Matt Walker - CTO at SimonData

(Joel Beasley at 00:00:00) Hello, my friends. I am very excited. Some special news today. We just finished recording our first master class, which covers the five biggest actions you can implement today to kick-start your leadership transformation, and is jam-packed with insights that you can leverage to propel yourself and your team forward. Visit leaderbits.io.

(Joel Beasley at 00:00:22) You can click on it. It's on the homepage, and it's absolutely complimentary. Today, we are talking to Matt Walker, the CTO and co-founder of Simon Data, and we discuss connecting the engineering team back to the customer, building a strong remote culture, and removing emotion from mistakes to grow and learn from them. All of this right here, right now on the Modern CTO podcast. Here we go.

(Joel Beasley at 00:00:51) This is the Modern CTO podcast. It's Matt.

(Matt Walker at 00:01:04) How are you doing?

(Joel Beasley at 00:01:05) Dude, I am excited. How are you?

(Matt Walker at 00:01:07) I'm doing well. It's been a pretty busy last couple weeks rolling into the holidays here.

(Joel Beasley at 00:01:12) Right? Well, look at that. You got some, like, look at your background. You got some kids, I assume, or you draw those dinosaur pictures.

(Matt Walker at 00:01:19) No, I didn't draw them. I have four kids.

(Joel Beasley at 00:01:21) Wow. Boys, girls?

(Matt Walker at 00:01:23) Two of each.

(Joel Beasley at 00:01:25) Two of each. Nice. Yep. I love being a parent, man. I've got a little girl who's about 16, and then a boy is on the way. We'll be here in April.

(Matt Walker at 00:01:34) Oh my goodness. Congratulations. That's exciting.

(Joel Beasley at 00:01:37) Yes. So we're getting started. We've only been married two years.

(Matt Walker at 00:01:40) Yeah. So we were aiming for three kids, and our third kid was twins. So that's how we got four.

(Joel Beasley at 00:01:46) Are you kidding me? That is the conversation we had in our kitchen last night. I said, yeah, you have a boy and a girl. I said, watch what will happen. We'll have boy, girl, twins as the third. And she's like, yep. Now did you have boy, girl, twins as the third?

(Matt Walker at 00:01:59) Yep. It was boy, girl. So we had a boy and then a girl and then boy, girl, twins. Yeah.

(Joel Beasley at 00:02:03) Get out of it. Dude, my wife's gonna freak. She's like, that's some universe stuff right there. How weird is that that we were having that conversation in our kitchen last night, and then you come on and you have that mix? That's insane.

(Matt Walker at 00:02:15) Twins are definitely insane, but they're a lot of fun.

(Joel Beasley at 00:02:18) You still have all your hair, though, so that's good.

(Matt Walker at 00:02:20) I'm slowly getting rid of it, but it's not just the twins that I have to blame for that.

(Joel Beasley at 00:02:28) So can you, first of all, the site's beautiful. I love your brand and everything. That's what instantly hooks me. If they have a good brand, I'm like, I'll at least pay attention to what it is. But can you give me a high-level overview of what Simon is?

(Matt Walker at 00:02:40) Sure. I mean, it's interesting because the brand actually ties into what it does. It's basically a place for retention marketers to collaborate with their data team, but you can view it as them having a data engineer in the box, and those Simon glasses represent that data engineer. So that was the brainchild of our CEO, Jason, just a sort of random image that he came up with and worked with our designer to create, and it really sort of captures what Simon does, I think.

(Joel Beasley at 00:03:10) That's, I, so the website's beautiful. And even on the home page, Tommy, the CTO of TrueCar.

(Matt Walker at 00:03:17) Yep. You know Tommy. That's right. Yeah. I've actually met him up in Austin a couple of times as well. Yeah. They have an office here in Austin.

(Joel Beasley at 00:03:25) Oh, really? They just, I got a note from Chloe, our PR, and you know, I get copied on the invitations for the future guests. And we booked him a week or two ago. I came through and said, oh, yeah, Tommy from TrueCar. I'm like, oh, that's awesome.

(Matt Walker at 00:03:41) Yeah. He's an amazing guy, and we've worked with that team now for over a year, I think. And David, one of their head of data engineering, was a colleague from Etsy as well.

(Joel Beasley at 00:03:56) Oh, and you're over at Etsy before too, right? I read that in your profile.

(Matt Walker at 00:04:00) That's yeah. That's where I cut my teeth on big data was Etsy, I think.

(Joel Beasley at 00:04:06) Any cool projects you can talk about?

(Matt Walker at 00:04:08) Yeah. So I was the first hire at Jason's first company at Tuitive, and we got acquired by Etsy. So we were kind of brought in as the search and big data team. At the time, they weren't keeping their web logs at all. They were expiring them and had no analytics infrastructure whatsoever. And so we really sort of walked in the door, set up a beacon in Akamai to collect web traffic, and then built the big data platform off the back end of that. And so that was pretty much my entire focus there for about three and a half years, was instrumenting the site so that we actually had behavioral information and then join it into data that came from their sharded MySQL. And so that was the sort of flagship data asset that they had for analyzing AB tests, and then we just automated everything downstream of that. So whether you're doing data-driven development on datasets that were produced from behavioral data, or whether you were actually just analyzing your AB test, that was all sort of automated for you through the big data system.

(Joel Beasley at 00:05:22) That's brilliant. So they were empowering people to do AB test very easily throughout the organization?

(Matt Walker at 00:05:31) Exactly. And it plugged directly into their feature flag system. So basically, every feature built at Etsy goes out under a feature flag. And if you use the feature flag to launch a feature, then you got an AB test for free out of it. So it's just a great way to democratize big data across the company.

(Joel Beasley at 00:05:47) Oh, that's so smart. And now Jason, that would be your co-founder at Simon Data?

(Matt Walker at 00:05:51) Yeah. He's the CEO.

(Joel Beasley at 00:05:53) He's the CEO. And so you met him on this previous project, wrapped that up, and then you started Simon Data right away, or were there some projects in between?

