six-millisecond fraud transactions? And it was some company doing recruiting, and I was like, maybe. Let's see what this whole world is because I was like, you know, they were saying like six-millisecond fraud detection.

(Joel Beasley at 00:26:35) I'm like, well, that's interesting. I want to see what technology is being deployed there, you know? This is really creative how people are finding different people. One of the, I was talking this morning to the CTO of Asana, and they're out your way, and they just passed, I think, 100 engineers. Are you familiar with Asana?

(Fred at 00:26:54) Oh, yeah. Absolutely. We're all Asana users.

(Joel Beasley at 00:26:57) Oh, really? Yeah. So yeah, I was talking to their CTO, like, an hour, two hours ago, fantastic individual, and we were talking about how they're scaling and what they're doing. And I was sharing with him that some of the CTOs in that larger space that you're entering in right now, you're going from the 50 to 100, right? So the way that he's saying that there's big bottlenecks with recruiting great talent and leadership to scale the organization. Because you start to connect those teams of teams, you need great leaders that have strong technical backgrounds. A lot of people with strong technical backgrounds are a little hesitant to move into true leadership roles, right?

(Joel Beasley at 00:27:34) They don't mind being team managers, but true leadership roles, it's hard to get them to move into them, you know, at scale, right? To move as fast as your company is moving. But the point would be that a way I've seen many companies do this successfully is by getting involved in the community and having their leaders speak in their different sort of micro cultures. Like, for example, you're at the Merchant Risk Council, but maybe

(Joel Beasley at 00:28:01) some of your leaders, if you're primarily in, I believe you said, HBase, right? That's something that you're in. They'll have people in that, you know, culture as well doing speaking and outreach and communication because I'm finding that when the leaders are out, the best people look at the leaders and say, I wanna work for that person. I like their attitude.

(Joel Beasley at 00:28:22) I like their culture. I like how they see things. And I get all the time people that come on the show, and then they'll call me up, like, a month later and say, hey, an engineer heard me on the show, and they reached out to me. And it's been one of the most fantastic hires because they wanted, because they align, the way they think aligns. When you're out there more and people can see who you are as a human, that'll attract, you know, your tribe.

(Fred at 00:28:46) Yeah. I totally agree. I mean, coming back to that statement that I was making is, you know, for engineers today, there's no shortage of opportunities where they can go and work. And I think, you know, as engineering leaders, as CTOs, we have to work really hard to make clear why working at our companies, working on our visions and our businesses are ones that are really exciting and ones where they'll get to learn a lot. I think, you know, if you go down to kind of what motivates a lot of engineers that I've met, is they want to work on hard problems, they want to be acknowledged, they want to learn, and they want to have that opportunity for growth in their career.

(Fred at 00:29:24) I think if you go and work at one of the big five tech companies, the room for growth is kind of mired in, let's say, complexity. It might be politics or bureaucracy, but it's really complex versus the notion of working at a startup where there's really no shortage of hard problems to work on. But the question is, will working on those hard problems get you to where you want to go? And so I think as tech leaders, our job is to kind of provide the clarity in why working at our companies is really exciting, is really rewarding, is really fun and valuable, and how we support the growth of the individual. And, you know, we're really lucky in that Sift has grown as a business.

(Fred at 00:30:06) It's always kind of hard to fulfill that promise if your company or your business is, let's say, not performing well. You can't give your engineers more. But as the company grows, the best thing to do is, you know, farm from within, as they say. Let the people who have been there, who have kind of been through the muck, understood the problems, seen the scaling firsthand and how the technology solutions we built while they work for, you know, 10 QPS to 100 QPS are falling apart at 10,000 QPS and beyond. Let them be the ones that help direct teams at larger scales and say, like, oh, I've seen this problem before.

(Fred at 00:30:43) Here's what we did. Here's how we encountered it. And then coming back to the earlier thing you were talking about as far as, you know, going out there into the industry and talking about it, I agree 100%. I mean, we spend a lot of time at meetups. We've spoken, I think, every year at HBaseCon because we're one of the

(Joel Beasley at 00:31:01) Oh, hey, there we go.

