Media/Appearances/YouTube

Skip Zapier, Go Straight to Code: Inside the Future of Sales & Marketing Ops

Cody Schneider/Benyamin Holley/

Cody Schneider/Aug 4, 2026
Transcript

0:00:00 So what is a GTM engineer and why is it one of the fastest growing roles right now that's coming into the market? Why are so many companies hiring for it? GTM engineer is a go to market engineer and they basically use software to build pipeline for companies. They use tools like Claude Code and all of these different APIs to basically wire together automations and workflows and even agents that can do tasks or run autonomously. Now on this episode today I'm hosting my friend Ben. Ben is a GTM engineer who's worked at multiple startups and he's going to be talking about what the actual job role and function looks like and how companies can be successful by implementing these people at their organizations.

0:00:50 Specifically, some of the things that I found really interesting about this conversation is we're going to talk about how API completeness is now the buying criteria for software. An example of this is how Salesforce is actually better now than HubSpot because you can do more through the API. We're going to also talk about how you can build UIs just in time for your team for internal tools. And this interface can be anything that you need it to be because you can just vibe code it. We're going to talk how you should just leap over entirely from Zapier or N8N. Don't even touch them. Go immediately to code. The benefits of that and why it's the better way to start building these solutions. And then we're going to talk actual workflows.

0:01:44 What does Ben's day-to-day GTM engineering work actually look like? Where are the opportunities right now on cold outbound? And how you can actually build pipeline for your sales team. And this episode is brought to you by graph.com. Graph.com is the platform to deploy marketing agents with. It's a data pipeline, a data warehouse, and then also a cloud server that you can push code to or agents to. With that stack, you can build basically anything that you need for marketing engineering. And the best part about Graph is that four deployed software engineers can do these implementations for you on top of the platform. They're basically like Palanteer, but how you do if you're trying to implement marketing agents for your business, like a Facebook ads agent, a Google ads agent, an SEO agent, a cold email agent, go to graph.com to learn more.

0:02:43 All right, let's get started with today's show. Ben, what's up, man? We've been uh chatting online for like months, but I'm I'm so stoked to actually have you on and be able to share. You have so much knowledge on the GTM engineering side. I feel like you're one of the few people that's really actually doing this in organizations. Like a lot of these guys are just like, you know, lone wolves. Maybe they run it like a a, you know, an excl. They've now moved everything to a super basin, like a railway server, right? But I feel like you're one of the few people I see that's actually doing these GTM systems within companies.

0:03:30 So man, you you I read you said this the other day online and I uh it just always resonates with me. You're like Salesforce is now the best CRM for go to market. Uh why do you believe that? Like why did you say that? Uh I I'd love you to go deep on that for me. Yeah. So um so first of all it is I don't care what anybody says. Like listen adio is cool.

0:04:01 We used HubSpot at AirOps, which I was not a super huge fan of, and I don't think it's like a big deal to say that, but um all these other CRM like they they have cool features, like don't get me wrong, but actually what's happened uh now that you can just plug the Salesforce CLI into Claude Code, all of that like legacy complexity and all of those like crazy endpoints and everything um that used to be like a huge pain in the ass are actually like an asset in the AI era. and Bitcoin say I'll give you an example of something that's really stupid probably to most people but is a uh symptom of how good Salesforce actually is now which I know it's crazy that we are having this conversation like we're out here selling Salesforce like we should be getting kickbacks on Salesforce hope for this hopefully but um when my uh my previous company Sift you could literally provision user security settings from the terminal Um, and some people might say, well, who who cares?

0:05:15 Well, first of all, if you're a Salesforce admin or like an IT person, you definitely care. Um, but it's that sort of like attention to detail that makes something actually usable by agents. Um, that's so important now because think about like you could have you could build your own onboarding flow. Like if you wanted to set something up in cursor with all the endpoints, it's like okay well the rippling API is like this person's onboarded. What what's their uh user group? Oh, they're an SDR. Well, we already have all that provisioned. Um, and just go do this [ __ ] like automatically. And I'm sure there's services out there that do this, but for enterprising teams who want to run lean, who don't want to buy a widget for like every [ __ ] business process, like this stuff is a huge asset.

0:06:14 Not even not even getting into any of like the actual sales and revenue generating like potential of that. So, of all those features, right? So, that's that's my like 10-second bit on why Salesforce rocks and everything else sucks.

0:06:30 I I I think this is a larger like theme too, right? I mean, this is basically how honestly I got to tell you this like my how I think about software and choosing purchasing software now is based off of what you you saying this like six months ago. Like I I am now all I think about is like how robust is the API and do I have to like if I can't do everything through the API that I can do within the UI of the website like I am not buying that software at this point because it's like why would I ever try to touch a UI and and really what it comes down to even more so is like if I need a UI I'll just build a custom one like we're doing this more and more now where it's like if I need some type of visual way to interact with like whatever it is that doing I will just use my like I I will build the custom solution that I need like just in time right cost of code has gone to zero and I can basically build the exact solution that I need um like if I need to click a button or like view you know have it build a flowchart to understand what's going on this I think it relates it's funny I've been having a lot of conversations with companies now where they're like why why wouldn't we use Zapier or like NM to do this like why would we why should we go straight to code like it's hard to visualize it I'm like just ask it to visualize it just ask it to make a work like it's just a it's a flowchart that's that that's not complicated.