(Matt Walker at 00:06:00) So Jason's an old colleague from graduate school, which is how I wandered into his first startup. Between Etsy and Simon, I actually worked with another colleague from graduate school at Apple.

(Joel Beasley at 00:06:14) So I

(Matt Walker at 00:06:15) was working as a data scientist in fraud detection here in Austin, though I was actually pretty much remote to Cupertino during that time.

(Joel Beasley at 00:06:23) Nice. So this retention marketing stuff, is this a new category?

(Matt Walker at 00:06:29) I wouldn't say retention marketing is, but sort of our approach to it is fairly unique. So there's an emerging category called customer data platforms where you're just sort of accumulating all data about a customer and making it accessible for various purposes. The reason that we do it is for retention marketing in particular. So it's things like abandoned carts, new user registration, upsell, cross-sell, churn prevention, you know, post-purchase series, you name it. It's anything where you're marketing to your existing or former customers. And so it's all built upon having this very rich profile of information about the customer.

(Joel Beasley at 00:07:11) So Simon Data, they act as a consultancy where they go in and they take a look at all your data and then build you dashboards, or just give me, what were you doing with TrueCar? How did that relationship go about?

(Matt Walker at 00:07:25) It's not as much of a consultancy as a data platform and a suite of products over the top of it to allow retention marketers to do their daily workflow. I guess where the sort of consulting aspect of it comes into play is sort of in the initial integration and then getting people trained up on the platform. But what we do is, you know, a new user walks onto the platform, we sort of predetermine what their data inputs are gonna look like, and they just go down the list hitting the OAuth buttons and submitting the API keys and giving us access to the data, basically. Then our integration team pulls all that data in, and it's made available through the product directly to the marketer. So the marketer's building audiences through our segmentation tool. They're building campaigns through our flows product. They're crafting their email templates and what have you, and then they're actually running their own campaigns through the platform across multiple channels.

(Joel Beasley at 00:08:23) Interesting. So you're like a little data agency marketing platform for companies, and they plug their data in and they can use it internally?

(Matt Walker at 00:08:35) Plug their data in, and then they get their marketing out, whether it be email, advertising through custom audiences, push notifications, text messages, you name it. With TrueCar in particular, one of our big sort of, you know, primary differentiators for them was actually just passing data through the system multiple times using our segmentation tool. And so they actually have these workflows where data will flow through the platform and then go back into their system and then flow back through the platform again for messaging. So it can get fairly sophisticated when you're dealing with a data-savvy user.

(Joel Beasley at 00:09:09) That's interesting. Because I don't think about that very often. They have these, TrueCar would have a large audience with hundreds of thousands, if not millions of users, and then they have to coordinate their efforts across all of them. And a lot of the tools that you can self-sign up for are usually for B2B or B2C sales, right?

(Matt Walker at 00:09:31) Yeah. Exactly. Or you can kind of imagine them being cases where your subscriber list fits on an Excel spreadsheet. And we are really talking about, you know, subscriber lists in the millions or hundreds of millions of contacts.

(Joel Beasley at 00:09:46) Wow. You guys send a lot, you guys use Postmark or SendGrid or something like that? Or you do integrate with whatever they want?

(Matt Walker at 00:09:52) Yeah. We support a wide variety of email service providers, but SendGrid is one of our preferred ones. Yes.

(Joel Beasley at 00:09:58) Nice. What was the, was there a specific moment or inspiration that you and Jason were like, we're doing Simon Data?

(Matt Walker at 00:10:06) So it really came from experiences at Etsy and then in the market where, you know, a lot of what we were trying to do was apply data science to real-world problems and letting it be driven by, you know, sort of the use cases that we found in the market. And a lot of what we saw was that marketing teams had sort of more baseline data problems to begin with. They just weren't well-served by their data engineering teams, or their sort of internal data assets weren't aimed towards marketing, and they weren't really utilizing the data that even their company had curated into a data warehouse to actually power the marketing. So just like at Etsy where we saw this building tools on top of data platform to democratize access to the data, we did the same thing for retention marketing teams, you know, out in the market, essentially. And it was funny when we were sort of shopping around for data science problems, the majority of the work that we found was data engineering in support of these marketing teams. So I think what we've done here is sort of built our ideal data platform to solve the data problem, and now we're able to actually layer data science applications in on top of it as well.

(Joel Beasley at 00:11:25) Nice. So the real hard part was getting that initial platform of, let's get everything connected in here, and then now let's do really great stuff with it.

(Matt Walker at 00:11:32) You got it. It's connected. It's cleaned. It's QA'd. It's making inoperable data, in some cases, operable. And then finally, it's going the last mile over the top of that and actually delivering, you know, predictive models to really get the last mile of optimization for these marketing campaigns.

(Joel Beasley at 00:11:50) Nice. So you're friendly with, when a company, they have some data engineers, you're usually, they're the ones that are heading up the integration with you, and then you're making their lives easier too?

(Matt Walker at 00:12:02) Yeah. We can, we interface in a variety of ways. If there is a data engineering team, then they're often sort of a very close colleague at the client account, and we get them into our SQL development environment so that they can build some of the datasets. But with many of the clients that we work with, they don't have a data engineering team, and so we provide the data engineers to be able to actually ingest data into the platform for them.

(Joel Beasley at 00:12:29) Nice. How do you feel about sales? You like sales or no?

(Matt Walker at 00:12:33) Sales at Simon is a great crew of people. They are extremely bright people, always very interested in, you know, sort of learning technical concepts because I think it helps them in the sales process. One of our AEs, um, knows what Hadoop is. So Josh, we were riding in a cab, and Josh put him on the spot and was just like, hey, what's Hadoop? And he actually accurately described it to me, which I was totally impressed by. So they're a really cool crew of people, and I think that they do a very good job in this, in our sort of sales process, which has a lot of steps. Because you are selling a data platform, there's a lot to it, a lot of features, and so there's a large number of conversations with many different stakeholders required to really actually make the sale.

