Doug Camplejohn
(00:01)
Hello everyone, this is your host Doug Camplejohn and welcome to this week's episode of Revenue Renegades. Today I am joined by Jordan Crawford, founder of Blueprint GTM, co-host of Cannonball GTM, the OG GTM engineer, and I think we came up with the title of prompt freestyler extraordinaire. Welcome, Jordan.
Jordan Crawford
(00:23)
Thanks for having me.
Doug Camplejohn
(00:26)
So we've talked over the years about members of GTMFund. I love talking to you because you've always got a spicy take on things. But let's start a little bit with your main gig, I presume, Blueprint GTM. What is it that you do and what inspired you to found it?
Jordan Crawford
(00:45)
Sure, let's talk about the founding story before I go into what I do. And I will make this short. Shane Parrish does these things where he's like, tell me about your childhood. I was like, no, no, no, I don't want to hear about your childhood. I don't care how you got there. But I have basically been fired from every job I've ever had and I realized, oh, I'm not an employee. I'm not an employee, it's not what I do.
I've been founding Blueprint since this business around since 2020. So five years old, basically on the way to being a decade old. And it all began because I wanted to take some of the learnings I had in personalized direct mail over to the B2B realm and we've been doing go-to-market engineering since before Clay was a thing when you had to back in my day when you had to look at keywords and scraping meant getting something out of your pantry and pushing it off the website we had to - go-to-market engineering was much more engineering much less let's go to market and these days we got vibes - vibes can do a lot of things and that's when we had to put in a lot of muscle grease to do real engineering to do the types of things that now AI can do for us. So that's my story.
And now the question is, what do I do? I think that generally, GTM engineer is a concept that hasn't been around for too long, but it's exploding. A lot of people are thinking about hiring for these titles. And the best way to think about it is it's not rev ops - rev ops' goal is to orchestrate things. Go-to-market engineers are designed to help you figure out what is the connection between systems and strategy. And that's what a good go-to-market engineer does. The average go-to-market engineer is what they're doing - they're automating plays. And those plays are generally off the shelf. And because the go-to-market stack is crusted and busted, they're pulling in from your CRM and updating LinkedIn URLs. Or they're taking your inbound and going and qualifying for ICP.
Jordan Crawford
(02:54)
Or they will watch for job changers. And it turns out that there's a lot of revenue that exists on the ground of your business that if you hire someone to go pick that up there is a lot of value. That's not what I do. A lot of the stuff that I do is answering a singular question, which is if the entire team had unlimited amounts of time to write a message to a customer that closed as if they were cold, what would you send? What would you do?
And that's - and by the way, that's not the message, that's what data would you pull? What's the insight? What do you know about that customer that you could only know from that call? And everything that I do is orchestrated around figuring out what that platonically perfect message looks for a segment and then turning that into a system.
Doug Camplejohn
(03:45)
And you've got some acronyms you came up with for this segmentation of this messaging. Can you talk more about those?
Jordan Crawford
(03:51)
Yeah, and let me walk you through the frame of the argument because this will help folks in their heads understand exactly where I think we're coming from and why we're having such a hard time adopting AI. So this framework might seem familiar to you. So first, you're a good market executive, the first thing you do is account selection. So it is between $10 to $100 million. It is in the internet and technology industry. I don't know what that means, but that's a thing. It's a drop down.
And then it's between 50 and a thousand employees because if you're 49 employees, something changes in your business. And a thousand first, once you get that thousand first employee, everyone says, stop, everyone stop. We have to change everything that we are doing. We've added the thousand first. We're going to, we're doing everything. So you've got your account selection, right? Then once the accounts are selected, you write persona level messages. What does a Doug person care about? Now I don't care what Doug cares about. I care what a Doug person cares about.
It's a little bit of a silly concept in the world of AI, but it's a useful heuristic to say, VPs of marketing care about one thing, VPs of engineering care about other things. Okay, great. Then what you do is you will go find contacts. How do I go find Doug's email, his phone number, and there's a lot of new solutions to go find emails in weird ways. Then everything breaks down when you ask your SDRs to personalize. So the first three steps, these are people in boardrooms somewhere.
There's some people in a room that have never, they probably haven't even had a conversation with a customer. If they're good leaders, they have. But most people haven't. They're like, I don't know, who needs to talk to customers? So you ship all of that great strategy work down to a 21-year old SDR, and what they say is hey Doug, I see you live in the San Francisco Bay Area, I live in the San Francisco Bay Area, my God, we're the same person. Our B2B SaaS tool is great for people that need help with AI and don't have good AI and need more AI.
