Prebid, Music Royalties, and Leading Through Transformation with Tom Kershaw

00:00:02 - 00:02:32 Stephen Koza
Welcome to TechPod Talks. I'm Stephen Koza CEO at EverOps. This is the show for engineers and technical leaders in the trenches, the people building platforms, running cloud, scaling infrastructure, and figuring out what's next. Every episode I sit down with somebody who's actually doing the work to get the real story. No theory, no fluff, just what's working right now. Let's get into it.

00:02:32 - 00:02:37 Tom Kershaw
Thanks Stephen, really appreciate it. Thanks for inviting me. That's a interesting resume now that I hear it out loud.

00:02:37 - 00:02:59 Stephen Koza
Well let's let's get into let's get into the resume since you brought it up. So you've you've got a really cool background and you know pretty diverse. So let let's start with one of the earlier ones. you ran a pretty big PL at VeriSign and then moved into ad tech, two totally different industries. give us a little bit of the background and tell us about you know some of those moves you made.

00:02:59 - 00:04:28 Tom Kershaw
Well, I think the theory a around Verisign at the time is I came out of the telecom space. I most of my background was in is in internet and mobile telephone networks. And at the time, VeriSign's one of their biggest businesses was network solutions. They ran the DNS tree for North America. So all the dot com, dot net, DNS resolution. So when you typed an address, you know, back then you used DNS to resolve where that was going. And in telephone we used this horrible idea of a telephone number.

00:04:28 - 00:04:54 Stephen Koza
funny i can relate i i spent six years selling to the telecom world and put it this way i my next job i wasn't looking for something in telecom tons of fun but yeah change is always good when you've moved across industries what are the things you're like how do you learn the new space what what are the first things you're trying to understand before you go fix things or break things

00:04:54 - 00:07:09 Tom Kershaw
Well, I mean, I I I just find it that after you know two min too long of a period, four or five years in a business, you stop invading and you start just kind of repeating what you've already done. And that's to me boring and that's not how you, you know, grow as a person. So or as a professional. So, you know, I've always wanted to kind of challenge myself, but it's

00:07:09 - 00:07:23 Stephen Koza
Yeah, I love that. So when when you find a thing, does that mean you're you're delegating away everything else? And then the way you spend your time with whatever the team is on that one thing looks a little different?

00:07:23 - 00:08:13 Tom Kershaw
you don't delegate away necessarily. You just focus your effort and you have to trust the people around you to know what they're doing. And if you have the right people, they will they'll ask you when they need to know an answer. And if they already know the answer, they will just do it. And that's what I look for in in, you know, people for in engineers and product leads around me is is don't ask questions just to hear yourself ask them.

00:08:13 - 00:08:40 Stephen Koza
Yeah,

00:08:40 - 00:13:49 Tom Kershaw
Well, I think like most things in in my career anyway, it's it most innovation is is founded in desperation. And certainly this was that certainly this is one of those cases. I I do I I do have this strong belief in open sourcing almost everything. And I think the idea that my I I'm so smart and my ideas are so better much better and my code runs better than anyone else is the most ludicrous thing I've ever heard. Success in engineering is ninety percent happenstance and ten percent luck.

00:13:49 - 00:13:53 Stephen Koza
That's funny. yeah, I I've never wished for a subpoena.

00:13:53 - 00:14:04 Tom Kershaw
It

00:14:04 - 00:14:31 Stephen Koza
yeah, how cool. Love it. well let's talk about leadership a little bit. I know you've run some big teams at Erickson, you had a couple hundred people across multiple functions. When when when you think about scaling a team and it it gets large, you know, presumably you've got multiple level levels of leadership. And how how do you think about that? What what are your what are your what are your what are your thoughts and principles for running larger teams?

00:14:31 - 00:16:35 Tom Kershaw
I think there's two essential ingredients to a scaled engineering team. It's leadership and management. And nobody can do both. You can pretend you can do both. And I've I've rarely seen people who are good at both. and it's important that you understand your strengths and weaknesses and you surround and you do everything you can to mitigate those. Like many engineers who enjoy creating and building stuff, managing people is not fun. I don't like it. Never liked it. I don't know.

00:16:35 - 00:17:06 Stephen Koza
Yeah, a hundred percent. I I like what you said about inspiring people or you know, casting the vision. So early in my leadership career, somebody asked me to define it or something, and I think I gave a similar answer. Yeah, it's like you gotta set the vision, get people excited about getting on board and aiming towards the thing together. so it was nice to s hear a a version of that and like maybe I wasn't totally wrong when

00:17:06 - 00:17:39 Tom Kershaw
You could see it when when you see it in action, like I don't see it in myself necessarily, but you mentioned Verasign Los Stratenslavos was a leader. I mean he was he he inspired everybody. Now managing, you know, that was you know, that can be harder, but when you see leaders like that that are and and when I was at Google, or Sudzla, the act exact opposite. Very you know, managing was great. The you know, it's it's it's knowing

00:17:39 - 00:18:00 Stephen Koza
So I know at BMI, you you've been there a couple of years now and you were the first CTO. what what was what was that like coming into not just new company but new role and what what did you see in terms of gaps and what were some of the first things you look to go and fix or change?