(Joel Beasley at 00:13:22) And so you've had, over here, we have a company outside of the that came out of the podcast. And so we have about 10 people now, so we're dealing with all that, developing your ideal customer profile and paying attention, because it's weird because you see the whole market, you know. It's like, oh, somebody has a technology team, it would be great. They could use our thing because anyone who has a technology team. But then we found that there's a certain part that responds or a certain part that replies, and they have this very specific issue. And it's like, so now we've, like, coordinated down, it's only engineering teams of between 20 and 100 engineers. And they're at this level of growth. And for some reason, we don't know why, but they are experiencing the problem that we solve the most, and so they're purchasing the fastest.

(Matt Walker at 00:14:03) Yeah. And that type of market exploration is one aspect of what they do. They also, so we've just sort of, as a company, tried out different market segments over time as well. Right? So one trend that we've seen is we've sort of been moving towards larger and larger companies because we support their scale.

(Joel Beasley at 00:14:23) Mhmm.

(Matt Walker at 00:14:23) So, you know, we can walk into one of those conversations, and it's just very natural. Yes, 120 million contacts or what have you. No problem. And a lot of our competitors have issues with that. The market for retention marketing is pretty diverse, though. There's a quite a range of sort of tech savviness in that position from, you know, down-market shops where the retention marketer is doing it all, towards the higher end where they're sort of one component of a larger marketing team that's a component of a much larger sort of, you know, growth team.

(Joel Beasley at 00:15:02) So what's the growth been like at Simon Data? How long have you guys been around? Where are you at today?

(Matt Walker at 00:15:09) Yeah, so we've been in market about three years. And one of the interesting things about the first year was that I was acting as our integration engineer. So we have this really cool program where we hire junior engineers out of coding academies, and then we pair them up with a core mentor, and they go through a yearlong boot camp where they're learning core engineering while delivering clients work through the integrations team, and then they graduate onto our core team. Well, the success story with that team is that they automated me.

(Matt Walker at 00:15:46) Previously, that was my job. I was actually the one who was building the data pipelines in the client conversations, figuring out which data to pull and how to query it and all this sort of stuff. And now that team has completely taken that off my plate and does an amazing job at it as well.

(Joel Beasley at 00:16:03) So how large is engineering as a whole there? How many people?

(Matt Walker at 00:16:07) So we're at 26 engineers, not including the integrations teams right now, and we're looking to essentially double that in the next year.

(Joel Beasley at 00:16:16) It's really smart how you solved your pipeline problem with the pairing. I like that. I've heard about it before. I've read about it, and I've actually seen it, but it's amazing that so many people don't do it.

(Joel Beasley at 00:16:32) That's the amazing part.

(Matt Walker at 00:16:33) It just happened to fit with us very nicely because we needed this integrations team, and it's a really cool opportunity for a junior data engineer because in a lot of the jobs that you get fresh out of training in the industry as a data engineer, you really don't get to get thrown into the deep end. Whereas our integrations team is just casting you directly into the deep end and letting you figure out how to operate data systems at scale, which is the fun part of data engineering.

(Joel Beasley at 00:17:06) Now customer success, really important to you. Right? How does that look there? What's important to you? What do you do for customer success?

(Matt Walker at 00:17:14) Yeah, so I think one thing that differentiates our approach to others is that our customer success managers, we call them TCSMs for technical.

(Joel Beasley at 00:17:25) Mm-hmm.

(Matt Walker at 00:17:25) And so they are very tech-savvy people who can understand the ins and outs of the platform. They're often competent SQL authors, so they can write SQL queries. So they can go pull data on their own. They can do analytics on their own. And now that we've built a strong suite of tooling for them, they can often integrate accounts entirely on their own. So it's been another success story for us in that they're very fast at integrating client accounts, and they're able to do that while not actually being programmers, but they can also manage the client relationship as well.

(Joel Beasley at 00:18:05) That's pretty awesome. They've got client relationship management skills, and they're also SQL-savvy.

(Matt Walker at 00:18:11) And they're a great crew of people to work with because they are good communicators and high-energy folk who are very friendly with our clients as well. And then I guess the other thing is that my brother is the director of the team right now, so I have direct affinity with them.

(Joel Beasley at 00:18:27) That's awesome. You gotta, with the holidays coming up, you've gotta keep the conversation at home about data science. Yeah. So, mistakes. We're all making mistakes, right? We're always learning. What are some of the lessons that you've learned recently in the past couple years from early on in founding the company?

(Matt Walker at 00:18:50) Yeah, I mean, I think one of the chief mistakes that we made earlier on was just not anticipating the scope of the product. So we started out building the segmentation engine effectively, sort of learned by entering it into the market that we needed to expand the feature set more towards a CDP. And recently, we've learned more that we need to expand the feature set more towards a marketing cloud. And so this has just been naturally revealed through interacting with our customers. The things that make them happy are making very simple workflows over the top of the data platform. And so we've gotten a lot of wins in the last couple of years by just delivering product over the top of the data platform that really either automates or just makes simpler their retention marketing workflows.

(Joel Beasley at 00:19:44) That's interesting because what they're having to do is they're getting sold all the time on these features of things you can do, then they don't have the data backing to actually execute on it, or it becomes a ridiculously large project. But you guys went the opposite way. You're like, let's go in there and give them the infrastructure, and now they have — and so that they can deal with all the sales that they're trying to make happen and all the features they're trying to bring in. But then you realize, hey, we already have it. We already have the relationship with the customer, which is the hardest thing to get anyways. Let's just put these flows on top of it.

(Matt Walker at 00:20:16) Yep, you got it. And I mean, it's just becoming more and more easy to use the platform all the time. As you identified, it looks nice. It's pretty user-friendly. It's not very intimidating, but you can encode some fairly sophisticated marketing campaigns in the thing. So the abilities of it just keep growing all the time. So it's a very cool platform upon which to build that type of product. And that's sort of our central problem is taking these data-heavy applications and delivering them to nontechnical users.

