E6: "Clients don't pay for crazy, they pay for effective" | Patrick Spychalski

Okay, what's up, GTM builders? It's Jared and Mateo. And welcome to GTM Engineer School, the podcast where we grill the sharpest GTM engineers for the principles, the workflows and the tools that are shaping this discipline. Today we're going to be hanging out with Patrick Spychalski, co founder of the Kiln, which is a skunk works collective of GTM experts, data scientists and lots of former Clay employees that turns messy revops data into revenue engines for SaaS teams. Patrick is a bonafide clay OG. You'll catch him dropping marathon length clay table teardowns on LinkedIn. So if you're a workflow nerd, smash that follow button. Definitely smash that follow button. Before going full on into his agency, Patrick's Bank. Two years working at Clay running their early partnership program. And besides that, he's also founded and still runs Unique Market, which is a web store and showcase of men's clothing, shoes and accessories focused on vintage designer and avant garde fashion, which is really, really cool. So Patrick's not just about crazy plays and workflows, but also taste and style. We're pumped to unpack his playbooks and unique approach to go to Market engineering. Patrick, welcome to the pod.

Appreciate you guys having me on Pumped to do it.

It's going to be fun. Patrick, I'd like to start just defining what is this? What is a go to market engineer? Can you tell us in your own words?

Yeah, I mean, I feel like everybody has their own definition at this point. In my mind, a Go to Market engineer is somebody who has both like highly technical ability and the tools required to run go to market systems, specifically automated go to market systems, as well as the like, general strategy and intuition of somebody who would be in a go to market leadership position, or at.

The very least like a more senior go to market role in the company.

So somebody, in short, who knows all the tools associated with the GTM engineer world, but then also knows how to.

Use them to actually get results for people or companies.

It's all about the results. Show me the money. Patrick, do you remember life before go to Market Engineering? What was that like and what's changed?

Yeah, what's funny is I actually wasn't in sales prior to Clay existing and so I actually got into sales at the same time as discovering Clay. I mean, I started working at Clay very early on and I was actually running a, like an organic marketing, you know, kind of freelance gig prior to that. So I started using Clay to go get myself more contract gigs in the early days. And it was funny because I would, I would talk to people who had been in sales prior to clay and go to market, engineering, existing, and it seemed horrible. I mean, I can't really imagine having to go and individually prospect and research.

And reach out to people.

I mean, like, I've actually never had to do that. So I feel thankful that, you know, I entered the space at the time I did.

But yeah, it's not, I would say.

A situation I'm very familiar with, actually.

Really funny.

Really funny.

All right, Patrick, thank you for that. And now we're going to break into the Lightning Round. Matteo is going to lead you through a bunch of quick questions.

So, yeah, very similar to, you know, like the Lenny's podcast Lightning round, if you've heard of that. Simple, easy questions, simple answers, and maybe just add some color as to the why. But let's start with your favorite CRM.

HubSpot, Salesforce or currently HubSpot. I think Attio will eventually beat it.

As it develops as a, as a company.

But currently HubSpot because I think it's.

A combination of like enterprise grade support while also being good at integrating. But again, I think Adio will eventually beat it.

Very interesting bet. Top LLM Claude or ChatGPT?

I like Claude specifically for what I.

Use it for, which is often writing things. I think it has a better like quote unquote voice or tonality relative to ChatGPT.

So I also just like Anthropic's branding a lot more.

So kind of partial a fan.

Likewise, Top enrichment tool.

It's gotta be Clay.

I mean.

Is that, does that count as an answer? Can I say clay? Because I feel like it's like.

Does it feel like cheating or for.

Some reason it kind of feels like cheating? It's just because it's like so it's like, it's obviously in my opinion the best one. Like, I don't think there's a close second. So like there could be individual data points maybe where you could pick an enrichment tool. But like, I mean, it's just an aggregate of every enrichment tool. Like fundamentally that's what clay is.