0:08:31 It can do it can show you that logic. So yeah, that's so funny. The portfolio well you you guys can't see in the chat but I was sending my portfolio earlier and like it's literally all these mermaid charts. I'm like hey I have all these like old repos make mermaid charts out of these. Like there's nothing uh sophisticated about this stuff. This is like an amalgamation of like all the systems that I was running and this is like some some of the uh smaller subsets but like all this stuff is like generated by cursor like um and there's a ton of different like pieces of code out there that it can like then string together for like that big mermaid chart. Yeah.

0:09:22 Um, you know, it's funny that when you talk about like NAN or something, somebody was um telling me that make was hiring for a job in Raleigh or whatever, which is kind of near me. And and I was like, dude, like how are these companies still in business? Like if you're at Zapier or Make like I'm not trying to be a dick, but like how are how are these companies still in business? Um because when I used to use those things, dude, it was such a headache. And like I like to think I'm pretty smart. I just I think I have like, you know, ADD or something and I'm like I can't I can't get this [ __ ] workflow to work.

0:10:13 Um, and when you're in cursor or something, it's like, "Hey, fix my workflow." And you know what's crazier about this too is like NAN had an opportunity to actually really get people more embedded into their tools when AI started like coming out because think about it like the stuff that we were talking about with Salesforce, like NAN already has all that. They already have all their endpoints and they have metadata about what these endpoints are supposed to do. And like why did it take them so long? I don't even know what they're doing now, but like I assume by now they've solved this problem. like why did it take them so long to like act upon the market forces and like actually take advantage of that like I don't know

0:11:07 totally I mean I think and this was the major learning like in the epiphany I came so I was using n for a lot of the stuff that I was doing on like the marketing engineering whatever you want to call it like this this category that we're in um but I found myself I was using Claude Code to create uh or I was basically using a a state-of-the-art model to create JSON files that I drop into NAD and I was like the exact same thing. Exactly. And you're like I'm like I was like I'm vibe coding here. Why am I doing this? Why am I not just going straight to a server? Right. Like why am I not just pushing code?

0:11:57 And I think that it leads into the next thing I want to talk about with you. You you you say this all the time um about really like code is the truth now of like what you're doing as a go to like for go to market like it's almost a log of your actions and your activities and you know everything that you're doing at a company. I I think you advocate for this a lot. Like if you don't have a GitHub repo showing the stuff that you've done, right, for the work that you've done at companies like like that that is what the future of this looks like, you know, for everybody. They you need to learn how to build these systems. What's your thoughts on that?

0:12:50 Like do you do you are you seeing I mean more I I I think we're both kind of on the forefront of this. This is how we think about everything now is like I'm not trying to hire a marketer traditionally. I'm trying to hire like somebody that uh can actually build systems and use all these AI tooling to like build out processes, not just like do individual, you know, campaigns that are one-offs. But I'd love to think about I'd love to hear about your systems design and just like how you're approaching like I am writing code that build, you know, and trying to create this log of all my actions if you would want to go down that that rabbit hole with me.

0:13:41 Yeah. Yeah. So um so first uh I think when we talk about like hiring people um it like revenue operations people uh go to market people engineers um I don't know where this is going to end like I guess like you don't want an enterprise account executive to have a GitHub rep. I mean, it'd be cool if they did, but like I don't know if that's super relevant for them, but like even like the SDRs at AirOps were like vibe coding stuff and um like that's one of the things that I enabled some of them to do is like hey we have the growth team GitHub if you have automations put them here and like we can like build upon this stuff like over time and uh we even had some of like the less technical marketers.

0:14:38 I was like red pilling them on this stuff. I'm like use the GitHub so that like I can help you with this stuff, other people can help you with this stuff. Um, I am seeing like the top tier of like GTM engineers like have GitHub repos and I think that a lot of the I don't want to say like whatever middle tier I because I don't know like how talented these people actually are. Um, but I'm seeing more people adopt it and to me like not having a portfolio like there's so many other jobs where you'd have a portfolio and like that's your I don't know stamp of approval or or calling card or whatever. And uh when like people are trying to get into GTM engineering, they're like, you know, I'm interviewing for GTM engineering roles, but I'm not having much success.

0:15:37 Like, what should I do? I'm like, well, you need a portfolio and you need to like be working on these problems over time because it's not the same as just being like a random like campaign manager where you're like, "Okay, we sent out an email campaign. Uh we'll see if it works." It's like, no. like um I had uh a folder where it it had all of my email campaigns that had ever been sent, all of the copy, the background, the context on each one.

0:16:13 And this was not something that I finished, but it's something that I need to do is like build like I haven't gotten into loops yet. Maybe maybe you can educate me on those later, but like uh and Eric Noski talks about this too. I think there's probably some other people who are a little bit smarter than me about this particular point where like every time you send an email campaign, you should recursively use AI to like improve it over time based on like response rates like improve segmentation like things like this.

0:16:51 And so that's like something that would sit in a in a GitHub repo. But yeah, I I mean the bottom line is if you're not like keeping a record of your work either uh for yourself or for your company to like that's the that's the real value in it is I think like when a GTM engineer is like building something and leaving it there for the company when they leave like Aerops has all the stuff that I built, you know, in the current GitHub repo. So yeah. No, absolutely. I I it's funny you say the loops thing. I feel like it's people over complicate it so much and it's like just what a human was doing previously but now we're just like using agents within that process.

0:17:44 Like for example, say you're doing a cold email campaign like how would you find success? I would be like cool like if you're starting out from zero we cold email a bunch of people we create pipeline we like have calls we look at who closed and then we're like okay companies that have this shape like they're the best companies. Let's go find more companies that have this shape.