(Fred at 00:31:03) Yeah. We're one of the larger HBase clusters out there, I would say, given the volume that we take on. And then we have our own homegrown events. It's called Turn Up the Bays, where we dive deep into our machine learning infrastructure, how we do evaluations, how we mix the many models, global knowledge, local knowledge, how we've effectively deployed RNNs, what it took to do that. And

(Joel Beasley at 00:31:30) Oh, I'm sorry. What's an RNN? I haven't heard that before.

(Fred at 00:31:32) Recurrent neural network.

(Joel Beasley at 00:31:34) Oh, yes.

(Fred at 00:31:36) Yeah. Sorry. We speak in TLAs at Sift. Yeah.

(Joel Beasley at 00:31:39) Elon Musk would not be happy with you. Yeah. Have you read his letter about acronyms? No. No.

(Joel Beasley at 00:31:47) He is very outspokenly against acronyms because apparently all his engineers when building the rockets, you know, early on because it's rocket science, they would call like thrusters, like these long acronyms and then you'd have to learn all of these acronyms. He's like, no, that's the thing that makes it go faster. Like just call it what it is. Like use plain English. And so he's got this famous email.

(Joel Beasley at 00:32:10) I read it. I read his life story in his book. He's got this famous email that he sent out that's like, no more acronyms. And so when I hear acronyms, that just triggers.

(Fred at 00:32:18) Really hard thing to overcome. I feel like engineers create culture wherever they go. And, you know, as CTOs, it's our responsibility to shape and hone it, but I've never put one second of focus on the TLAs. I always thought that, you know, yeah, acronyms, you know, I feel like everything's an acronym, but it's an interesting tidbit of knowledge about Elon Musk.

(Joel Beasley at 00:32:41) There's your next talk, quantifying the cost of acronyms. There you go. How do you measure that?

(Fred at 00:32:48) Yeah. Yeah.

(Joel Beasley at 00:32:49) Oh, that's amazing. So have you thought about bumping over to the other coast to pick up the talent on that side of the world?

(Fred at 00:32:56) Yeah. Absolutely. I think, you know, the Internet is a powerful thing and you don't quote me on this.

(Joel Beasley at 00:33:03) I'm gonna quote you on that. It's going out in history.

(Fred at 00:33:06) Yeah. The Internet's really powerful. I mean, you don't necessarily need to be in the same office. You guys are somewhere in Boston or on the coast, right?

(Joel Beasley at 00:33:12) Yeah.

(Fred at 00:33:13) We're having this interview with at least seventy milliseconds lag, and we're doing just fine. I think there's a lot of opportunity to tap into the really rich, the really strong and experienced engineers that are around the world. While I worked at Google, it was kind of just the norm. It was kind of expected that when you had meetings, you could very well be meeting with people that were distributed across not only the country, but even the world. And with Sift, that was a really hard learning—learning to operate in a mode where, oh, I have a meeting with someone. Okay. They're in Seattle. Okay. We gotta get onto Zoom to have this call. We need a conference room.

(Fred at 00:33:48) There's a lot of tools and kind of infrastructure you need to make it happen. But once you do, what we found is there's really, really strong talent in Seattle, really highly experienced, motivated, and just passionate people. And we believe that they exist not just in Seattle and, you know, California, but the rest of the world. And our plans are very much to expand globally to tap into that talent, not just engineering, mind you, but other parts of the company as well.

(Joel Beasley at 00:34:20) And it's great to have all that culture together. So yeah, I would like to—if you guys did East Coast stuff, that'd be really—I'd be interested to hear about that. Just when you do, I—well, hold on a second. I'll say it like this. I know you're going to, so when it happens, just ping me and say, hey, we're doing some East Coast stuff. But I'll actually be out there. So I'm meeting with the CTO of T-Mobile, who was on the show last week. Very cool guy. So if you want a really cool person that's out there, his name's Cody, CIO or CTO of T-Mobile.

(Joel Beasley at 00:34:52) And then Kevin Scott is the CTO of all the CTOs at Microsoft.

(Fred at 00:34:57) Yeah. I know Kevin.

(Joel Beasley at 00:34:58) Oh, you know Kevin?