So it's kind of just. It feels like cheating almost.

Patrick, what. What's your favorite API or Enrichment Within Clay?

Oh, good question. I really like their HDInsights integration. It's expensive, like it's 8 credits per run, but you actually couldn't use HDInsights.

Otherwise if you didn't use it in clay.

It's not accessible unless you had an enterprise plan with Them and it's super.

Powerful if you have the right use case for it.

That's technographics, right?

Yeah, yep.

Tech stack.

And what's your overall technical top GTM engineering tool?

Yeah, probably clay, second place. Probably like N8N.

So I'd say those are probably the top two by far for me.

And what's the most underrated GTM engineering tool in your opinion?

Probably Exo AI. I think it's an incredible way to.

Source really niche leads.

It's not, doesn't require like a lot of engineering per se. Like I think anybody, not just a.

Gtme, can use it, but it's, it's still, in my opinion, in the ecosystem and super powerful.

How do you spell it again?

Exa.

AI. Yeah.

In short, you use natural language to source leads. Like you can give it a very.

Specific prompt on what specifically what lead you want to source and it will just find them for you automatically. Super cool.

Thanks so much for sharing. That was fun. All right, let's switch gears and let's talk about a bit more deeper stuff. Like when you're building new tech stacks for your clients to set up these go to market motions, how do you approach designing them? And can you talk us about like the building blocks and the underlying principles?

Yeah, totally. So obviously we assign different tech stacks.

Or like a combination of different tools to every client that we work with.

But I think the best way to approach it initially is just figuring out what the best tool for any given.

Use case would be.

And it very quickly allows you to like assemble the avengers for a specific client or not. So for example, maybe some clients have massive sales teams of people like you.

Know, maybe 20, 30 SDRs or BDRs.

And they want to do like individual.

One to one outreach instead of doing.

Like the kind of automated approach, then you would probably want to recommend them.

Like an outreach or a sales loft instead of a smart leader. And instantly.

And so there are cases where like.

It'S very obvious what specific tools you need to use for a given use case. They're like, hey, like we want to.

Be able to connect this to clay.

We want people to kind of do.

Gut checks of every, every email we send out and we want people to.

Be able to make changes. Like that's, that's very obviously a more sales engagement sort of tool than an automated one.

So I think it's, it's worth finding.

The best in class for every tool.

Of course that takes a fair amount.

Of time to do and really can only be solved via experimentation.

But after you have the like, long list of best in class tools out there. I think from there it's pretty easy.

To assemble a list of them.

So that from there there's. There's a couple situations in which it might take more nuance to figure out.

What tool to use.

So for example, like, should I use clay or N8N to build this workflow? Like that could be actually a difficult.

Question to answer in some cases.

But like, for example, I would say if you're running like an enrichment workflow.

Clay would be an obvious answer. But if you're doing more of like an AE based, like I want to get a response and like draft a response to that response to this email, like, maybe NADM would be a better approach because it's is better at encoding AI agents for.

So it's tough to answer. I mean, I think it just requires.

Like a fundamental understanding of how these tools work and then you can then kind of piece together, you know, the web of tools a company should use.

I love your approach and especially, you know, like that you start from, you know, like the what's the best tool that you can shop for? Depending on a use case, but obviously just a double click on that. You know, it's so hard to a know you know, what's best because there's new tools coming up probably every week or every month. But also how do you really pinpoint what's the specific use case that you need to solve for and work back the best tool for it?

Yeah, totally. And I would also just say, like, if you're not deep in the weeds as a GTM engineer, I would just.

Go ask somebody who probably knows better.

Like it's, it's way easier than going.

Trying out every tool yourself. There's probably somebody who can answer that objectively for you. Like, have people ask me constantly just like, what tool should I use for this use case? And I'm like, x tool. And then it probably solved like it probably saved them a fair amount of hours of like, you know, toying around to do that.

That's pretty much what we're doing with this podcast anyway.

Yeah.