00:18:00 - 00:20:08 Tom Kershaw
Yeah, it's it's been a real challenge and learning experience. It's it's great that you can, you know, be, you know, a certain part of your career and still learn every day. but you know, my most of my career I was in technology companies or the product was technology. And clearly, you know, engineering was, you know, at Rubicon, 80% of our staff was engineers, you know. And you come to come to like BMI or Travelport before, these are not technology companies. maybe by is a mu is a music company, learning to use technology.

00:20:08 - 00:20:36 Stephen Koza
Yeah. So let's talk about BMI a little bit. we've known each other for a little while and you know, our teams have spent time together. So around here, if if I ever talk about BMI, you know, somebody at EverOps that doesn't know you guys, I say BMI is the company that sends a check to Taylor Swift when her music plays in a bar. give give me give me your version or your description of what BMI does.

00:20:36 - 00:24:44 Tom Kershaw
well, I'll take that one actually. It's it's it's the easiest to explain. And and it's the one of the best things about my job at BMI is is I don't have to explain to people what I do. I just say like Taylor Swift, Miley Cyrus, you know, end of story. But the the reality is this business is transforming just at at a rapid, rapid pace. And and and when someone writes a song, especially if they didn't perform that song, we have to make sure they're compensated. We have to make sure that

00:24:44 - 00:25:12 Stephen Koza
Yeah. So tell me more about the the challenges with the scale around what you do. So you said billions of performances. To me, that sounds like a a big, big, big data problem. And you know, I imagine your platform and infrastructure and data engineering has to be built in a way to handle all that. How how do you guys do that? And what what are the things you're still trying to solve for?

00:25:12 - 00:29:37 Tom Kershaw
It's both. It's both scale and structure. You know, I think scale and structure is being separate, you know, two sides of the same coin. Structure being the complexity of the data problem and scale being the size of the data problem. And and often we've got parts of our business that are very, very complicated and structured and some that are are are simpler but have high degrees of scale. And I think the structure problem is is the big one. I mean, f well first what I'd say is that billions of performances a day by internet terms is tiny.

00:29:37 - 00:30:37 Stephen Koza
Yeah, clever, clever. our I imagine some of our listeners are lighting up when you're talking about the stuff they work on every day, Postgres and Kafka. And so thank thanks for getting a little nerdy with me. I appreciate it. You you you re mentioned AI a few times, so definitely want to ask you about that. for those of us not in the industry, you know, I I think we all see it's it's changing things, it already has, it will. you know, you can

00:30:37 - 00:34:31 Tom Kershaw
Well, I mean, I think it's it's a very complex topic and I think I don't think there's any answers in anyone in the industry has, but it certainly does elicit emotions and strong reactions up and down the community. There are songwriters that feel that AI generated or computer generated music will replace their works. There are those that use it as a as a development tool. There's lots of different opinions on this and and I can see and recognize

00:34:31 - 00:34:50 Stephen Koza
I tend to agree. I I I'm always saying the genie in the bottle thing, like you said, you know, can't close Pandora's box. yeah, I think it's pretty important. Our company's gotta figure out how to harness it 'cause it has so much to do with, you know, economic growth and superiority and all the other stuff and can't just shut it down too late.

00:34:50 - 00:35:19 Tom Kershaw
And and it's also not like, I mean, whenever I say like the Pandora's box and you say Genie the bottle, it's like it it it it basically implies we're in a horrible situation, but we need to make the best of it. And the reality is technology has done some good things for the world. Let's not be too Luddite-ish about it. I mean, the world will continue to advance. We'll continue to come up with ideas and and it's on us as a society to make sure that those are used in in in a proper way.

00:35:19 - 00:35:38 Stephen Koza
Yeah, for sure. yeah, I think people sometimes tend to have recency bias and you know, they kind of forget how we all got here over the last hundred years and you know, there's been a lot of similar revolutions, maybe not as big or you know, this may have a bigger impact than a lot of them or maybe all of them put together. But we we've seen this movie before.

00:35:38 - 00:36:00 Tom Kershaw
We've seen this exact we've seen this movie before and it all you know, there I remember, you know, there's how can you let your kid have a cell phone and you know the internet's gonna rot everyone's brain and then you then social social media wasn't exactly the you know, the greatest thing in the world in some in some ways, but you know, I think we we collectively will find ways to harness these s technologies and regulate them correctly.