(Fred at 00:34:59) Yep.

(Joel Beasley at 00:35:00) He's on the fourteenth, so he's on in less than two weeks.

(Fred at 00:35:04) Oh, cool.

(Joel Beasley at 00:35:04) Yeah. So I'm going out to your neck of the woods to hang out with Kevin, Cody, the CTO of Asana, and then it would be cool to stop by and see your offices too.

(Fred at 00:35:14) Yeah. Let's make it happen.

(Joel Beasley at 00:35:15) Yeah. I'll be out there for a week, so I'm just gonna bounce around and say hi to everybody for a day because it's so cool to—I get to hang out with you and find people I really like through talking, and then it's just cool to see the offices. Everyone's so different. Then what I do is I go around and collect information from everybody, and then I just share it and help them out.

(Fred at 00:35:31) Yeah. That sounds great.

(Joel Beasley at 00:35:33) Alright. So what is on your mind today that you're really excited about that you're working on as a CTO? What's your focus right now? When you woke up this morning, other than being excited to hang out with me, what big initiative are you working on?

(Fred at 00:35:53) Yeah. So there are a couple. I think, from my investment of time perspective, there are things I work on for the business, things that I work on for the technology and product. From the business perspective, we just raised a recent round of financing. You can go look up the details, but it was a fair amount of cash.

(Fred at 00:36:14) And some of the things that we're working on right now are trying to formalize how we want to invest that. It kind of goes without saying, but when investors give you money, they're not looking for you to return a modest amount. We're not savings accounts or, you know, CDs.

(Joel Beasley at 00:36:28) Right.

(Fred at 00:36:29) We have to deploy that capital aggressively and growing. And so from the business perspective, you know, as Sift kind of evolves its platform of different fraud and abuse products, we're trying to figure out how to invest that. Secondly, I've been spending a fair amount of my time actually doing outbound work with sales team. So evangelizing how our product works, how our technology works, and kind of pulling the curtain back a little bit with some of the enterprise customers that we work with, sitting in with their data science teams, their engineering teams, and just taking the direct questions like, hey, you say you do this on your website. How does that actually work?

(Fred at 00:37:05) So we'll do a whiteboard. We'll draw it, and we'll just kind of get into the details. And, you know, candidly, that's something we're learning is really, really valuable for the enterprise class of merchant that's still kind of unsure what is machine learning, how could it fit into my business.

(Fred at 00:37:25) And so one of the things I'm working on is trying to make that more clear and more kind of easy to accept because while you and I could talk about it and say, hey, these algorithms have been around for dozens of years and, you know, you can go and download any number of examples that will be a perfect cat classifier. Enterprises don't operate in that same way, and they need to have a deeper, more clear understanding. And so that's where my investment on the business side has been. On the technology side, one of the roles that I play is kind of an experimenter, a tinkerer.

(Fred at 00:38:00) And so there's a very, very small team of us. It's just two of us. And we are experimenting with what I like to call, you know, wild-ass crazy ideas. We have no idea if these things are gonna manifest as anything meaningful.

(Joel Beasley at 00:38:13) It's like your labs.

(Fred at 00:38:14) Yeah. It's kind of our labs. We try not to put too much expectation on it because, you know, if you set the bar low, anything you do, you can kind of do well.

(Joel Beasley at 00:38:24) That's how I do my marriage, my friend.

(Fred at 00:38:27) Don't tell my wife. The—but the goal here is to kind of do things that, you know, we wouldn't put a first-class investment of a team or, you know, people who are already working against a clear area of ownership. These are net new kind of experimental investments. And as those investments kind of become more clear, then we start putting a deeper and more meaningful investment behind them. So those are very coarse-grain kind of the areas where I'm working on.

(Joel Beasley at 00:38:53) So that ability to not interrupt your processes that are executing in order to try something else is—you've got it. It's working. It's going. That's like, that's my best person in this area, but I gotta try this. It's like you can't rip them out of their routine. It doesn't work.