So thanks for being so transparent. And let's say you, like, if you could only keep like three tools in your GTM stack, what would they be and why?

Great question. And I'm going to sound like a broken record. I think Clay, N8N and Lovable, I would probably pick. I think Clay is incredible for enrichment. N8N can build essentially any workflow, no code.

And even if you need to code A little bit. You can incorporate that into your workflow.

And Lovable is a great way to.

Visualize these workflows for people to actually use. Like you can, if you can connect.

The workflows you're building on the backend.

To like a lovable, Lovable built software or dashboard or tool, it can allow you to pretty much make anything.

So I think with those three, I honestly could probably do 95% of of.

The work that we do today if necessary.

So you use Lovable for like internal dashboards and tell me more about that.

Yeah, so Lovable for those who don't.

Know what it is, it's a vibe coding tool that allows you to build.

Websites or dashboards or apps or SaaS.

Products using natural language.

And so I think it's most useful for in the GTM engineering context is it allows you to simplify workflows or.

Create an interface for workflows so that people that you're building these workflows for don't have to actually go into the underlying tools.

So for example, you could build a dashboard in Lovable and connected to Smart Lead via their API that allows you.

To, you know, visualize the results of all of your campaigns.

Or you could create an enrichment table in Clay and instead of having somebody have to go into Clay and put a company into the Clay table to run the enrichment, you could have a SaaS product or even just like a.

Small website on top of it that.

Allows sales team members who don't, who aren't familiar with clay to interface with the clay workflow.

And same with N8N. Like if you can connect, you can.

Connect lovable and N8N to create pretty powerful software tools very easily. If you can create the workflow on the back end and then connect the.

Two via just a web hook or an API call, then you essentially have a, like a bonafide SAS tool.

So just webhook, no MCP server needed.

Depends on what you're building. If it's just if like the smart lead example I just talked about, you.

Would need an MCP server to do that.

But in like the Clay example I gave you, I just built a lovable project recently that connected back to a.

Clay table and all that was required.

Was a web hook that went from Lovable to Clay to feed Clay the data that it needed and then a like a get API call from the.

Software or like a post API call from Clay to send the data back to Lovable. So it didn't require any sort of MCP server.

But yeah, Again, it depends on how.

Complex the thing is that you're interfacing with.

Yeah. Yeah, great. Patrick, would you, next, would you walk us through either your Favorite, go to MarketPlay you've built and I may have an idea which one that is, or the most common play play that you build when you first start working with a new client.

Yeah, so the most common one I would say is probably there's two and.

I'll probably talk through both of them because I think it's actually like a dead tie.

The first one is CRM enrichment and CRM data cleaning plays. The reason this is often the first thing we do for a client is because they might have this laundry list.

Of projects they want us to work on.

But ultimately all of them usually route back to a source of truth of some kind. And if their source of truth data isn't good, then the projects we're going to work on, you know, after that would be essentially useless. Like for example, if you were to build like an inbound lead enrichment thing that routes all your like, scores all your inbound leads and sends it back to your CRM, it's useless. Like because half the leads in your CRM might be scored, the other half might be completely outdated. And then you have this like brick that doesn't actually work. So we do a lot of like initial like CRM cleaning and CRM enrichment.

Builds for, for clients in the beginning.

Because I actually in some cases like require we do that in order to.

Build other things to make them more useful.

The second one that we've seen a lot are inbound lead aggregation and scoring builds. Like a lot of clients of ours just have a bunch of different marketing initiatives, kind of disparate marketing initiatives running concurrently and they don't have a way to action upon and enrich and filter through these leads without doing so manually. Like usually there's somebody on their team that's just patchworking together a bunch of.

Different spreadsheets from different initiatives and trying to find the people worth reaching out to.

So aggregating those all into kind of.

A singular clay table or clay workbook.

And then scoring them, enriching them, and then sending them back to the CRM.

Has been a pretty big value add.