0:18:12 But I think the granularity that you can get to now is really interesting cuz it's like, oh, I can go and I can scrape their similar web traffic to be like, okay, what is their what is the similar web traffic of a good client look like for us? Okay, I can go and scrape their um for example um uh like job post, listing history. Like I can scrape like there's all there's all these different things that I can use to enrich like who what does our target customer look like? both from like a firmographic standpoint, but then also like you can start to uh uh like mine the sales transcripts, understand like where were they at at a given moment in time.

0:19:03 Can we find other businesses that have like that psychoraphic that are you know they're that look like for example like um you know we can go into signals probably all day and there's people that are way more advanced on this than I am but I mean one of the easiest signals is literally like or two of them that we see the most success with is like did they interact with content on LinkedIn that's related to the product that we're trying to sell.

0:19:37 Um and the second one is like are they hiring like for specific job roles that we basically like augment or totally replace that job role, right? Um and then from that it's like okay is the company size that's appropriate. But I think on the loop piece that's the thing like it's just it's exactly what a human was doing previously but just didn't have time for but now we can just do it at a scale and out of granularity with also like more metrics than we previously could. So

0:20:14 yeah. Well, uh, that's what I was going to say about the loops, too, is like you actually have to know what information to enrich, which is kind of like a crapshoot. So, it's kind of like you could you have to go get everything like a hundred, you know, different data points. And, you know, at that point, I feel like sometimes like the granularity and detailed like nature of all that information can actually be like distracting. Um, also the funny thing about like people say this about like sales conversations and stuff too, but dude, listen, I I mind thousands of conversations from Gong uh at AirOps and there was a lot less information there than than you would hope.

0:21:03 And this is nothing against the sellers at AirOps. I love you guys, but it it's not always like people don't ask the best questions for for me, you know, which is not the point of them having a sales conversation to be fair. No. Yeah. doing a talk track. It's like how different can it be if they know like if you know a talk track works and it actually closes like you're going to do that same talk track you're not like how are you feeling today? Like how many people did you hire last month? Like nobody's going to [ __ ] ask any of that stuff. No, totally. Totally. Yeah.

0:21:49 I mean, we just like all that information we're just collecting anymore from like Reddit or from like social ch listening channels like we're using uh more and more just like scraping like Tik Tok and Instagram to find like what are the people talking about that are in this category and like can we figure out different or unique ways to talk about the product. Um because I I feel you you said this before we jumped on the call.

0:22:21 It's like everything is becoming red ocean right now. like whether it's paid ads, whether it's SEO, whether it's cold email, whe like any cate like any any everything is becoming just absolutely cooked. Um the I I think the only way that worst thing that you can find like edge is you have to basically map out what is everybody doing in the market and a lot of the times that your competitors are all positioning themselves in the exact same ways that you are or like you know they're all like they're basically like you know if you think about it as a a matrix right where it's like here's you know five different ways I can position and here's like 10 different situations I could be in. they're all kind of they'll they're going to be bunched together.

0:23:18 Um, and what you'll find is that like rather than competing in that red ocean space, if you go over to this blue ocean space where like nobody else is positioning themselves or like talking about the product in that way, it's so much easier to compete there. And we see this within the ads all the time where it's like, okay, can we find some type of alpha within the way that we're describing the product or the pain that we're talking to, the outcomes that we're talking to, etc.

0:23:54 Um, but anyway, I think it's it's also like it's something what we're seeing and that's really top of mind for me is like certain categories are extremely competitive. Like I I I imagine with air hops like you guys there's so many companies that are doing like AI search observability like how do you differentiate in a category like that gets really challenging but in contrast like we're working with a company right now they're like a motion graphics company um and for them uh like there's just not a lot of players in the category yet and so it's just so much more blue ocean and I think this is just you know anyway I'm curious your thoughts there like how you're thinking about it or how how are you like trying to extract or find that arbitrage or find that edge within the current market?

0:24:54 Also, just what are you seeing on the front lines from a competitiveness? Well, okay. So, yeah, this is a great conversation to have because um Owen Clancy, who is my the VP of growth, he left, but he was the VP of growth who hired me at AirOps, amazing guy. If anybody's on LinkedIn, you should go, he doesn't really post anything, but you should go hit him up. He's he's doing recruiting now. So anyway, um he would send me emails that like he would get because people, you know, oh, VP of growth, they're going to send him emails even though we literally sell what they're they're trying to sell him. And uh it was very interesting.

0:25:41 Like some people did a really good job with their emails. Um and some of them did a really bad job. But um I don't think I think from a like a to zoom out a little bit, I don't think that most people understand the market that they operate in. And I think about this all the time, right? Like people are like, "Oh yeah, I work at this ventureback company. We just raised a series B." I'm like, "Okay, well, you just raised a $20 million series B or whatever it is." Let's remember that the GPS in all those funds uh are LPs and other funds and they're derisking their investment by investing in five other companies just like yours who have the exact same positioning, who are selling the exact same product.