(Joel Beasley at 00:20:50) So I just wanna get a little bit geeky here. I'm sorry. I wanna understand it just a little bit better. So let's say you went to TrueCar and they were using it for their advertising world, but then sales wants to start leveraging that data. I heard you use the word "democratizing" a lot. How does that look for the company? They come in, you get all multiplied up, their marketing's going, advertising world's going really great with it, but then sales is like, "Hey, we wanna tap into that data." How does that work?

(Matt Walker at 00:21:16) Yeah, so we have a couple of approaches to that. First of all, we can be very incremental in the way that we onboard data. So essentially, when we go in to integrate an account, we bring in the smallest amount of data possible to actually drive a campaign. And that way, we can go end-to-end on one use case immediately. Then as we identify new marketing use cases, we just add new data to the platform. So the platform is very good at just adding incremental data. So what happens when we start to go multi-team within an organization, we have a permissioning framework to be able to sandbox the different subteams. But it looks very similar to a new integration. So if you got everything working for retention marketing and sales wants to actually use the tool, they can start by just getting a user account and seeing what's available. Then you can come up with use cases and sort of bring in new data to support those. And then pretty soon, you've got people across multiple teams using the same data platform.

(Joel Beasley at 00:22:19) Nice. I like how you continuously — outcome-driven, right? You're always use-case first. That prevents...

(Matt Walker at 00:22:25) We do that across the board.

(Joel Beasley at 00:22:27) Yeah. That's how you prevent from getting the data swamp. You got it. So what stands out to you as signs of a great engineering culture?

(Matt Walker at 00:22:38) Sure. I mean, I love our team. So I mean, I guess one thing that's relatively unique about our team is that we are distributed.

(Joel Beasley at 00:22:46) Okay.

(Matt Walker at 00:22:46) And that sort of owes from the fact that I've been remote since 2007 in my career. And then I've been remote specifically from Texas to New York City since 2009, so a long, long time. So we have a very good culture of daily stand-ups amongst teams, weekly meetings across the entire team, and quarterly off-sites where you get to meet face-to-face and do planning and also just meet and greet. The mentorship I talked about before is another sort of core principle where we're constantly learning and teaching each other. And like you say, it sort of establishes our pathway for junior engineers onto the senior team. And then I think last is just that we're highly effective, in part due to the flexibility that remote work allows us to have. So people tend to keep office hours, but if you need to take a break in the middle of the day, go eat lunch with your wife or what have you, and then sign back on in the evening to do your engineering work, you can. So I often find myself doing my management work in the day and my programming at night when things are quiet, and I think that makes this pretty effective.

(Joel Beasley at 00:24:01) What sort of leadership style do you have?

(Matt Walker at 00:24:05) I'm totally a lead-by-doing kinda guy. Like I say, right now I'm filling the roles of VP of engineering, CTO, sprint team lead, and one of our lead ICs. So I'm writing code, I'm building the team, and I'm building technology.

(Joel Beasley at 00:24:23) You're still writing code?

(Matt Walker at 00:24:25) Absolutely. I think it's very important for me to be writing code because it's the way I maintain context on the whole system.

(Joel Beasley at 00:24:32) Okay. Now do you have plans to reduce your amount of code writing?

(Matt Walker at 00:24:37) No. I have more plans to reduce my management. Okay, cool. Yeah. So I'll continue to manage the senior engineering organization, but not from an HR perspective, more from an architecture perspective, right? And then I think part of doing that high-level architecture is just actually having your hands on the code so that you know how the system works. You know all the moving parts. I wanna be involved in the hardest technical decisions across the entire stack, and that's what I really like to do.

(Joel Beasley at 00:25:10) Okay. Yeah. That's the beauty of being the cofounder, right? Yep. Yep. You get to — I love it because I talk to all of these different CTOs, right? Huge spectrum. And what I find, you know, that they have in common is they all hang out in the area that they're strong in, and then they hire out around the other stuff. So some people are really strong people — like leaders type — personalities, and so they really crave that, and that's what they like to get in and figure out how that's working. So then their SVP of engineering will be the one that is the extremely technical one. Other people are really technical and they really wanna be on the hardest problem engineering-wise at the company and kinda close out everything else, hire out everything else. One of my examples is, Steve Jobs is still a people person, but he totally shut the world out for the iPhone. That was the only project he was working on. He completely closed everything out and was just obsessed with that even though he was the CEO of the company at the time.

(Matt Walker at 00:26:11) Yeah. But as the CEO, that's pretty impressive. I mean, I have some flexibility being the CTO, but I definitely — I do go through those cycles where I really have to get heads-down, and those are some of the most fun times for me. But like you say, it's a good investment from my perspective building out this team of people I can trust so that I really can go heads-down while the company stays afloat. The engineering department keeps cranking away, and it can kinda run without me.

(Joel Beasley at 00:26:39) So do you have a main person that's helping you, like your right-hand person?

(Matt Walker at 00:26:46) I don't have a specific individual, but I have a team of people managers that are extremely talented and have been with the company for a long time. And I'm starting to develop a team of team leads that are sort of the senior engineers that really organize the actual execution of the engineering teams. Nice. And so between those two groups of people, I've uncovered, basically.

(Joel Beasley at 00:27:13) Nice. It's smart. I always like to see — this is a fun conversation for me because it's always interesting how people structure things around them in order to have the outcome that they want. And you're doing awesome.

(Matt Walker at 00:27:23) Yeah. I mean, it's been a very iterative process. It's interesting how much it's changed over time, but that just is completely described by how much we've grown. I mean, we doubled last year, and we'll double again the next year, and the actual structure required to manage it is different.

(Joel Beasley at 00:27:40) And you're early. I mean, you're — it's like you're three years old, right? Now, your platform and you being involved in those decisions is really important, especially with the way that the — the thing that's interesting about the field you're in is it's really cutting-edge right now, right? There's so much — there's so many different areas to pull your focus.

