GTM Engineering: The Technical Bits — Everett Berry, Clay

AI Engineer · 19 min · 183 sentences · from YouTube's caption track

Each timecode opens YouTube at the start of that sentence. Line anchors (#s42) are the cue ids in the WebVTT, and every line carries its start and end seconds. All transcripts has every talk, and the whole corpus as one file.

00:01[music]
00:12Okay, good morning everyone.
00:13I'm stoked to kick things off this morning.
00:17Um, today's talk is going to start off with just a brief look at GTM engineering and what it is, and then I'm going to get into kind of the areas that I consider to be the most technically interesting and challenging
00:29um, within this space.
00:30So, GTM is pretty new.
00:33Um, it's arisen out of a couple factors, but one of the main motivating ones is that GTM teams have kind of realized that it is now possible to ship
00:42as fast as a product and engineering team.
00:45And so, the best GTM teams that I work with are generally um, pushing changes to their GTM structure almost at the same cadence that an engineering team might be doing releases.
00:57And so, at Clari what that looks like is something like this, where every 2 weeks we're pushing new data to our teams, we're pushing new automations, we're of course running new campaigns,
01:07and we're constantly iterating on the things that we're shipping and trying to keep pace with the speed that is um, that our engineering team is working at.
01:15So, um, to do this you need a couple fundamentals in place, and in my view GTM engineering at its heart is really about removing the constraints that have historically stopped GTM teams
01:27from shipping at speed using technology.
01:30Um, and the role itself has exploded.
01:33So, um, I would venture to say that most advanced GTM teams are now hiring GTM engineers or looking for this role, and in my opinion it's one of the first roles that actually is an index on the advances that we're making in AI.
01:48And so, as models have become more powerful, GTM engineers have gained more leverage within their organization and become more valuable, and we've seen tremendous growth in this role.
01:58So, um I kind of break it down into four areas.
02:02And what I hope to do with this talk is kind of lightning round go through some of the technical problems and solutions that I see for this.
02:08If you're not familiar with Clay, we are obviously building infrastructure that kind of attacks these areas.
02:14But this is not meant to be a sales pitch.
02:16It's more meant to be if you're an engineer or GTM engineer that's working on these things, some ways to think about how to structure it.
02:23And then if you're a founder in this space, some of the interesting problems that I think are worth uh tackling.
02:30So, the first is data.
02:32And more than most teams within a company, data is the lifeblood of GTM.
02:37And the core goal that I think we're trying to accomplish with our data is to create a perfect virtual copy of the market, the ideal customers, the accounts and contacts that you're going after.
02:49So, um most teams start out with something that looks like this.
02:53You have accounts, which are the companies that you're targeting, and contacts, which are the people that you're going after.
02:59The problem though that makes this challenging is that accounts at least exist in a state of constant change.
03:05As an organization, a company might be getting acquired, it might be spinning up new offices, it might be launching new products.
03:13So, the company itself is always changing.
03:15In addition, you as the GTM team are doing things that company that is making it change as well.
03:20So, you're marketing at them, you're selling towards them, you're trying to book meetings.
03:24That's changing the state of the account.
03:26And then also, the company itself is hiring people and firing people and doing things that provide signals for you.
03:34So, as GTM engineers, we need to manage all of this state within our accounts.
03:39And um we don't have all the data that we need to do this kind of right off the bat.
03:44So, um this is what a an example from Clay's CRM looks like.
03:48You can see I have a ton of fields that are simply telling me what state this account is in, whether they're a customer or not, whether they're expanding or churning,
03:56how big they are, how well they're scoring.
03:58And so, one of the fundamental things that we need to do is make sure that our records are updated so that we can actively accurately action on what's happening here.
04:08And so, as a you know, in a brand new CRM or a or a CRM that I'm going into, the first thing that I am doing is actually filling up the account with relevant contacts.
04:17And then I'm layering in third-party information that is going to help me figure out which accounts to target and which ones are in in market.