0:26:36 We're all just a bunch of monkeys in a knife fight for the venture capital firms, which is fine. I'm not complaining. That's that's what we all signed up for. or I don't know if you signed up for that. I don't know if you're bootstrapped or whatever, but yeah. Yeah, we we raised a preede in October. Um and we we just had to to like make this like we infrastructurally we couldn't do it bootstrapping. So yeah. Well, uh we signed a Fiverr contract. We love you Fiverr. Thank you for all everything you do for us. But like yeah, there was no way we were going to be able to handle that. So

0:27:22 yeah. Well, so you know this uh raising money is an obligation, not an accomplishment. Uh it's the the beginning of the work and not the end. But um but yeah. Uh uh well there's a couple things I can't really talk about that we were doing at AirOps but that they wouldn't want me to talk about but um positive things like uh we were trying to get a differentiation on our positioning and go uh deep into one vertical that we really believed uh we could conquer um and especially like an AI search the thing about like Aerops is like that we weren't just doing visibility but that's what you know right like and that's the problem that's that's that was our problem that we were dealing with right because

0:28:15 profound was our biggest competitor they sort of I think we can say that they won the visibility round for like the last year. Um so you know hats off to them uh for sure but they also raised an insane amount of money and I don't think that their revenue is like I just think that they have a very long road ahead of them. Um so best of luck to to them and they're dealing with the same stuff.

0:28:48 But um yeah finding differentiation is really tough. I will say like one of the things that I worked on right before we left was um this lead magnet which I think I can talk about this I don't think it really matters um where basically we went really deep on analyzing um everything about the company's a company's website all of their AI search performance and we got like the snapshot in time about like what how they were actually performing against competitors and things like this right so the lead magnet was great um it generated a lot of pipeline for us in a very short amount of time and it was really what I thought of it as like a V1 one of a sales enablement product and like marketing product that you could actually spin off into a bunch of different directions which was going to be like phase two.

0:29:49 And that's what I would tell people. I'm like, "Okay guys, like I know the lead magnet was a huge pain in the ass for us to build and maintain and all this stuff, but it's V1 of a platform that we're developing for our sales and marketing team." And um I don't think that every industry has like a tool that they could build like that that would generate alpha. uh if anybody wants to know about it, I can talk in more detail about it after, but like um message me. But um trying to find alpha and like leveraging it in interesting ways and presenting it in interesting ways um is something that like I think about a lot. The problem with this lead magnet though is like okay, think about this, right? Like you get an email.

0:30:46 We're like, hey, we did some analysis on your search AI search performance. SEO offers are so commodified like VPs of growth, web masters, I don't know if that's a term that SAS bros use, but like the directors of digital marketing, I don't know, whatever. They get so many emails where they're like, "Hey, your your uh traffic is down, you know, all this like [ __ ] like commodity sort of offers." And so, how do you actually present that in a different way? It was a much different question and like a much harder question to answer. And I think we did a pretty good job of it by like naming specifics from the research that we actually did. said, "Hey, you're not doing so well against this competitor. We actually have a bunch more uh we actually have an amazing data set on this.

0:31:45 Um we would love to walk you through it and uh you know, see if it could be useful." And so that anyway, I'll just like pause there. Um but that's sort of like how we were thinking about it. No man, I think it that's such a like the insight there is like you have to have some type of proprietary information like everybody can go find anything that's public, right? So like that historically that was just like can you systematize the process of getting unique insights and sending them at scale right

0:32:23 um but now it's you really have to have some like secret sauce and that's the only way that you can actually compete I also think that there's this something to be said about like like the cost of software has gone to zero. You can use that as like a lead magnet now. So I I have a friend who's running this company is in the mental health space. um they have this mobile application and like their data looks like super good from a retention standpoint all this extremely cheap and I'm like why is it why are you charging so little like this is crazy and he's like Cody this is marketing for me cuz like if I get all these people into this app and like have them you know basically using this constantly then I

0:33:19 can go and it creates pipeline for me to like you know sell mental health like counseling right like to like whatever better help etc and I it kind of was this epiphany to me of like okay you can make these like micro bro, whatever you want to call them, like like features and they become almost the lead mag in itself. Like it's like the software or this like custom thing that you built is is what you use to get people in the door. And I think that gets like pretty exciting when you think about that at scale.

0:34:00 And it's it's really fascinating, man. I I think that um like this is also like a part of like this new like go to market you know your GTM engineer marketing engineer role is like what what free things can you like what what can you give the market that's relatively zero cost um but historically would be impossible to actually ship because like you you couldn't actually get you know product used to be the hardest thing to get out the door because engineers were the most expensive you know asset that we had an organization I mean that's kind of flipped it's like I you can build anything now now you Now, how do you build the thing that is actually useful and like have that filter for that?

0:34:53 Um, I think everybody's calling this uh like right now it's it's basically taste, which I think is like such a dumb way to think about it. It's more just like what does the market have appetite for? That's actually the better way to think about it. I think it's almost like expertise though because think about like companies like Clay and Aerops was somewhat moving like in this direction. Um, and I here this I'll explain it. Jesse over at Lead Magic has this like bit that he does.

0:35:29 He's like, "There's no such thing as SAS. Uh there there's software and then there's services and they're not the same and you have to sell both of them." And I think that's kind of like where we're moving like with something like the lead magnet because like like to be fair like a lot of the stuff that Aerops does like you could do on your own but if you don't have like alpha and expertise to execute on these projects like you may put yourself in a bad hole. like a lot of the stuff that you do. This is why we wanted to have you on one of our webinars. But um I think the the decision was like um if Cody comes on the webinar, he's going to show people how to do air ops for free or whatever. But like

0:36:27 that was the problem. I was like, "All right, maybe I can figure out how to like fit this together where I like the tooling is a part of the process." But anyway, yeah, go ahead. But but no, I mean I still think we should have had you on because I don't think that stuff matters because what happens is what like uh okay, let me try not to go off on another tangent, but like HubSpot just implemented their like AEO uh tool. It It's not good. It's actually really bad and it's going to lead a lot of people astray and they're going to end up wanting to buy something more mature and purpose-built. And so like if somebody tries to run the Cody Schneider playbook, they're probably going to shoot themselves in the foot.