(Matt Walker at 00:28:01) The focus is one of the scarcest commodities but most important things for the team, in my opinion. So I mean, in general, you're right. But definitely, given that we are fielding a data platform and a marketing cloud, there's just an infinite number of things we could do. So we're very focused about picking our priorities and then just incrementally building against them to make sure that we're on the right path.

(Joel Beasley at 00:28:29) And your customers help you a lot with that too, as you mentioned. Yeah. The CDP, that was customer data platform? Customer data platform. Yeah. And then you were looking at that, and then you found out you were more of a marketing cloud? Yeah.

(Matt Walker at 00:28:42) We're essentially a marketing cloud that has a baked-in CDP. So a marketing cloud with a really strong data platform underneath it. Okay. And I think that's — it's the combination of those two things that really meets the use cases for retention marketers.

(Joel Beasley at 00:28:56) Yeah. And so how would you classify a tool like Segment? I've seen that, but I wasn't really — I played with it a little bit. I piped my Intercom data into it, and I...

(Matt Walker at 00:29:06) Yeah. We integrate to Segment. They're actually an upstream event provider for us. So Segment makes it very easy to deploy multiple beacons essentially into your web app or your mobile SDK. And so they have access to the event stream data and can provide you with downstream analytics on it. We sort of would view them more as an input, and we can take their events, curate them within our marketing data warehouse, and then provide access to their data plus other data. So in particular, just like we learned at Etsy where we were putting behavioral analytics data side-by-side with this sharded MySQL that had all of our production database data in it, it's really the combination of the two that's required for retention marketing.

(Joel Beasley at 00:29:59) That's awesome. I love this stuff. It's like a whole new world, man. It's so exciting. Have you heard of — I don't know why this just popped into my head, but lots of conversation around productivity of engineering. And then I saw this thing come up called GitPrime. Have you come across this yet?

(Matt Walker at 00:30:20) I have not, no.

(Joel Beasley at 00:30:20) All right. So what they do is — I found it so interesting. I screenshot and then I sent it to Chloe, and I was like, get this person on the show. I wanna talk to him. So I think they're gonna be on the show in March or April. But...

(Joel Beasley at 00:30:31) They go through your Git repositories, and they give you some real measurable KPIs or indicators of your engineering based on what code they write, how it goes back and forth, how it gets into production. And then how many bugs it has once it's in product. It's got this whole life cycle so you can quantify the effectiveness of your different engineers, your engineering.

(Matt Walker at 00:31:00) Right? I mean, so we have accumulated metrics for that exact kind of thing over time, but it's definitely not automated in any way like that. We use Asana for project management, and so we're tracking throughput through Asana. We do have some kind of vanity metrics that we derive from GitHub, but we don't utilize those to really gauge throughput. It's more Asana that's the primary source of information.

(Matt Walker at 00:31:30) But, you know, we get metrics out of the code base for a variety of reasons, you know, to understand complexity, to understand security vulnerabilities, all sorts of stuff like that. So that is a pretty interesting thing to do static analysis and be able to tell you the story of your team through metrics.

(Joel Beasley at 00:31:47) Do you guys check out the episode with Prashant, the CTO of Asana?

(Matt Walker at 00:31:51) Amazing. I will totally do that. I love Asana.

(Joel Beasley at 00:31:54) I flew out there. I've got a picture, a selfie of me in front of the Asana sign in San Francisco. Nice.

(Matt Walker at 00:31:59) Yeah. Yeah. We're heavy users across the board from management tasks to engineering to client success to sales. We use Asana all over the place.

(Joel Beasley at 00:32:10) So you actually will interface with your clients and use Asana?

(Matt Walker at 00:32:13) So the clients aren't there, but the PS team manages their time through it.

(Joel Beasley at 00:32:19) Okay. Yeah.

(Matt Walker at 00:32:19) Yeah.

(Joel Beasley at 00:32:20) It's definitely one of those things where we use it here too, and it's like, why rebuild it? It works, and it's pretty, and it's functional, and it's real time. It's like, why even—it's good enough.

(Matt Walker at 00:32:36) Yeah. I mean, for us, the only thing to do is sort of build some procedure over the top of it and some structure to make it analyzable. So that's really the only investment we've had to make is come up with the desired structure for our sprint boards that we use uniformly across the teams. And then we just have a script that dumps the sprint boards out and then can produce metrics off the back of that.

(Joel Beasley at 00:33:01) Oh, wow. So you guys integrated with it on your side?

(Matt Walker at 00:33:04) Exactly. Because it doesn't quite do the reporting that we need specifically for our sprint process, so we just make up the difference with some code.

(Joel Beasley at 00:33:12) Oh, nice. Have you showed that to him?

(Matt Walker at 00:33:14) We have not. No.

(Joel Beasley at 00:33:16) You should show it to him. Here's what he'll do. I'll intro you to him, and I'll be like, you gotta check out what they're doing. They're integrating with you and doing sprint stuff and analyzing reports. That might give them some inspiration about, you know, what they should be working on over there.

(Joel Beasley at 00:33:29) Absolutely.

(Matt Walker at 00:33:29) If they build that kind of sprint velocity reporting into the tool, we would definitely use it.

(Joel Beasley at 00:33:35) Yeah. But it's always interesting to see because that's the thing about with APIs. You don't really know what they're doing with it.

(Matt Walker at 00:33:40) Yep. Right.

(Joel Beasley at 00:33:41) Yeah, I hope they're doing something good with it. Okay. So any challenges with leadership as far as being remote, or is it similar to in person? I know you've been remote for a long time.

(Matt Walker at 00:33:53) Yeah. So it's definitely something that I have experience with. There's definitely challenges that can arise from being remote. People have hesitancy to jump on a video call or pick up the phone, and you kinda have to get over that very quickly. But I think it's something that remote workers intuitively learn very quickly. It's when to escalate the conversation from Slack, which is not super high throughput, to just being on the phone where you can have the conversation much more real time. And with technical escalation, it's very obvious typically what the priority of something is. So it's easy to just jump through the escalation hoops there. With management issues, it can be less obvious when something's really important to someone. So you just have to be super verbose and ask.