00:36:00 - 00:36:33 Stephen Koza
Yeah, yeah, totally agree. one more question on AI content generation. So if if make AI makes it just like code, AI makes it easier, presumably, for people to create things. I how how do you respond to that? Do you need 10 or 100 or a thousand X times the infrastructure to keep all keep track of it all if it goes from

00:36:33 - 00:39:05 Tom Kershaw
I don't know. I mean I I I'm it that's it's a a judgment call on how you handle the same as w as I mentioned the Rubicon situation where all of a sudden the volume of requests and page loads we got went up by about twenty five X, like in a matter of a month. So your choices are your choices are don't process them all and figure out which ones to drop and which ones to process, or you drive your marginal cost to zero. Those are your two options, right? And and and so my my preferred option was driving my marginal cost to zero.

00:39:05 - 00:39:48 Stephen Koza
I wanted to ask you about transformational sort of initiative. So I I know you've done this a number of times, different organizations, different cultures. And we I've talked to other folks on the podcast and a bunch of our listeners. you know, we we've all got them. You know, there's there's some big thing that needs to happen and it's messy and high stakes and failure rates high. Are there any patterns that you've seen across different transformations that you've run?

00:39:48 - 00:41:11 Tom Kershaw
No, I mean I think with transformation, it comes down to, you know, to to patience and perseverance and and the ability to fail. when you're in a transforming business, you are going to get it wrong a big chunk of the time. If you think that you can make the right call in an uncertain environment in a changing circumst in a changing circumstance, you're wrong. I mean, there ain't no way. You the only way you do it is you you you set a course.

00:41:11 - 00:41:18 Stephen Koza
Yeah,

00:41:18 - 00:41:26 Tom Kershaw
I mean there's there's nothing wrong with that. I mean i it i it's it's the inability to admit that you screwed up that's that's that that's dangerous.

00:41:26 - 00:42:18 Stephen Koza
Yeah. I I remember a a senior leader I worked for at one point. We like we launched a product and yeah, it was kind of a failure. And you know, he got up and said, This didn't work and you know, it was my call and here's why, da da da. But here's what we learned from it, and here's what we're gonna do differently. And I I really respected that. I was like, you know, a way to like get up and you know, put put that on your shoulders. I that's that stay with me for a long time.

00:42:18 - 00:43:25 Tom Kershaw
I I don't know. I there's I I'm trying to think if there's a specific one. but I I think you know the hardest part when you're an engineer and a product person is knowing when to to stop doing stuff and when to end products and move on and what to focus on. It's just so easy to say, I'm gonna, you know, I'm gonna bill for this market in in or I wanna go international. What does that mean anyway?

00:43:25 - 00:43:49 Stephen Koza
Yeah, there's something about psychology. It's it's super easy to come up with your list of priorities and what you want to get done and stuff, you know, all all the things. but it's it's really hard to decide what to say no to, to look at the list and say, you know what, we're we're not gonna do that. We're not gonna do this product, we're not gonna do this initiative, we're not gonna set this goal because otherwise we can't focus on the stuff that matters the most.

00:43:49 - 00:43:56 Tom Kershaw
And yeah, you just you said the two key words in the same sentence. Focus is no. That's what focus means. It means the literally means the word no.

00:43:56 - 00:44:04 Stephen Koza
Yeah.

00:44:04 - 00:44:10 Tom Kershaw
Yeah. Which one's more important and the and the business team says they're all important and what what are you gonna Yeah.

00:44:10 - 00:44:41 Stephen Koza
We

00:44:41 - 00:45:29 Tom Kershaw
well, I I think, you know, being in the music business, I would argue everybody should understand like, you know, the the how AI music is being used and you know, I think you need to look at Suno and Udio and what they're doing, look at the various tools that are out there and you know, rather than recommend a book. I mean, if you haven't, you know, really actively div d dove into things like Claude and Gemini and and OpenAI, y you have to understand, even if you choose to not use it.

00:45:29 - 00:45:42 Stephen Koza
Yeah, that's a good one. I c I could do a soapbox speech about AI fluency and, you know, why you gotta learn it and understand it. And books are probably not the right way 'cause they're gonna be outdated real quick.

00:45:42 - 00:45:48 Tom Kershaw
all the books I read are are autobiographies of musicians, so probably boring to most people.

00:45:48 - 00:45:54 Stephen Koza
Cool. Love it. if people wanna track you down and find you, how how how do they do that?

00:45:54 - 00:46:02 Tom Kershaw
I'm all over social media. So whatever, LinkedIn, Instagram, whatever. I'm sure you can find me.

00:46:02 - 00:46:05 Stephen Koza
Cool, man. Well, it great chatting. Appreciate you spending the time today.

00:46:05 - 00:46:15 Tom Kershaw
Yeah, thanks so much. I really appreciate you g yeah, whatever Office is doing and you guys have helped keep us on our toes and thinking about what we should be thinking about and we we we need that help.

00:46:15 - 00:46:17 Stephen Koza
Well thanks again, Tom. Hope you have a good one. Yeah, likewise, man.

00:46:17 - 00:46:47 Tom Kershaw
Yeah thanks Stephen. Really appreciate it.

‍