04:24Primarily, I'm looking at the account hierarchy, the firmographics that uh describe how big the company is and so forth, the technographics that describe what technology they're using, and then I'm trying to create signals on top of those accounts
04:37that are going to allow me to go after that um that company.
04:42And so, because uh there's a lot of third-party data involved here, we need to actually go out and source and procure that data.
04:49And within GTM, there is literally hundreds of vendors that you can turn to to um to get the data that you need, but none of those vendors is going to have a complete picture of all the information you you desire.
05:01And so, the key technique here is called waterfalling.
05:04This is where I'm going to actually go and look into multiple providers to try to fill in all the information that I need.
05:10If you see here, if I was just using Forager to get phone numbers for this um this set of countries, I'd only get halfway there.
05:17So, instead, what I need to do is layer on all of these other providers.
05:21And that's not only true for phone numbers, but for most the other data points that we care about within GTM.
05:26And so, either you or the vendor that you're using needs to run evals against these data providers in order to obtain the most accurate information.
05:35And so, not only do I need to sync in third-party information to my data layer, I also need to incorporate first-party information, and I need to keep that constantly up to date.
05:44It's also incredibly expensive to update data all the time, especially if you're purchasing it.
05:49So, I can't just update all the fields.
05:50I need to kind of selectively choose which fields to update.
05:53And not only am I pulling information from various sources, but if I'm running a signals program, information that I need is getting pushed to me all the time as well.
06:01And because I'm using multiple third-party sources, the actual representation of a single account in those different sources is going to be different.
06:09So, I need to resolve the entities um in between them.
06:12And so, a great data layer will tackle all of these things and allow me to kind of move on to the more interesting work, but without this, it's really, really hard to execute automated GTM plays.
06:24So, the next piece that I need to get right is orchestration.
06:27And this occurs because within go-to-market, there are literally dozens of tools that most teams are using.
06:33So, I might have a sequencer or a dialer or a bunch of places where my reps are living or my field marketing is living.
06:40And in most cases, the view of the world that those tools have is different depending on where you look.
06:46And so, orchestration, in my view, is really the act of keeping all of that up to date.
06:51And so, what happens to our data model then is I'm um on top of my enriched information, I'm sending emails and calling into my contacts.
07:00The account is generating events that I need to keep track of.
07:04And I now need a system to actually plug all of this into my data layer.
07:08And so, most teams are going to have something that looks like this as their basic stack: CRM, data warehouse, sequencer, a dialer, a note taker for call recording, and some chat interface like Slack.
07:20However, um I usually see like 10 or 20 or 30 tools that actually these teams are interfacing with.
07:27And orchestration needs to kind of keep that information up to date.
07:31The problem with orchestration is um all of these different systems have different data needs.
07:35So, some systems need kind of like real-time updates one record at a time.
07:40Other systems are going to need hundreds of thousands of records, maybe updated once a day.
07:44I might need to schedule updates on a monthly or weekly basis depending on the data I'm using.
07:49Some data points like employee count change all the time.
07:52Other data points like headquarters location change very rarely.
07:55Um and I'm not going to be making updates to all the fields at the same time.
07:59So I need logic within these systems that helps with that.
08:02I also need to be able to take a single um system that I'm working with and actually fan its information out to multiple different systems.
08:10And because I end up with this kind of distributed setup, I also have failures that are happening all the time.
08:15So this turns into a fairly complex data engineering problem that we need to resolve.
08:21A classic example of this is um when I'm working with these different systems, usually they are not fully orchestrated, which means that one system is talking to each other while I'm trying to talk to both of those systems at the same time.
08:33So for example, if you have a salesforce uh connected to outreach or a sequencer, usually that CRM that seek sequencer are syncing independently of your orchestration system.
08:43And so if you create contacts in your CRM, you actually need to wait for that contact to sync to the sequencer before you can then take action on it.
08:50This creates some difficult problems where you actually need to introduce things like weights and loops to check if information is ready.
08:58So Clay, we've iterated on this problem um quite a bit and we've ended up in a place where we are basically taking a graph-based view of the orchestration problem
09:08with a series of general purpose nodes that are executing various things.