(Matt Walker at 00:34:49) Yeah. Right?

(Joel Beasley at 00:34:50) I love that word, by the way. It's one of my favorites.

(Matt Walker at 00:34:53) Yeah. I mean, I think we do a good job. I think last week I had something like 2,000 lines in Slack or something like that. So we're extremely active in Slack.

(Joel Beasley at 00:35:06) So are you doing any public speaking for Simon Data, anything like that?

(Matt Walker at 00:35:10) We really haven't trotted Simon out on the engineering side the way that we have on the marketing side and sort of sales side. But we have plans to. A lot of the people who live in New York attend meetups, and Jason does a lot of sort of tech talks there as well. But it is the kind of thing where we're looking to start getting on the conference circuit and sort of spreading the word about Simon.

(Joel Beasley at 00:35:37) Nice. Off topic question, is that a Nike shirt?

(Matt Walker at 00:35:41) The internal one? Yeah. No. It's some kind of just generic, soft threaded shirt that I wear under my fleece here when it's cold.

(Joel Beasley at 00:35:52) I have a Nike shirt that's that exact pattern. Yeah. I'm wondering how close you are to me in a parallel dimension with the whole conversation last night. Are you wearing the shirt I was wearing yesterday?

(Matt Walker at 00:36:04) Yeah. I'm in the alternate dimension here. Yep. Yeah. So in Austin, it's weather change time. So the reason my voice is kinda funny is that cedar season is starting up. So it's time to be allergies for the rest of the winter, essentially.

(Joel Beasley at 00:36:19) Oh, man. We got it—we just got a little bit of cold here in Florida too. It's like—

(Matt Walker at 00:36:25) Oh, you're in Florida? Yeah. Very cool.

(Joel Beasley at 00:36:27) Yeah.

(Matt Walker at 00:36:28) Or maybe warm. It would be the—Yeah.

(Joel Beasley at 00:36:30) Well, no. It's like this week, it's been 40 and 50.

(Matt Walker at 00:36:33) Yeah. Same here.

(Joel Beasley at 00:36:34) Yeah. Usually, it's 75, 80. So it's sweater weather. But it's hard because in the morning when it's 40 or 45, I'm so used to my morning runs. And it's like, I don't wanna run with the cold air, so I drive ten minutes to the gym because I just—for some reason, I feel like when you're running cold air, your potential to get sick triples.

(Matt Walker at 00:36:56) Yeah. For me, I also run in the mornings, and I just lose my voice throughout the day with meetings because you get so much congestion from running like that.

(Joel Beasley at 00:37:10) So let's bring it back to Simon. Yeah. What are you most excited about today?

(Matt Walker at 00:37:17) That's a good question. I'm excited about the team, but I feel like I've kind of got the pattern there. And so now it's really just executing, growing this team that I have in mind over the course of the next year. So, really, what I'm excited in doing is sort of turning my attention to the next round of technical problems.

(Joel Beasley at 00:37:37) Mhmm.

(Matt Walker at 00:37:37) We just have some immensely challenging problems to solve, and that's what I'm really passionate about. So, you know, whether it be in data science or whether it be in scaling the platform the next 100x or what have you, that's where I really wanna spend my time in the next year.

(Joel Beasley at 00:37:55) Yeah. Like, how do we go from a billion to a trillion events?

(Matt Walker at 00:37:59) Exactly. I mean, we get to scale by different orders of magnitude across the stack while we're scaling the number of clients that we're onboarding by some multiple as well. So it's a very fun problem of scale.

(Joel Beasley at 00:38:14) Lots of Redis queues. Yeah.

(Matt Walker at 00:38:16) Lots of Memcache.

(Joel Beasley at 00:38:17) So lots of Memcache.

(Matt Walker at 00:38:18) That's a good point. We don't use Redis. We use Memcache. I'm very content to sort of make up the application tier difference over the top of Redis using Memcache because it's tried and true and very easy to operate. So Memcache is one of those things where we put it in service three years ago and have never had a problem with it.

(Joel Beasley at 00:38:39) It's old. I mean, I was using Memcache in the early 2000s. But it's such a stable technology.

(Matt Walker at 00:38:46) I know. We utilize it extremely heavily, and it just stands up to all the punishment we mete out. Like, it is a good technology. Yeah. Doesn't mean bad.

(Matt Walker at 00:38:55) Yep. Right? Old means stable.

(Joel Beasley at 00:38:59) Great. Well yeah. Right? Alright. I gotta know. You're Matt. You're Matt Walker, and he's Jason. Who's Simon?

(Matt Walker at 00:39:09) Who is Simon? Who's the persona?

(Joel Beasley at 00:39:11) Yeah. Your third cofounder?

(Matt Walker at 00:39:14) Yeah. That's—no. Josh doesn't have glasses, so it can't be Josh.

(Joel Beasley at 00:39:19) Is it the person—if you were to blend you two together?

(Matt Walker at 00:39:22) Yeah. That sounds about right. He leans more to the data scientist, and I lean more to the data engineer, but I will point out that I have the glasses. So yes.

(Joel Beasley at 00:39:31) And he is not in Texas with you. He's in New York, you said?

(Matt Walker at 00:39:34) No. We were in college at UT Austin together. He was in Texas, but he lives in New York now.

(Joel Beasley at 00:39:40) Oh, he lives in New York now. Yep. Nice. He just wanted—he was like, we're too close, and now I wanna work remote, so I'm going to New York.

(Matt Walker at 00:39:49) No. So he went there primarily to start his first business, which was an advertising business, and that's a great place to kick off an advertising company.

(Joel Beasley at 00:39:57) Yeah. Well, what I'm finding being in, you know, Tampa, Florida—I'm an hour south, an hour outside of Tampa.

(Matt Walker at 00:40:04) Okay.