0:37:22 Like just being real. Uh, most people do not have the expertise to do. It's the same thing with cold email. People are like, "I sent a million emails." I'm like, "Nothing happened." I'm like, "Okay, well, let's like let's examine what you actually did and maybe we can talk about like why it didn't work and maybe why you need some level of better expertise because doing the actions at volume is like we said, it's not the problem." Um, so anyway, I man, I think everything you're saying is exactly what we're experiencing and feeling like on our side of like the aisle.

0:38:04 Um, so like, you know, I can't tell you how many conversations I had where it's like, I had Claude Code running our SEO and it looked great and now it's absolutely nuked. And it's like, yeah, it's cuz you weren't, you know, you you it's still like the tool is only useful by a person that knows how to wield it, right?

0:38:29 Um, like I and I think it's that same principle like all this is is just a tool for for execution and and there's ways to create those guard rails. Um, but I mean we're, you know, we're we're still figuring out like how do we do this like in a self-service motion, right? Like with graphs and what we're finding is that there is a subset of people that are like incredibly competent with and know how to do this already and maybe they're running it locally and like that's actually where we're finding this opportunity right now.

0:39:09 I'm curious your thoughts here. It's basically the infrastructure to deploy marketing agents cuz it's like everybody's monkey patching this together right now with like okay I I've got some data stream maybe I got access to like you know I I I begged my CTO to get access to the data warehouse um and uh I'm now like deploying random agents to railway but I have no observability and like consistency.

0:39:38 Um, and so basically what we realized is like, okay, can we just unify that where we're just going to give you literally a CLI you can give your agent and you're like, go and spin up this workflow I just made locally, put it into a cloud with a live data stream and it can go do that, right? Um, but anyways, um, I first off curious to get your thoughts there in a second, but I I think just to come back to that that core idea like the expertise is still the most valuable part, especially now with like everybody like trying all competing, right?

0:40:21 This is I think this idea is related to what we just talked about where all all of the marketing channels are becoming more hard to compete on because everybody can just do the average route, right? Like so how do you have differentiation? How do you stand out? um that you you it's becoming more and more valuable like if you really are a good you know top 1% direct response copywriter and you know how to get um you know an AI to write in this in your style for for example email or cold email or you know even scripts for for ads right the the likelihood that you're going to be able to out compete with the exact same tools is way higher right so I think that that expertise like there was this moment where it's like oh you could just throw anything up because it, you know, it was the 0ero to one, but we're past that now.

0:41:26 It's now moving into okay. Um, it's the one to one to 100, which is way way harder because it's just it's no longer blue ocean. Like there's way more competition there. So, but anyways, I'm curious your thoughts on like, you know, where we see this is where the market's all going and um that like everybody is trying to deploy marketing agents. I swear to God, a new one pops up every day and it's like whack-a-ole at this point and it's hilarious to me. And it's like you look under the hood and you're like, "Holy [ __ ] you're just spamming Reddit threads." Like that is the software.

0:42:11 Um, in my, you know, in my view, like what what is an actual agent? It's like, okay, some type of like job to be done and you have like a thinking loop that's within it with like an LLM making decisions. Um, and then you have a data stream, you know, that's basically coming back to the agent so that it can understand the outcomes based off of the actions. And that's just like it's very it's very simple, right? Like I'm not trying to create God and put him in a box and like have him have total like control. Like I I don't really want an agent to make decisions and things. I I want it to run a process. And I think that the the thing that we're realizing is like the infrastructure for that is kind of hard to like get built quickly.

0:43:10 You can do it like for example like you can build a data pipeline, a warehouse and you know have a server that you deploy agents to. It's like you could do it with air bite with uh click house and then you could have you know your agents running on railway but that's still like engineer it's just hard. It's just like we've tried to do it

0:43:38 and maybe this gets easier as like frontier models go along and and and you know we're totally wrong but our thesis is that that that like everybody's going to want to go to code immediately like anybody that is building some UI to like connect together sequences like I've seen this with all of these marketing like engineering startups. It's like it's like this is just a Zapier clone. Like you're you're ending up in the same space. That's not how this is going to work. Like how the the best are performing right now that are like you know and I talked to heads of gross at companies like every week like they are all running their systems out of Claude Code and the challenge that they're facing is they're like they're trying to deploy that into like

0:44:34 something that can either run on an hourly or a daily cadence. I'm curious just to get your thoughts on like what you you saw, you know, or what you're seeing on the front lines if that feels like it's in the right direction or if you're like, "No, this is like, you know, something entirely different. Is it is it going in a to a different place?" Um, yeah. So, uh, a couple things. Uh, so first of all, like the, uh, marketing agents that, uh, spam Reddit for you, uh, that [ __ ] is not going to last. Like that that door is going to be closing very quickly. Um, and there was some news about Reddit uh possibly not renewing their contract with Google that allows Google to scrape them for AI overviews. Um, which is going to screw up a lot of people.