09:12And so we have nodes that run agents, nodes that make tool calls, nodes that handle our conditional logic, nodes that run code, and then nodes that run effectively this map reduce system to fan out the information and bring it back.
09:25And so a great orchestration layer will handle all of these and whether you buy it or build it, I think this is like fundamentally the the kind of modern way to set this up.
09:36Here's an example of this um in in clay.
09:38And again, you can use other tools for this, but basically I'm going to have some sort of event that kicks off my orchestration, some trigger or some schedule.
09:46I'm going to then talk to a couple different systems.
09:49I'm going to combine that information back together, and then I'm going to push it out to different interfaces that, uh, my reps are using.
09:57Okay, so now that I have a data layer set up and I have my everything orchestrated, I now have this system which is starting to look a lot more complex.
10:07And if you spend a lot of time in go-to-market, uh, I think this is actually like one of the simplest views of what is happening with an account, where I have multiple signals that are occurring, I have data that needs to be updated, I have actions that my agents or reps are taking,
10:20and then I have meetings that are happening and feedback that's occurring.
10:24And so, um, the state of the world today is that we are relying on sales reps in a lot of cases to manually sort through this, but one of the great advances
10:32in the last year or so is that we can now use agents to take care of all of this context.
10:38Um, however, if you use agents, you actually run into some of the same problems that you are dealing with if you just use LLMs.
10:45So, um, in GTM in particular, we are trying to build very long-running agents, agents that run over a course of weeks or months that keep track of the state of an account throughout a deal cycle.
10:55We also have a high bar for error because a lot of the results of our GTM work is customer communication and getting that wrong can have disastrous con- uh, consequences.
11:06And then finally, the agents are often doing unstructured work and pushing that into systems that are highly structured like a CRM.
11:13And so, the the mapping of what the agent is producing is is super important.
11:19So, the architecture that we've landed on for this that I think is most powerful is an agent that exists for each account and maintains a persistent state of that account.
11:28It's always going to execute, and because it's executing over a of weeks or months, it's often going to be dormant for most of the time that it's available.
11:37So, we need to use smart triggers or a heartbeat or something to wake it up.
11:42And then when it wakes up, it needs to kind of ingest the current context of the account from our data layer and our orchestration layer.
11:48And because the agent is making decisions for us in many cases automatically, we need to allow for feedback on the agent as well.
11:55The um cutting edge of doing this is the learning phase where as the agent works on an account or a series of accounts, it updates its own view of what's working.
12:06Today in GTM, this is not fully solved yet.
12:09And in fact, the continual learning um effort and the next best action suggestions are kind of one of the cutting edge problems that that we're working on.
12:18So, here's a look at um a a kind of a basic agent.
12:22You can see here that I'm pulling in information from a couple different sources.
12:26I'm using reasoning steps here.
12:28And then critically, I'm also updating different values in my CRM that are just for the agents.
12:33So, I always recommend separating the fields that agents are updating from the fields that deterministic systems are updating or that people are updating.
12:41And to look at a full uh view of an agent, this is a closed loss free awaken agent.
12:46So, you can see it's talking to Gong, email, CRM, and um my data warehouse.
12:52And so, it's pulling together multiple pieces of information.
12:55And this agent is triggered on a time basis.
12:57So, if if we lose an account, we're not going to kind of like immediately go after that account again.
13:03We're going to wait for a little bit of time in order to um in order to attack it again.
13:07So, there's a number of kind of timing and context issues that we have to address when we're building agents for GTM.
13:15The last step um that that is important here is the actual execution step.
13:19So, now that we have a data layer that contains the perfect copy of the virtual world, we have an orchestrated system that's um sharing context with all of our systems.
13:28We have agents that are making decisions and reasoning about what to do.
13:31We actually need to get in front of customers and execute our messaging.
13:36Uh unfortunately, uh messaging and execution is one of the hardest problems in GTM engineering.
13:41Here's a look at some of the um email open and reply rates over time.
13:47Generally, what we observe is that cold email works less and less well as the years go on.
13:52This is a trend that's been true forever.