And like a very low hanging fruit.

For most clients that we work with.

To answer the favorite workflow question, I don't know what your hunch was, but.

It was probably right.

The lovable table I built recently where you could create custom software or custom apps in clay at scale was probably.

In my opinion, like the most powerful.

Thing I've built and there's not a really great use case for that specific table. But I think the general idea that tables like that can be built and.

Can exist is really the powerful thing.

Like the fact that you can build custom software for a prospect or even.

Just for list of companies programmatically at.

Scale with the right prompting is a very powerful idea. And I also kind of like it because lovable gave me their internal API to do it. Like, you can't actually build that table. Like nobody other than me and the lovable team could probably build that right now. And so there was some degree of like pretentiousness, you know, like pretense associated.

With that table as well.

That was such an amazing project to read about and yes, that was my hunch. So congratulations on that. What a great showcase.

Appreciate it.

What about like, Patrick, can we double click on this and just like share Also, what's the kind of like credit consumption or the costs for running these things at scale? Just like to give our audience like a bit of a hunch of or perspective on this.

Yeah. So skilled clay tables can cost a.

Lot of money if you.

Well, they can cost a lot of money, full stop. And it can cost even more money if you don't, you know, set them up correctly. And so we've had clients that have enriched millions of records from their salesforce.

In a series of clay tables.

And like, you know, you can imagine, you know, 5, 10 different data points that they're enriching for already.

It's adding up to the millions and millions of credits.

And so those projects, if you just use clay credits, oftentimes can cost upwards of 20 grand, 30 grand for an.

Enrichment project, sometimes more. I mean, I've seen quotes as high as 60 or 70 for clients of ours before.

But it's also worth noting that API.

Keys can vastly, vastly reduce the cost of the majority of these things.

And sometimes it requires a bit of creativity. Like, for example, you might have an enrichment in clay that costs three credits, but you could, and you could technically get an API key for that, for that software. But let's say hypothetically, you can't get.

An API key for it.

Then it could be worth trying to figure out if Clay Gent can find.

The same data if you prompt it.

Correctly, buy an OpenAI API key and.

Then use that to reduce the cost.

So you can vastly reduce the cost.

Of these workflows if you want to.

I mean, one time we had a client that was quoted at like, yeah, we, we Once I had a client that was like about to spend 60 or 70k on an enrichment and we.

Just found a bunch of API keys reduced the cost to like, I think it was like 7 to 10k.

So like you can really save a lot of money by buying API keys for these projects, but it can get hefty regardless. I mean 7 to 10k for an enrichment is still not cheap. I would say it's worth it for the right company though.

And what you're saying is half of being a go to market engineer is being a, is being a Clay credit engineer?

Yeah, I mean you can literally if your clients planning on using Clay, regardless of whether they hire you or not, you can make the money they're paying you back just in recommending a specific set of API keys. Like the client in question just saved 60 grand. Ish. Like just from our, our consultant consultancy. Like I mean like that in itself paid for us for a little while.

Right.

So like, you know, I think being a great GTM engineer also involves being able to budget really well and, and also find API keys that the client otherwise wouldn't have found. Like for example, we have an API key that can replace Clay's LinkedIn enrichment that is a fraction, like a very small fraction of the price and also is less expensive. And so like if I can just use that API to replace all the.

Functions, I would have used other credit Clay.

Credit Clay credits for otherwise, I mean.

It'S, it saves a fair amount of money.

Can you say what that is?

I don't know if I want to drop it on them just yet. I'm gatekeeping this because they're going to raise their prices. Yeah, maybe I'll drop it one day. I don't think I can do it.

It's so good.

It's too good. I'll say. Actually I will say the competitor they have though, which is also good crust.

Data, which is a little more well known.

But crust data is a pretty good.

LinkedIn enrichment as well. But they have pretty high contracts so if you're going to use them, use it. Like I would definitely recommend doing it for a larger enrichment.