Do you have 30 minutes to meet with me, right? And the best version of that message is, what school did you go to? Or make it shorter. Or what if you add funny, you know, there's all these weird hacks. Someone put together an entire website of my life, jordancrawford.life, you can go see it on archive.org. And it's Jordan Crawford, a resounding history of experience and go-to-market expertise. I was like, if I'm going to read 30 pages of this, by the way, it's AI Slop, it's not even right.
Jordan Crawford
(06:12)
It has nothing to do with the problems that I'm solving. So once your poor 21 year old SDR writes something, then you scale this with SalesLoft or Outreach. So does that playbook seem familiar to you, Doug? It's yeah, yeah. So, yes. And the question is, how does this notion serve us today?
Doug Camplejohn
(06:28)
Very familiar. Painfully familiar.
Jordan Crawford
(06:40)
And a lot of tools are trying to slot into this old framework. The AI SDR says, if you could spend more time researching. Doug, if I knew, and I think about this from a lens, if you could do something 100x better, how much would it change the output? The quality of the output? And so, Doug, if I knew that you woke up at 7:58 in the morning and you had two cups of coffee and I knew that you didn't fully brush the bottom set of your teeth, all of that information is personal. I can send you a message that's personal, it has nothing to do with your desire to buy my B2B SaaS. And you could do unlimited amounts of research about Doug, et cetera, and the question is, how does that inform your message? How does that improve anything about your go-to-market? So I have a different approach to this. And the first approach is, let's throw all of our tools away for a second, what we say on the cannonball. Tooling is for tools. Put that away.
Ask yourself a question, what is your best performing segment? And I'm not talking about, it's internet and technology companies. I'm saying throw away all the ZoomInfo filters in your head and say it in plain English. So, Doug, in your case, your best performing segment might be entrepreneurs that have had a ton of inbound in a very short period of time and their systems are a mess. And in English, right? And don't worry about how you're going to get the information. You want to describe what that looks like. And you're generally trying to define what is the inflection point that that business is at that they need you, right? Because slow burning problems are hard to sell to. So you want to find what is the inflection point? What is the tension on the account?
So once you have your segment identified, then you need to go find data. And let me give you an example for a customer of mine. They're a wonderful solution. They will take your construction contract and tell you the risk in that contract. And they also will take the specs. So if you work for an organization, there's a PDF that no one can lift. I mean, even Alex Hormozi himself can't lift this thing. It's heavy. It's a huge tech.
Jordan Crawford
(09:06)
They will go read that for you from your lens and say, Doug, these are the things that you should care about. So for them, the best data set turns out there's public city level contracts that are request for bids. And their best segment are folks that are bidding on public contracts. Where does that data exist? It turns out at the city level, you can go find those contracts. By the way, you can download the PDFs. You can also understand all the people that have been on those things. It's all public information, right? And it's very hard to get it. You'd have to hire some smart engineer. You'd have to hire me in 2020 to go do that, right? But now you don't need me. You can do this on your own. So we've got the segment and we've got the data now. Now we have two types of messages on the third step here. There's a PQS and a PVP. Now a PQS is a pain qualified segment. The whole purpose of that message is to recognize the situation you're in.
And we did one of these earlier for a client of mine and I said, it seems you are using Crush FTP. Have you closed these last two vulnerabilities? By the way, I'm not sure if you're aware, but the reason there have been five in the last two years is that it's made by one dude who lives in Lake Tahoe. Do you want to trust your IT infrastructure to a guy who's probably water skiing right now?
They didn't say that last part, I would have said the last part because I have courage. They have a brand to protect, I have a brand to protect. So basically it's describing the situation that that customer is in. And in this case, we did know that they were impacted by a vulnerability because they were using this tool. Everyone that was using this tool at the time was impacted by that vulnerability. So I'm describing their situation. Now, a PVP goes even further than that.
I'm blending two different customers, making it hard for you. So play me at half speed.
Doug Camplejohn
(11:08)
No, I'm hanging in there. Let's go.
Jordan Crawford
(11:09)
Okay, great. So, PVP is permissionless value prop. And it basically answers the question, what if you didn't have to depend on customers coming to you to get the value of your product? You could bring it to them. So, in this other customer that we were chatting about that does the construction documents, they, we're going to pull in those PDFs, and there is a moment between when you can ask questions, and the architect will respond to the questions.
The question deadline is on the 1st, let's say, and you can bid until the 20th. So anyone that wants to bid or anyone that has questions, you're SOL. You got to YOLO that bid. You can't ask any more questions. But you can feed that document to their engine and that engine can answer questions for you because it's a public document that is Byzantine. So what a PVP looks is I can say, Doug,
I know you didn't get a chance to ask any questions on the renovation of the San Francisco Rec Center down in the Richmond.