(Fred at 00:39:11) No. I agree with that 100%. I think engineers in particular, because it's this balance of creative and very highly intellectual work. So when you break that cycle, there's this activation energy that it takes for them to reestablish their mindset and their thinking. And so when you break it, it's actually not just the cost of the disruption and the amount of time they're not working, but also the amount of time it takes for them to reengage into that, into the zone, as it were.

(Fred at 00:39:40) Right? People doing more transactional kinds of things, like, oh, I can stop and go do this. Okay. Now go back to that. The activation cost's a lot less.

(Joel Beasley at 00:39:49) Yeah. So I've noticed this too. So what I found that I was doing is I'm programming a system, and then the individual has quantified how they're valuable by their output. So they say, alright, I do this and I'm valuable. And so if you requantify that output, there's a lag and there's a whole cost you're talking about associated with them figuring out, alright, what am I doing now and then how do I go back to bringing you the value that I was before? And that's a process that I see that I noticed. I'll just be upfront about it. I was doing it, and then I noticed I had to stop doing it because it causes issues.

(Fred at 00:40:28) Yeah. Absolutely. It's very expensive context switching. And while those teams are working, I just want them to hit their goals and have the space and room to deliver against what they're doing. I don't wanna disrupt them.

(Joel Beasley at 00:40:39) So I've got a couple quick questions for you that have come in. And the first one is struggles during the acquisition or anything that just comes to mind about your experience. You said you've had a couple acquisitions in your career. And if you were mentoring a CTO that was maybe a half or a third of your size and they were going through an acquisition, what just comes to mind as what you would tell them?

(Fred at 00:41:07) That's an interesting question. I mean, the acquisition—Google acquired us seven or eight years ago, and that was the most opaque process. So I'll just give you a little more context. At the time of the acquisition, actually, Facebook was at the table and Google was at the table, and, ultimately, it turned out that Google acquired us. And just for a little more context, you know, Jambool was in the business of doing payments for games. One of the reasons that we were really attractive was we had this native Flash widget, and so you didn't have to bust out of the game and hit an iframe to do your transaction. You stayed native in there. And Facebook was coming with Facebook Credits, and they were gonna kind of destroy the world of payments and virtual goods. And then Google was trying to create Google Me or what became Google Plus, and they were gonna have a games thing. So that's how the whole conversation got started. And because, you know, I think neither Google nor Facebook was really clear about what their vision was, there was this opacity to the whole process that just made it very, very strange.

(Fred at 00:42:11) And so, you know, for us, we were brought in to do interviews. It was as if we were standalone engineers going through the process ourselves, and then they did a team-based code review. So we had to show our lines of code, and, you know, an architect from Facebook or an architect from Google would point and say, like, why did you create this abstraction? How do you deal with logging? Where do your logs go? What is the lifetime of the logs? How are they encrypted? How are the keys managed? How do you rotate the keys? And so on and so forth.

(Fred at 00:42:34) And it was just kind of like probing conversation. And the only thing I remember was this was so opaque. They did not want to share any reasoning or justification for anything that they were doing, and we were just at their mercy. So if they would say, hey, now we need to go and, we need credentials to your AWS account so we can kind of poke around and see how you set up your infrastructure, we just kind of said, okay. We can kind of give you this access and you can view this. But we had no control, and I don't know if this is a learning more so as an FYI. Be prepared for that.

(Fred at 00:43:05) These big companies, they get to call the shots, and I think the way acquisitions have gone over the last five or six years, it's kind of more in their favor. They get the decision. Right? You really don't. The other thing is you can't lose sight of what you're working on and the value you need to bring to your customers. It's probably the case that your acquisition is gonna not go through. That's more likely than not gonna be the outcome. It's not to be, you know, disparaging about the effort you put forward or the product or the business or the people, just the way things tend to go. So you have to really balance and manage the demands that the acquirer is gonna bring to you in the context of what you're still trying to do because it's a very low probability that things are gonna turn out. And we effectively stalled all development. We went into strict maintenance mode for about four months that the acquisition took place.

(Fred at 00:44:13) And so, you know, we would come in and we'd say, oh, you know, the MySQL cluster is still working. We didn't get paged. What do we do today? And the CTO would say, look, we don't wanna disrupt anything because we don't wanna change the conversation we've had with Google as far as what we're building, what that product looks like. So we'd work on really simple things that were not very exciting. And I think if the acquisition hadn't gone through, we'd all feel really demoralized.