0:45:33 Um, and I don't know, I don't care if they do or not, but if that's what your product is built on, um, you need to you need to find God pretty much. Um um as far as like what the best heads of growth are doing and stuff like this um I don't know if this answers your question but basically like the way that I was I was running a lot of my processes was out of Versel um with just the code that we've we deployed and so like you can almost imagine a future where like instead of you know where your directory of vibecoded slop revenue agents are is like inversal And instead of logging into like Salesforce every morning, you're like logging into Versell to like make sure that none of them errored out overnight and um you know, you click on the little Versel tab and it opens the app and you go to the the dashboard and you you know, you see what's going on.

0:46:42 Um I I don't see people like buying more apps. Like it just think about like some of the most successful things that I did at AOPS was around like job changes. Um, and it was 100% vibe coded. Like, could I have gone and bought user gems or something? Like, probably. Yeah. But there's like so many little nuances around like account ownership and you know, all these, first of all, vibe coding a job changing bot like that automatically sequenced people and enriched stuff like was like way harder than I felt like it should have been. Like maybe I'm an idiot or something, but it seems so simple. And um I remember one time I got into like a a fight on LinkedIn with the CTO of user gyms because I was like joking around about how anybody could vibe code user gyms.

0:47:40 I don't regret it. I wasn't wrong. But I will say if you are listening to this, you are right. It was not easy. Okay, I admit it. You concede. Tell me how you did that just so the audience gets some nugget there. So, what just for context for everybody who's listening as well, like what what uh Ben is talking about is basically uh like come person switches to a new role and you're trying to as quickly as you can get to them basically with your offer because like a job change is typically a high signal that they're going to try to implement some type of new solution at the company that they're at.

0:48:27 Uh tell me about uh like what was it what was the actual tactical like what did you use in the stack to do this? Yeah, so uh we used common room for the actual signal. Um, and the common room folks, I've had discussions with their product team, so this they're not going to be mad at me if I say this, but like Common Room is great is a great tool, but there's a lot of shortfalls for it. I I probably would have used something else to be totally honest if uh if we didn't already have Common Room. Um, but they're trying to do great things over there, so hopefully they do it. But, um, uh, all all you need is a web hook that, uh, you need something hooked up to your CRM. So obviously it has everybody's LinkedIn URLs.

0:49:26 Um then you need a database that or a you know enrichment provider that's going to tell you that they changed jobs. Um I mean that it sounds really simple but the the problem was like all of the stuff downstream from that like all of the exclusion rules like did were they have an open up? Are they a customer? Because I had two different sequences. I had former customers and then former prospects. Former prospects is really kind of overstating the matter because they may not know anything about us.

0:50:02 Um, and all I wanted to do was like en enroll them in a limbless list sequence. The the hard part like like the part where like complexity started getting in was like account ownership. Um, you know, if an SDR I felt like if an SDR had helped book the meeting that the original customer came from, like they should get the downstream of that. But that's not really how we wanted to do it. Um, just making sure that if the account wasn't because this was the other problem, right? We have an account scoring model. if an account's not tiered, we're not going to work the account.

0:50:46 And so, like, if it was a new account that wasn't in HubSpot, I had to send it to a clay table to get scored. And then I had to like make a web hook that would like wait for five minutes for it to like either show up in HubSpot or not. And then I, you know, then I had to like route it to a different person based on the taring and like all this [ __ ] Um, and so that's why people buy user gyms, I guess. But, uh, I think that it's probably explains the the full thing. Totally. And then it would go into a limitless sequence. That's that that was the end of it. like it goes into a sequence. That's pretty That's pretty simple.

0:51:39 Totally. Uh I got somebody that is like the very Midwit version of this, but it's basically just like their LinkedIn profile. Um you just all of the you know again your ICP, you have them in a database. Um, and daily you just use an appy to scrape it or weekly you can do this like go to each profile and pull out the job listings and then just have an LLM that's like a super cheap LLM be like is there a new job roll on here from last time and that is a way to get signal. What kind what kind of scale does that work on though? Like how expensive is that?

0:52:25 You can get profiles for like like fractions of a penny. So imagine if you're doing 10,000 like it would be cheaper than what you would pay one of these signal companies for for sure. Yeah. I mean we had like 200k contacts in HubSpot at least. Probably more than that. Like the week before I left I had like 50,000 more contacts. So we were on our way to having a lot. But

0:52:55 I don't know. I mean I think it would even be at that cost. I mean it depends on how the frequency you want. But like I I mean you could do this weekly, you could do this uh you could do this daily depending on your list size. It just depends on like I mean I don't know what a contract for like common room looks like um on a yearly cadence but like I I again it's just that's that's all we're doing right it's just seeing like is the LinkedIn did they change right so anyway

0:53:35 there was a bunch of other stuff we were doing with common rooms for sure it was just what I had at my disposal but yeah there's probably a midweight version of that that would have saved me time but like now I did it and I'm like next time I'm doing this I'm [ __ ] doing the midw version like this [ __ ] was too complicated for Sure. For sure. So, um, okay, man. What What am I not asking you about that I should be? Um, like we've we've been here for a minute. Like there's so much I feel like changing in this category. Also, like for the uninitiated, like GTM engineering, what is how do you think about it? What is your definition of that? I feel like it's just like this thrown around term.

0:54:31 And it started out as like, you know, people doing waterfall enrichments. That was like the origin. And now it's evolved into like, nah, man. I'm writing like custom software and you when you hire me, I literally have agents I bring with me to the team. What do you actually view it as? And then what am I not asking you that I should be?