But CrustDate is another great like LinkedIn enrichment tool and it's also a live scraper. So at least to my knowledge. Clay right now their LinkedIn enrichments are using static lists that they're scraping like periodically. But Cross Data does a live scrape.

So when you make an API call it actually goes and finds the LinkedIn data as it is increasing the accuracy. So I definitely recommend checking it out.

Bring it on.

Skills and trends.

Skills and trends. Yes.

Pick your brain about what, what's coming.

Up, what are new skills that you think are exactly like emerging engineer GTM engineers like should learn more about.

So the first one would probably be improving your N8N abilities if you haven't already.

I think N8N is a super powerful workflow tool. I think it will replace a lot.

Of functions of clay if not just add to your arsenal.

Like there are some things like N8N can do that clay can't. And I really think it is the.

Best workflow tool to learn. Would also recommend checking out like building MCP servers, like doing research on building MCP servers.

So like watching YouTube videos on how to do so or even trying to build a few of your own with basic APIs like ones that you, you.

Know only have a few different functions. Another one would just be checking out.

The vibe coding tools that are available now.

So there's kind of a spectrum like lovable, super easy to use but then like cursor is really hard to use and then for engineers. So I'd recommend checking out the gauntlet and figuring out which one you have the most say like intuitive familiarity with.

What are some of the things that you can that you Normally do in any 10 that you're not able to run on clay.

Yeah, so I mean I feel like N8N and clay are vastly different tools. Like, like N8N is a broader workflow building tool.

It is meant to build workflows for really anything.

Clay is really growth and GTM related and so you know, they have their own crossovers. Like they can both do, they could both do like you know, API connections and they can do like formatting really well. But there are some things that are.

Better suited for anyone.

So for example building AI agents that do specific tasks for you based on pre trained data or data that you.

Want it to do like an action in. I would definitely recommend using N8N for.

Especially AI agents that use tools of any kind. Like if you're having an agent like do something in notion or send an email like significantly better to do an N8N like conversely clay. I think any enrichment based workflow should.

Just be housed in clay. It's. It's way easier than connecting all the API keys separately and super intuitive.

Also N8N is great for kind of like any sort of trigger based one off run I found like if you for example get on a call and like after the call the call transcript.

Gets like sent to an N8N agent to like, you know, give you a summary or something if you wanted to.

Build something, something like that. Like that's a very N8N friendly use case. Also anything just like really customer technical.

I feel like N8N is. Is better suited for like you could really build anything. And if you, if you have like a high enough skill level at the tool, like it's again, just a kind of a no code wrapper for, for building software or code activities. So it's really good. Big fan.

Do you have a preferred call transcript application for. For pulling transcripts in.

I've been asked this recently. I actually have. We've tried so many.

They all, in my opinion, suck.

The one we use is Fathom and it's annoying because Fathom only integrates with.

Zapier for some reason.

Like, so we have to do a Fathom to Zapier to N8N for any of the stuff we're doing. But I actually, I can't recommend a single one. Like we've tried a bunch of them.

They're all pretty bad, especially for teams. And if anybody has one, they recommend, please DM me.

I'd love to know what is the.

Thing that they're failing at or what's the thing you wish they did.

They all just have their own problems. Like some don't have a good API at all. So there's nothing you can really do with the data. Like for example, Fathom doesn't have a great API at all.

You have to go use Zapier to get any sort of workflow done.

Some of them are just really intrusive as well. They kind of boot into your meeting.

And they're all over your screen and just very annoying when you're on a call.

So they all have, I feel like just their own problems. I would love one that like, I don't know, it's recording like it's nowhere on my screen.

It's just kind of doing it in the background.

Like Granola, for example. But then Granola really doesn't have like a great API or like a way.

To interface with, with call, you know, recordings or meetings.

And then so yeah, they kind of just. I haven't found one that I've loved.

Yet to date, but I'm sure they exist.