(Joel Beasley at 00:40:04) And so what I'm finding is as our business is growing over here, there is real benefit to living in a big city. Because you can just look at all the brands in the big city, you go to meetings with them. Now if I wanna go meet with, I don't know, Harvard, right? I was in Boston two weeks ago, or Mimecast—they do email security. I wanna go meet with them. It's $200. Like, to me, you'll have a client meeting. Right? It's—there's a lot of benefit to being in a huge city and doing business right in your area.

(Matt Walker at 00:40:36) So we kinda get the best of both worlds because we have a really strong presence in New York. So people have close access to all the things going on in New York City. But, you know, we have engineers in cities across the country. So we have people in DC. We have a person in Montreal. I'm in Austin. So we kinda get a blend of the different tech cultures across that. It's nice to have people representing you in multiple cities.

(Joel Beasley at 00:41:05) And then how do you do—so you said you're mostly distributed. Do you have an office there in Texas? How do you do—you earlier, you mentioned that you had this concept of the student pairing and—how do you—

(Matt Walker at 00:41:21) It's all done remote.

(Joel Beasley at 00:41:22) It's all remote.

(Matt Walker at 00:41:23) Yeah. So when I mentor someone, what that means is I'm essentially sharing a screen with them for an hour each week. Right? So I'm doing code reviews, showing them some code that I worked on, reviewing code that they have done in the last week, answering questions that they have. But then also what we do is we collaborate on producing technical brown bags. So the junior engineer will basically build a presentation under the guidance of the senior engineer and then deliver it over lunchtime to the rest of the engineering team. So it's a great way for junior engineers to sort of share what they've learned while also getting input from the entire rest of the team.

(Joel Beasley at 00:42:01) Technical brown bags, junior engineer—explain this again. I'm taking notes now.

(Matt Walker at 00:42:06) Yeah. So, I mean, our mentorship program is very well defined in that they're on a year-long path towards graduating senior engineering. They have these sort of quarterly checkpoints to sort of gauge how they've been doing, but the way that we gauge how they've been doing is not just tasks accomplished. Right? There's also some component of it that's how are you doing towards your progress towards becoming a core engineer? And part of that is these pairing sessions with a core engineer who's helping you learn what's necessary to become a core engineer. And so they pair up to code together. They pair up to do code review. They pair up just to do Q&A kind of stuff as well. But they also collaborate to build technical presentations that we deliver through what we call a brown bag. And so that brown bag is just a presentation given over lunch. Everyone dials in and eats their sandwich while the engineer is sort of giving a presentation about some technical project that they've worked on.

(Joel Beasley at 00:43:10) Nice. So they work on a little technical project, give a presentation about it, and then they get—do they get to pick it, or is it assigned to them?

(Matt Walker at 00:43:18) Yeah. So, typically, what we do, even in the hiring process, when we first are screening these junior engineers, we ask what they're interested in doing. And so we sort of figure out, does this person have an affinity towards full stack development, or are they really heavy with data science, or just what are their interests? And that's how we choose who we pair them up with. It's also how we give them sort of side projects. They typically are high priority but low urgency tasks so that they can work on them over a period of time. And that just gets them hands-on experience with the kinds of tasks that they're gonna be working on in the team that they graduate into.

(Joel Beasley at 00:44:03) Your whole pipeline, how you get people—I guess internship job ad post things. Right?

(Matt Walker at 00:44:09) Yeah. We do a lot of recruiting through the graduation events of the code academies as well. Because we have people on board who are from the code academies, we can go back to those events and recruit more.

(Joel Beasley at 00:44:23) Okay. So you go to these—you know these code academy people. You recruit at their graduation events. Yep. You assign them to a mentor with the area where they would have the most success. Yes. You have an outlined, clearly defined system for them to travel through this program, this mentorship path to becoming a core engineer program. That's a twelve-month program. They have these series of events, quarterly reviews, type deals, technical brown bags, code together, review together. They're getting paid. Obviously, it's like a job. Right?

(Matt Walker at 00:44:58) Mhmm. Yep.

(Joel Beasley at 00:44:59) And then they graduate into a core engineer. Yes. And then that's your pipeline for talent.

(Matt Walker at 00:45:09) That's absolutely right. And I mean, while they're in this program, the type of work they're doing, whether it be client requests or integration work, is extremely valuable to the company and to them. So they're getting a lot of very hands-on experience integrating accounts, understanding how the data pipes work, seeing the full range of technologies from data input to data output, and getting to contribute across the stack. So it's really a great warm-up to actually then graduate onto the senior engineering team and have a ton of context about the product and how it's used, so that they're amazing contributors from day one once they graduate.

(Joel Beasley at 00:45:49) That's interesting because then if you're bringing in senior people from the outside, they have to start at day zero.

(Matt Walker at 00:45:54) You got it. So booting up the senior engineer is actually more challenging for us. What we do is tend to put them through rotations in different parts of the stack to give them exposure so that they can come up to speed on the tech. But in terms of how they're actually exposed to clients, it's great that our sales team and our CS team are so collaborative because they invite them to actually sit in and shadow client calls to gain domain intuition. So core engineers are invited to attend sales calls and listen in and just hear how we're talking about the product, and they're invited to go to client meetings as well and just meet the people, understand what they're doing.

(Matt Walker at 00:46:39) I mean, it's one of the things that I took away from being an integration engineer in the first tranche of customers was that I felt like these people were like my colleagues, and I was working side by side with them to actually accomplish their marketing use cases. And it taught me a lot about what they're trying to do and how they want to do it, which informs everything that I do in terms of engineering today.

(Joel Beasley at 00:47:03) Yeah. Well, it's important because you don't want it to become too isolated. Like, this is engineering over here, we're doing the hard work. And you want to keep it open and keep people connected because that isolation hurts the business in so many ways.

(Matt Walker at 00:47:21) Right. It limits access to engineering from the sales and CS side. But from the engineering side, they lose context. They just don't know what's going on whenever they're myopically focused on engineering tasks, right? So I totally agree with you. I think it's one of the—as the team gets even larger, it's one of the problems that's going to be most challenging for us to solve from a team dynamics perspective.