0:54:57 Um, well, well, you're not asking me what kind of job I'm looking for because I'm looking for we can deal with that later. But um uh yeah for to me GTM engineering is a set of principles around applying software solutions to sales problems and a lot of people are going to be like well you know sales isn't a product problem and of course I agree sales is not a product problem it's not this it's why like you see like engineering founders really struggle and why they make bad first sales hires and why they want to hire VP of sales for 350k a year who like does nothing for them and things like Um to me it's about like finding uh friction points in like sales processes and this could be anything from like revenue operations, marketing, sales development is kind of like where my like specialty lies like finding friction in those processes and like working with like SDR teams uh marketing teams to like build lead magnets things like that.

0:56:08 Um and then like utilizing and measuring all that stuff but um it it is fundamentally different. I think the best way to explain it is why it would be fundamentally different from like revenue operations because first of all, nobody's asking me about finance stuff. Like nobody's asking me to reconcile billing. Like you don't want to ask me to do that. Like cuz I probably wouldn't do a good job. But um but it is interesting because I actually saw a job listing for a post sales GTM engineer and I was like, I've never heard this. What does this mean? And when I went onto the listing, somebody had, you know, sent it to me. um it was around like onboarding, CPQ, finance, billing, and so really it felt more like a bisops person with like a cursor subscription. But I was like that's actually really cool.

0:57:08 That would be a really cool application of these principles um and like ways of working that uh are GTM engineering adjacent. Um I don't think GTM engineering is just about spamming people at this point. like whenever I would run an email um I mean we would use email bison but like all of the inboxes and stuff you know I have a provider for that like and um there's a lot of things within email deliverability that require like custom coding like inbox rotation and like automatically rettagging an inbox that's like burnt to re be warmed up for you know the next 3 weeks and things like this but um I know that's kind of like a a long-winded answer.

0:57:57 It's not marketing exactly because like typically I'm not doing like creative or campaigns or or ads for the most part. Like I did uh add people to ad audiences based on like certain things, but that's um just like one little thing. I don't do too much of in ads.

0:58:18 Um well, I think uh if you're looking for a GTM engineer, you should hire me. No, I'm just kidding. Um, but uh if you're looking for a GTM engineer, like I think you should find somebody who has a real sales background and not just a a computer science background because to me it's more about the jobs to be done and understanding what actually closes deals and drives revenue and like building things in service of that as opposed to like you know a bunch of dopamine hit like workflows that like sound cool and like I talked to one guy who had hired a GTM engineer quote unquote GTM engineer and I was like okay well what happened cuz they're like we let him go or something and he's like, "Well, they just made like a bunch of dashboards.

0:59:14 Like they vibe coded some dashboards." And I'm like, "Okay, well, that's not really driving the business forward, but um okay." So, yeah, there's a lot there's a lot of things to look out for. I know there's some recruiters out there who help people find GTM engineers and like maybe can help you vet them better. So, that's that's what I'd probably recommend if you're looking for somebody like that. No, totally. I think that it's it's to your point. It's like this person is they're I like what you described. It's like it's better for them to come from the sales rep side of the spectrum and then move into this um rather than the engineering side and move into it, you know, from that that side. Um it can be possible and like depending on like the complicate, you know, what you're doing like I met Absolutely. Absolutely.

1:00:15 But I think for the general public when they think about like what is the actual like implementation of this hire look like in the company it's to your point it's like it's in service of the sales team right like it is not uh you know you're this person drives revenue for the company and has like directly has a lever to pull that I I think what you said is really interesting too on the dashboard side like this isn't like a person that is um you know you're not reporting on the activities you are initiating the activities and then reporting on them right it's like you're not an analyst This isn't an analist role.

1:01:02 This is like somebody that goes and actually does things. So yeah, absolutely. And they should hold a bag. I think GTM engineers should be responsible for a quota of some kind. Whatever that is, SQL's pipeline. That's great. Whatever. I think that's great. And be measured to that and maybe have bonus structures. That's associated like it this is a sales hire, right? And is it part of the sales organization as a function? Um I'm I'm curious your thoughts on this. I talked to I had another person on the podcast yesterday. Um they'll be like right before your episode comes out.

1:01:43 Uh he cold email guy talking about um he's like I'm starting to think about cold email now is it like onetoone marketing at scale. Um and I I really found this like a very interesting and unique idea of like oh you can now research people so deeply and like have like an understanding of the individual like at such a granular level that like and for example the example he used was where uh they're cold emailing these uh different agency businesses and maybe they have different specialties like cold calling or cold emailing or um you know LinkedIn DMs etc.

1:02:25 And so from that they can basically like craft this like pitch, you know, because the product, you know, it serves all of those individuals, but rather than a generic, it can be specific to them, you know, based off of their website, based off of their LinkedIn post history, you know, them talking about the business, etc. And then from that, it's like an email. He's like, I don't even care if they respond.

1:02:55 I just want them to see this. And like we know that that impression like is going to be a high value impression because it's like a onetoone like you know imp basically it's a one to one ad is like how he was thinking about it which I thought was really interest. I'm curious your thoughts there and like have you seen you know any experimentation with that or any you know any data that you're sitting on you know anecdotally that that feels related to that idea. Yeah. Um so first thing um I do also view cold email as a onetoone ads channel. Um, I think that's the best way to think about it.