13:54Look at a snapshot of this.
13:56Um if you look at the far left here, I actually think these are pretty uh elevated rates for some of these, but the relative differences between these channels is correct.
14:05So, you know, LinkedIn can be three to four times more effective than cold email.
14:09Cold calling and cold email are roughly the same.
14:12And then on the right here, um this is a series of uh statistics from Smart Lead.
14:18This is across, I think, 20 million emails.
14:20And so, you can see we've got somewhere between a half a percent and 1% reply rates.
14:25So, what that means is, of course, if we've got 100 contacts that we're sequencing, maybe one of them will reply.
14:30And so, that really raises the stakes for agentic execution within GTM because if your agents are doing the wrong things, then you're missing out on the margin, which is where most GTM teams
14:41are uh are having success.
14:45The other thing with execution is you have to solve some very human problems.
14:49So, for example, do you email on behalf of the rep, or do you let the agent do the emailing?
14:55If you email on behalf of the rep, what that means is like everett@clay.com is actually reaching out directly to customers.
15:01But, if I do that in the wrong way, or I don't get the replies that I need, that then means that my overall domain reputation is going to suffer for for my company.
15:11So, a common technique then is to use multiple domains to get in touch with customers.
15:15But, then if I do that, I actually need to find a way to route responses on that domain back to my main domain so my reps can process it.
15:22And that's just email.
15:23There's also multi-channel outreach, which is tricky as well because you then have to, for example, if you get a call connection and a meeting booked on your call sequence,
15:32you then need to suppress your email sequence and maybe unenrolled someone from a life cycle marketing campaign.
15:39So, the coordination of the execution of all of this is also a hard problem and also something that we can use agents to to help resolve.
15:50So, here's a look at one way that that we're tackling this.
15:53This is a setup for a kind of rep proxied view where we have a bunch of rep inboxes that are connected to our sequencer and we're actually sending that on behalf of reps.
16:03But like I said, we do this for kind of a portion of our accounts.
16:07For many of of our accounts, we actually use multiple domains to go after them and then we have to tackle the routing problem.
16:15So, that was kind of a lightning review of what I consider to be some of the like harder problems within GTM engineering.
16:22I think if you get these pieces right, you can end up in a place where you're providing a technical foundation for growth that is helping your company achieve incredible results.
16:31But I think a lot of teams actually overlook some of the harder pieces here and and don't necessarily design around some of the constraints that they have to deal with.
16:42So, that's my talk.
16:43You can see me on LinkedIn.
16:45There's a URL there and looks like I have about 90 seconds for questions if anyone wants to ask anything.
16:55[applause]
16:58Any any questions for Everett?
17:02Raise your hand.
17:02I can Okay.
17:09Just uh curious as to what your um biggest challenge is with all these new um, within the org.
17:18I think one of the harder problems is probably the interface between the human and the agent.
17:24Like I said, I think the most powerful use of agents within GTM is to act as the reasoning and decision layer for a lot of tasks that a sales rep was previously doing.
17:34Um, and so you run into a lot of instances where the rep might think that they should do something different or the rep might not know that the agent did something.
17:42So, um, because ultimately a human still needs to get on a call with a prospect, coordinating the, uh, connection between what the automated systems are doing and what the, um, what the human sales rep is doing and making that work well, I think is probably one of the hardest problems.
18:00Okay, well, oh, yeah, one more.
18:08Um, I have a couple questions.
18:09So, um, GTM engineers are fundamentally software developers that have this, uh, knowledge.
18:15And, well, my other question is that, uh, this is mainly for outbound or inbound as well is included in this product.
18:21For everything, yeah.
18:22So, um, like the orchestration problem is pretty acute in inbound.
18:26You have to get the routing right.
18:28You have to qualify the account properly.
18:30There's usually historical context on inbound that comes in that needs to be, um, needs to be understood.
18:35So, uh, yeah, this I think GTM engineering covers, uh, covers all of GTM.
18:40And, uh, if you want to chat with me more about this, I'll just be right outside, but I will hand it to the next speaker now.
18:45Thank you.