(Fred at 00:44:31) But keeping that balance against the vision, against the product, and your customers, and the business is really, really critical.

(Joel Beasley at 00:44:47) So the first time I faced that, it was just—it was mind blowing to me because I was the CTO and it was my first acquisition due diligence. And so I'm sitting there and the CEO is like, hey, Joel. So, you know, give them access to everything. I'm like, but we just started talking to them, like, two weeks ago. I'm like, are there NDAs signed? Like, did we sign stuff with them? Just give them the code. I'm like, but that's our intellectual property, man. Like, that's the only thing that we have. We give them our code, they'll see everything that makes us unique, and they're in the same business as us.

(Joel Beasley at 00:45:19) Like, this guy owns a portfolio of companies of which are our competitors. I'm like, we're essentially just going to open up our safe and say, hey, look at everything. And you happen to be related to our competitors. It is very interesting thing to see how that happens. So you don't imagine that that's even something that's possible, right? And then when you go through it, you're like, this is how it happens? This is not very secure. This is very odd. And then, yeah, I would say that my experience is very similar to you, where the people are talking. I was left out. Like, no one really knows exactly what's happening. You weren't the CTO. Correct? You were, like, head of engineering?

(Fred at 00:46:02) I was one of the principal engineers. Yeah.

(Joel Beasley at 00:46:04) You're a principal engineer. Yeah. So there was like—it was basically the two CEOs kind of going back and forth with information, and they would go talk to their team and come back. And it was very limited, and it was, yeah.

(Fred at 00:46:16) Yeah. Yeah. I think the other thing—you didn't ask this question, but I'll share—is I think a lot of people, at least, you know, five or ten years ago, equated the notion of an acquisition with success. And I would kind of say that that's not success. You know? If you're lucky enough to be acquired by a company that's doing the same thing, trying to solve the same problem, and reinforcing the effort that you've made with deeper resources, deeper investment, then, yeah, it could be a success. But the acquisition by Google, not to be disparaging about Google, was no. They hired our team. We went in, and we were given the mandate of basically rewriting Google Checkout, which was kind of a dead product at the time.

(Fred at 00:47:03) And so we rewrote Google Checkout. We redid their payments, and it's probably the payments thing you use today to buy stuff on Google. And so that wasn't our vision. And I would say for, you know, CTOs, for executives of companies that are looking to be acquired, beware because you may not get what you actually think you're going to get. And try to be clear with the acquirer about what your team is gonna do, what they're gonna get to own.

(Fred at 00:47:32) Don't just think about the price tag that's gonna come along with the acquisition. Think about what you guys will be doing years after that acquisition happens. I don't think we spent very much time on that, so that was a little jarring. We were originally acquired to work on what's called Google Plus today. Back then, it was Google Me. And then at the eleventh hour, they just—oh, no. No. You guys are gonna go rewrite Checkout. And so we didn't really think about that process in a way—or at least from my lens, we didn't think about that process in a way that was more than just the monetary, financial transaction of the acquisition.

(Joel Beasley at 00:48:06) Yeah. And for them, it was probably like an acquihire thing. Right? They're like, this is just a really expensive recruiting cost. We're going to deduct the amount of revenue they make, and then we're actually just picking them up for like three times the recruiting cost.

(Joel Beasley at 00:48:18) So we want that team. We'll pick them up and put them on payments.

(Fred at 00:48:20) I mean, that's what we rebuilt all of Google Payments. So—

(Joel Beasley at 00:48:24) Yeah. Well, that's pretty cool, though. That's an awesome project.

(Fred at 00:48:27) Right? That was a tremendous learning. Yeah.

(Joel Beasley at 00:48:29) One of my first projects in real estate. I saw, by the way, we were in real estate at the same time. You were in 2005 solving problems with Zillow real estate. Yeah. I was speaking at like Inman and things like that if you ever went to that conference.