(Joel Beasley at 00:47:41) I like it. Man, we're wrapping up here. This was a fantastic podcast. I love that. I think the coolest thing—well, I know your product's really cool—but I've never had someone explain their pipeline for talent that clearly, and that will help a lot of other CTOs out there.

(Joel Beasley at 00:48:01) They're going to be replaying that section of the episode over and over.

(Matt Walker at 00:48:05) I think that's probably a symptom of having an engineer build a hiring pipeline, right? We're fairly analytical about the way that we compose the team and the hiring pipeline too.

(Joel Beasley at 00:48:17) Right. And it's logical. I had—we did an investment round, and so I had the VCs that came and came into the actual physical location, reviewed it and said, "Oh, let me see your sales pipeline. Let me see how you produce this. Let me see how you produce that." And I've got documents on how we do everything. I'm like, "How else do you know how to do it?" And they're like, "This is unbelievable." I'm like, "What type of other companies do you see?"

(Matt Walker at 00:48:40) I mean, you're doing it right. One of the things I didn't call out is how heavily we use Google Docs.

(Joel Beasley at 00:48:45) Yeah.

(Matt Walker at 00:48:45) We just—writing things down is incredibly important, and we've iterated on them over the years, but it is extremely useful to just have all of that thinking written down, especially for new hires, just to be able to say, "Hey, here's how we do it. Read this document, and you'll understand," right? At least get a first pass on the way that we try to do things at Simon.

(Joel Beasley at 00:49:07) Have you ever tried to convince someone how they forget? Because I've had people that are like, "Oh, I don't want to write it down." I was like, "No, you're going to forget." We forget at the rate of one bit per neuron per second. We're essentially goldfish. We're going to solve this problem, and then we're going to forget the solution to it within minutes or hours, and we're going to be like, "What was the exact wording on that? What were the exact steps for that?" And then we're going to have to resolve the problem again, and that's going to cost us more time, and we're trying to accomplish a goal as an organization.

(Matt Walker at 00:49:38) I mean, that's interesting. I never thought about it from the standpoint of forgetting, but it does tie into our postmortem culture. The chief reason that we perform those postmortems aside from making the tactical remediations to process is to never forget. We've got the archive of postmortems. It's like, these are lessons learned. Don't ever do this again, right? And so we remediate it from a process standpoint, from a tech standpoint, but then also we have a historical archive of just everything that we've learned over the last few years, right?

(Joel Beasley at 00:50:12) This is interesting because I do this, but I haven't talked to other people about it, and then you bring it up. So yeah, I believe that our knowledge as an organization of what we learn—for example, when we're going out to talk to different parts of the market. CTOs between 5 and 10 people don't have any interest in us because their role is completely different than a CTO with 500 to 1,000 engineers, right? And while they're the same role and people would say, "Oh, that's the same," they're not actually that on a day-to-day. And so we've gone around and talked to all of these groups, and there's so many different types of them. And we know exactly—so that's just for our messaging information. But the same thing with how we deal with our actual customers, interfacing with our customers different ways based on what type of customer they are. But yeah, that's interesting that you talked about that. I like the postmortem thing. You have a list of all the scripts and the things we never need to do again.

(Matt Walker at 00:51:12) Exactly. And I mean, it's really ingrained into our culture that if an incident occurs, we postmortem it, and we have a very formulaic process for going through the postmortem from timeline to the five whys to the remediations. And that really helps sort of take blame out of it, take ego out of it, and just get down to the actual root cause. Like, what went wrong? Why? How can we fix it? Make sure it doesn't happen anymore.

(Joel Beasley at 00:51:43) Yeah. And I like that culture, and it's important to establish it early because I noticed I was establishing it with nontechnical people, and just as part of our thing with scheduling guests or whatever it may be, things that we don't do. And at first, the first couple times you do it, some of them were like, "Oh, you made me feel bad" or whatever. Then they realized this is just the culture. A mistake occurs, we identify the root cause, we put a process in place so this doesn't occur again, and we move forward, and we don't care. It's not about you doing it. It's like driving on the interstate with bugs. They're going to hit the windshield. Things are going to happen. We just need to make sure to acknowledge it, understand it, put a process in place, and then move forward.

(Matt Walker at 00:52:24) That's right. And it can be very awkward, but the process helps sort of get over the awkwardness of it. And on that note, both our sales team and our CS team have adopted the postmortem format that we use and have conducted really interesting postmortems that give us business insight.

(Joel Beasley at 00:52:44) Matt, I had a really fantastic time talking with you today. You guys don't—so you guys don't have an office, correct?

(Matt Walker at 00:52:50) We do have an office in New York. Yes. So the business side of the house is actually physically located in New York City. Yeah. But all of us engineers are distributed, so we have our home offices.

(Joel Beasley at 00:53:02) So if I'm passing through—you said Austin, right? So if I'm passing through Austin at a conference, I got a big speaking schedule for 2019, so I will put you on my list and get you to come out and say hello, and it'll be fun.

(Matt Walker at 00:53:14) Absolutely. That'd be awesome. I'd be very interested in it.

(Joel Beasley at 00:53:17) Awesome. Well, you have an excellent afternoon. Enjoy the holidays. And if you ever need anything from me at all, any relationships, any questions you want me to put out to the network, you just message me, and I'm available to you whenever.

(Joel Beasley at 00:53:29) Sure.

(Matt Walker at 00:53:29) Thank you very much, Joel. It was really nice to meet you.

(Joel Beasley at 00:53:31) You too. Talk soon, Matt. Thank you so much for listening. If you'd like to help, please take a moment right now to open up the iTunes app and leave a review of the podcast.

(Joel Beasley at 00:53:46) If you take a screenshot of the review and text it or email it to a friend who needs to listen to the podcast and then CC me, [email protected]. If you CC me on the email, I'll send you a copy of the Modern CTO book or give you a shout-out on the podcast, whichever you prefer. We're trying to get listed on the top 100 for iTunes, and I need your help in order to do this.