I've heard really good things of attention. I don't know if you've heard of it and apparently it has built in both agents that you can plug into or like that run specific tasks with the transcripts. And they have about like 200 integrations. I can see definitely like Slack teams, CRMs, but a bunch of others. So like I'm sure.

Oh yeah.

OpenAI clay native integration. So like I'll definitely like share it with you but haven't used it personally, but for really good things.

Sweet. Yeah, I would totally love to see it.

Sounds like we might all be using OpenAI for this soon anyway.

Yeah, it seems like it.

You might as well, right? How would you advise like aspiring GTM engineers to start building on these skills?

Yes, it's a great question. I think the variety of educational content.

That'S coming out now are significantly better than before. I mean like when I started learning clay, there were.

There weren't even docs.

I created the docs in clay.

So like you had to just poke.

Around on the tool to figure out how it worked.

But now, I mean there are obviously like incubators.

Like you guys had GTM engineer school. I know there's like a boot, like some boot camps out there that can help with learning clay specifically.

So there are these kind of like in person like sessions that you can do if you don't want to like have to sign up for a course.

Or like pay any money to do so. There are awesome free, free materials out there as well, like Clay University.

Eric, now Slowski's YouTube channel, Brandon Charleston's.

YouTube channel is great.

So there are, you know, plentiful resources.

To learn these tools.

And they're also no code tools.

So they have, they operate in a relatively closed environment.

So actually getting like the technical acumen down to use tools like clay or N8N or lovable or exit or like any of them, it is very possible.

To get majority of it down pretty quickly. I think the hard part is the subjective strategic side of being a gtme. Very good point, Patrick.

Leave our listeners with one practical tip that they can implement today.

And this is just any tip. Like I can just give any like.

It'S not GTM engineering.

Yeah, somebody they're getting into GTM engineering. They don't have a lot of experience yet. You know, maybe they've done a couple of projects want something that could just up their game.

I would recommend if you're a GTM.

Engineer to, to not get too lost.

In the sauce of tooling. Ultimately the only reason anybody wants to use these tools is because they add value to them or their business is a lot of fluff out there. And I'm sure I've even supplied some of the fluff of the crazy stuff.

You can build with these tools.

Clients don't Pay for crazy. They pay for effective. If you're getting hired in house as a GTM engineer, you're not getting hired to build crazy workflows that people think are cool. You're building workflows that actually add value. And so think about these tools as a value vehicle and not a, like just a Lego set that you build for fun. Unless you're being a GTM engineer as a hobby. You know, if you want to do this as a career, it is important to go back to why does this add value to people? And if you can't think of why, then it's probably not a super useful.

Tool or a useful workflow.

Follow the money. Start there and then figure out the tool set.

Yeah.

Patrick, is LinkedIn the best place that people can find and connect with you?

Yep. Yeah, I tend to be most active.

On there and, you know, feel free to shoot me in DMS with any questions and totally done to help.

All right, fantastic. Well, this was really super. Mateo. Any other last questions we've got for Patrick?

No, I think we're good.

Okay. All right, sounds good. Well, then we'll just wrap it up here. Patrick, you dove today into a lot of your secrets and we appreciate, and our listeners appreciate it. You highlighted three tools. Clay for enrichment, N8N for workflows, and lovable for no code Software creation. You shared online. You shared a fascinating project where you built custom software programmatically at scale. That is pretty eye opening and everybody should check that out just for inspiration. And today you emphasize the importance of smart credit management. That alone can potentially save your company up to 60k by strategically using API keys. Very valuable stuff. And then your key advice here. Always start by understanding the business issues where the biggest opportunity is. Don't necessarily start with the tools. Find that specific use case, then find the best in class tool. And don't be afraid to ask experts for recommendations. How can I get data for X, Y and z use case? I think that was all fantastic stuff, Patrick. So thank you so much for spending half hour with us and we want to stay in touch.

Of course.

Appreciate you guys having me.

Thanks so much, Patrick. That was awesome.