(Joel Beasley at 00:48:44) But okay. Yeah. There's this big conference in real estate called Inman and it happens once a year in San Francisco and then once a year in New York City. But my product got acquired in real estate and I didn't—I was really excited because it was the first time I'd ever licensed any technology or did anything like that. And I didn't realize that—I didn't think about it.

(Joel Beasley at 00:49:05) I saw money. They're like, oh, we're going to give you money for this. And I'm like, yes. I was like, let's do it. What do we need to do?

(Joel Beasley at 00:49:10) And because I was actually building a product alone, like myself. Right? My parents were in real estate, so I was solving a problem that they had. And then an agent went from one location to another location and then told them that, oh, this kid down, you know, in the East Coast, he built this. And so they flew out to see me and they're like, oh, we're going to buy it up.

(Joel Beasley at 00:49:27) I'm like, oh, this is very cool. But then I realized that there was—it was an acquihire. Like, there was nothing for me to do after that. And then to further it more is they ended up just shutting it down.

(Fred at 00:49:42) Yeah. That's—

(Joel Beasley at 00:49:43) They just wanted to buy it and own it so that someone else didn't own it.

(Fred at 00:49:46) Yeah. It's one of the darker sides of the acquisitions, but yeah.

(Joel Beasley at 00:49:51) But thank God it happened early on in my career because now I know what questions to ask.

(Fred at 00:49:56) Yeah. Absolutely.

(Joel Beasley at 00:49:57) I know how to position it. And I learned that there's this whole world where you don't just have to sell it for a dollar amount, you can get residual income from it and all of this.

(Fred at 00:50:09) You can—the thing is, that's actually a good point too, is you have to remember that this is two sided.

(Joel Beasley at 00:50:14) Yep.

(Fred at 00:50:14) That's the one being acquired. You do have some say. When you come to the point of Google or, you know, in the case of these big companies, you probably have a little less say than in other cases, but you do have a say, and you just have to know what to ask for. And so that's a really important point to bring up.

(Joel Beasley at 00:50:33) Yeah. And then knowing how the business is structured between companies—like you could be considered a portion and investment in a $50 million fund that's planning to flip it around back to a billion dollar fund.

(Fred at 00:50:44) Mhmm.

(Joel Beasley at 00:50:44) Like, you can be in a portfolio deal and you can know where you sit. There's—I did, so after I did a couple of deals, I ended up doing due diligence. I was on the other side of it. Yeah. I ended up doing technology due diligence for private equity firms because I just had made relationships.

(Joel Beasley at 00:51:00) I did like four projects back to back, licensed them all, and then each one there's investors on both sides so you get to know them by going through the process.

(Fred at 00:51:07) That's smart. Yeah.

(Joel Beasley at 00:51:08) Yeah. I'm sure you've made tons of relationships like that. Right? Because ultimately all these abstract things we're talking about, there's humans that represent them. So you get to know them during these processes and then everyone goes from project to project.

(Joel Beasley at 00:51:19) So, yeah. After I did that, I just got—that was right in 2013, 2014, 2015 when things really started to take off in the private equity market. And everyone was just, look at this, look at that, look at this. And that's what started this whole book and everything is all the trends I saw looking and I went from being on a project for nine months, you know, building something with groups of people and things like that, to seeing four projects a week and seeing inside of teams. And I'm like, whoa.

(Joel Beasley at 00:51:50) Well, thank you so much. I've got one last question. Elon Musk, time machine. You go to his house. He's got a time machine. Right?

(Fred at 00:51:58) Okay.

(Joel Beasley at 00:51:59) Actually, you're hanging out with him. He's doing the test run. Like, Boring. You're doing flamethrowers. He happens to have a time machine there at that event.

(Joel Beasley at 00:52:07) You go in. You get to give specific advice to yourself. The most useful advice you could possibly give to yourself, very short, very quick, your past self, what would you tell yourself?

(Fred at 00:52:18) Wow. That's a good question. I thought you were going to ask me a question about Elon Musk. So, yeah, I think the advice I would tell myself is to embrace failure. I've been really fortunate in my career to have worked at companies that have been really formative to my experience and they themselves have grown.