1:03:41 Obviously, we want people to respond. We want people to say, "Send me the booking link." That's not But here's the thing about email, and I think this probably is obvious when you really think about it, but I I don't think I think um the way that like cold email has been treated is like we used to think of it as a marketing channel, and then to some people, maybe it's just me, I don't know, we then started to think of it as like a sales channel where we're like, "Oh, if they don't respond to the email and book a meeting, then it's not working." And now after working at AirOps and being exposed to like a more traditional marketing organization, I've sort of like shifted back to thinking about it as a traditional marketing channel.

1:04:37 Um because like I do have data, hard data that showed like uh heavy influence on some of the larger deals that we booked. um where I had this like uh it's in my uh portfolio if anybody goes to see it, but like um this dashboard that showed like all of the deals that we booked and like it would manually allow me to attribute it to outbound or not based on like do I think that we actually had some effect? Did they reply to an email? Uh did we email the person who boo came inbound? Things like that.

1:05:18 Um and so there was an influence whether or not they responded to it. And the other thing is like think about like this like what hap like this happens to me all the time and I do this right like somebody will send me an email it's not really relevant for me but I'm like oh you know what the the other guy like in revops or the SDR leader might be interested in this I forwarded to them like there were gong transcripts speaking gong transcripts this was actually useful for me um in this deal attribution like uh app that I built I would have it analyze the gong call to see if they said how they heard about us and I would use that case it's amazing

1:06:11 yeah I would use it to attribute deals to outbound because Like there was a couple of examples where somebody was like, "Oh, well my boss forwarded me this email and told me to like check you guys out." So nobody responded to anything. They [ __ ] came in through a branded search, but like they told us on the call and they probably didn't even end up like we have a how did you hear about us form on like our demo booking. They would never [ __ ] say email. So I don't know why people are so reticent to just tell us how they heard about them. Um but then they'd get on a call and they' be like my boss forwarded me or called email.

1:07:03 So, I can't figure it out. You know, attribution is not a solved problem. So, it's something everybody else deals with. So, yeah, that's that's my spiel on that. No, man. I I I stand so firmly behind this.

1:07:18 We I did this test forever ago. I I said this in the last pod, so sorry if you're hearing this again, but we basically like cold emailed a bunch of people with no intention of responses, and then we just looked at the emails to see if they went and signed up for the product. and we saw we saw a conversion rate that was comparable to like what we saw with like ad traffic from like you know ad traffic to sign up and I was just over there and I was thinking to myself I'm like this is just literally a marketing channel if you think about it like this and so it's like how cheaply can I send it scale it's like and actually get it to land actually get it to be open so

1:08:14 dude think about the cost per click versus the cost of somebody reading your email totally like it's cold email is basically [ __ ] free like when You think about it like that. Yep. And so then it's like how do I stand out? And that comes to the copyrightiting piece. But all right, brother. We're coming up on time. Thank you so much for coming on today. I co-sign that you should go sign uh go hire Ben if you need a GTM engineer. He's exceptional at what he does. I respect your work deeply. Um also include link to the portfolio below and anything else that you want. But man, if people want to reach out to you, uh where where should they where should they find you?

1:09:06 Awesome. So I have a website. It's benignamine. I ben y a m i n.ai. You can check me out there. It's got links to my LinkedIn, my uh Twitter account, my YouTube, um all my other podcast appearances and stuff. So, a bunch of like long form articles if you guys want to go pick my brain. I wrote all these articles because people would always ask me things and I got [ __ ] sick of like answering their questions. So, I'm just going to write an article about this. So, anyway, that's where you should go check me out. Amazing. Awesome. We'll include that all in the show notes. Thanks for coming on, man. We'll have you back in like six months and tell us a state of the union, what's going on in GTM engineering and what you're seeing on the front lines.

1:10:05 So, dude, we're all going to be unemployed by then. But yeah, we're going to have a Chinese model ship and it's just going to it's going to replace every person. So, you talk about our woodworking projects that we're doing now that we've been replaced by the robots. Exactly. Exactly. I've just got into gardening. I'm happy to just like continue only doing that if this uh my job basically disappears. So, all right, brother. Yeah, thanks for coming on. We'll talk to you soon. Yep.

Cody Schneider asks why Salesforce is suddenly the CRM Ben wants. The Salesforce CLI plugged into Claude Code turns the old endpoint maze into something an agent can actually drive. Buying software now starts with a question: can I do in the API what I can do in the UI? If you need a UI, vibe-code one. On Zapier, Make, and n8n, Ben's story is that he used to drop model-generated JSON into n8n and realized he was already writing code. Skip the middle. Put the work in GitHub so the company still has it when you leave. His portfolio in the chat is a pile of mermaid charts of old repos. AirOps SDRs and less-technical marketers were committing into the growth repo. He mined thousands of Gong calls there and got less than people hope. Reps run the talk track that closes. Loops only help if you already know which enrichments matter.

Notes

  1. API completeness is the buying test. If the UI can do it and the API cannot, do not buy the tool.

  2. Skip Zapier and n8n. If you are already generating the JSON, push code.

  3. A GitHub repo is the portfolio and the handoff. He was showing mermaid charts of old work in the chat while they talked.

  4. SDRs at AirOps were committing into that repo. Less-technical marketers got put on GitHub so other people could help.

  5. Thousands of Gong calls produced less signal than the folklore says. Reps repeat the talk track that already closes.

Watch on YouTube

Get new episodes in your inbox

Let's talk shop.

We try to give away as much GTM alpha as we can on a first call. Tell us what you're thinking about, we'd love to help.