(Fred at 00:52:40) But I have experienced a lot of failure in my life, and I think at first when I was confronted with it, I was kind of dismayed. I was kind of saddened about it. But, you know, life is pretty long and I'm one of these people that's probably going to be working into my eighties. I'll probably be a terrible engineer at that age, but I'm excited to do it and take on new challenges. And so what I would tell people is embrace the failure and try to be clear and specific about what it is that you learned from that failure.

(Fred at 00:53:10) And if you can, write it down. Write it in a Google Doc or whatever it is, and say, like, in 2005, I worked on this project, and we thought it was going to be an acquisition, or I thought I'd have this tremendous product market fit, and I invested this much time and so on, but it failed. And be really clear and objective about it. Try to take the sentimental kind of like, oh, man, I'm such a loser, and this didn't work out, and if only I'd done this. Don't dwell on that side. Just be critical in an objective way about why you failed, and then write it down. I think the act of writing it down has helped me. So I'm kind of cheating by telling you this is something I actually do. This is the advice I would kind of—

(Joel Beasley at 00:53:51) Start earlier.

(Fred at 00:53:52) Yeah. I would just start earlier and be less kind of emotional about failures. You know, there's—I don't know if it's Warren Buffett or John Bogle, but they say something like, you know, there's no room in investing for emotion. Right? You can't be emotional about investing.

(Fred at 00:54:08) I would apply that same adage to failure. You know, you're going to fail a lot, and that's actually really, really valuable. Make sure you extract the value from it, and don't allow yourself to repeat it.

(Joel Beasley at 00:54:21) This is beautiful. Okay. You're going to have like two mentions in my next book. Yeah. So I'm not kidding you.

(Joel Beasley at 00:54:29) Unintentionally, the call has touched 60% of the outline for my second book. So—

(Fred at 00:54:37) Are you guys inspired to do this together?

(Joel Beasley at 00:54:39) Yes. You're going to get the pre-copy to look through it and judge it and give me feedback prior to its release. But the first one did very well, Modern CTO. Then I have all these CTOs I know now, which has just been—I guess I was one of the first people to stand up and wave my hand and talk about it because everyone's coming out of the woodwork to talk to me now. I love it.

(Joel Beasley at 00:55:01) I just love sharing and talking. But everything that I've learned in this past year that I didn't cover in the book, I'm putting it all into a second book called Everything But Code. Right?

(Fred at 00:55:11) Yeah. Yeah. That's a lot. Right? Yeah.

(Joel Beasley at 00:55:14) Yeah. Everything But Code. But it's geared towards, you know, the people in the transition. And there's many levels of transition. And then it's about sharing all of this information so that the next generation of people can do better than we did.

(Joel Beasley at 00:55:28) And that's what we do, even on a DNA-based level. DNA replicates and shares information down the chain, and so I figured it's one of the most human things I can do to contribute is just to share information down the chain.

(Fred at 00:55:39) Yeah. I think that's really great. I'm excited to see that book come out.

(Joel Beasley at 00:55:43) Awesome. Fred, thank you so much, man. We did it. We had a podcast, and it was amazing. You're like a superhero with the way you're saving the planet from the fraud and the destruction.

(Joel Beasley at 00:55:54) And then you were—I know you were born to do this, Fred, because your hair, you have superhero hair.

(Fred at 00:55:59) Oh, I think so.

(Joel Beasley at 00:56:00) You have Clark Kent hair. It's like, of course. Right?

(Fred at 00:56:03) Yeah. I did—that wasn't an intentional thing. I think I'm just overdue for a haircut, but.

(Joel Beasley at 00:56:11) Oh, man. Awesome. Thank you so much. I'll be out there in a couple weeks, and I'll say hello.

(Fred at 00:56:16) Okay. Perfect. Yeah. All right. Thanks, Joel.

(Fred at 00:56:18) Thank you so much, man.

(Joel Beasley at 00:56:18) You have a great day.

(Fred at 00:56:20) You too.

(Joel Beasley at 00:56:25) Thank you so much for listening to the Modern CTO podcast. Share this. Get the word out. Thank you guys so much. I couldn't do it without you.

(Joel Beasley at 00:56:32) I appreciate it. You guys are the absolute best.