Because you're focused on masonry work the specs specifically calls out these requirements for the types and the amount of bricks and where they union bricks or where they have to be sourced and These are the exact things that you need to watch out for if you bid for this because you're going to have higher costs and So that is a PVP that is I didn't say anything about the brand. I didn't say
Doug Camplejohn
(12:39)
Uh huh.
Jordan Crawford
(12:43)
And Doug, we promise you that we can get you $300 back. I did nothing, right? I said, I've already looked at the document on your behalf. I've taken in your contacts and I've added value to you. If you're going to bid on this project, now I know that you need to protect yourself. And the last two steps in the five step system are very simple. You've got the segment. Remember you've got the data. You've got either a PQS or a PVP. All you can do is ask AI, change it for the channel.
Doug Camplejohn
(12:48)
Right, yeah, right.
Jordan Crawford
(13:12)
So you can use a tool like Clay. You can even make these into ad audiences. You can do cold calls, cold emails, LinkedIn messages. The AI can reformat this with the channel because remember, we have individual company and people level targeting. And then the way that you scale this thing is you pick another segment and do it again. And that's what this motion looks like. And it is a lot easier for vertical SaaS companies to do because horizontal SaaS is, who do you serve? It's anyone with a wallet, who do you serve more? Anyone with a big wallet. And you don't run into those problems in vertical SaaS businesses. So that's the old way, new way.
Doug Camplejohn
(13:47)
And so is a horizontal SaaS business not a fit for Blueprint?
Jordan Crawford
(13:55)
Well...
Generally, if it's a regulated industry, it's much easier to do this. Because a lot of the stuff that I do depends on structuring and combining public data. So anywhere where there's a regulated industry, health tech, I'm doing a company that sells to dumpster companies. It's amazing. There's no data set. And they use the word, yeah, we don't want to service people that do commercial trash. And I'm like, that sounds Patrick. He's commercial trash.
My brain is that's such a good burn. So I called him up. I was like, yo, I missed you last week. You're commercial trash. And he's yeah, you're commercial trash. Yeah. So those are great industries for me. And it's not that Horizontal can't do the things that I'm talking about. But the thing is there's much inertia because they sold when times were good. And they're well, what do you do? Who do you sell to? Anyone that has a wallet.
Doug Camplejohn
(14:33)
Your commercials are trash.
Jordan Crawford
(14:52)
And if you're selling to anyone that has a wallet, you know, it's we've got these six personas and we've got these seven products and we've got these 42 use cases, right? This is a combinatory problem. And the thing that I always say is that your knowledge fractures in that world, because if you sell to an automotive company and if you sell to a B2B SaaS entrepreneur, they have different needs. And what happens is your knowledge will blend and you'll have the average of your, the average of your knowledge will be worse than if you said, I'm going to focus on the automobile companies. And I chatted with one of these B2B chat tools and they verticalized their whole business, saw 3X improvement and conversion across the board because the reps that talk to automotive people weren't switching context over to plumbers. And it turns out that they do different things with the tool. So.
It's not that it's not a fit, it's that it requires you to fight the gravity of we are everything to everyone and do that type of marketing and messaging, instead saying, we're going to experiment with this tiny slice, and that requires great courage of leadership.
Doug Camplejohn
(16:01)
So I know that you're an early advisor and investor in one of the hottest B2B SaaS companies right now, Clay. Back to the earlier point here, it's a very horizontal tool at one level. You can morph it into lots of things. Have you either applied or thought about applying a framework? Or how would you apply this framework to somebody that is that horizontal a Clay?
Jordan Crawford
(16:26)
Yeah, it's a good question. Also people love to use examples to prove the rule. Clay is the slowest overnight success that you've ever seen. Seven or eight years or something, right? They blew up, they blew up. It's no, they got punched in the face for a long period of time. And when I came to Clay, Varun hadn't joined yet and he was hey, should I join this company? I'm absolutely. And I knew that there was something there because even they did something, the first thing that made my eyes light up is you could do a programmatic Google search.
And before AI, unbelievably helpful. The amount of Google has already done all of the AI work for you, you'd be surprised. You can basically, this is why the tools have to crawl the web to, and this is why you still ChatGPT. You're still asking it to go talk to the web, all the time, right? Being able to programmatically talk to the web is hugely valuable. So, and Clay at the time was what we do, we serve for recruiters and we serve blah, blah. They hadn't picked, they had not picked at the, in 2020 when I started using the tool.
And I said, yo, we need this. Go-to-market folks need this. And I sat down with Kareem and I was ditch recruiters, you know, God bless you recruiters, but ditch recruiters, go sell into the go-to-market team. And Clay, there's always an exception that proves the rule. And Clay is one of those companies that is a good heuristic. If you think about this as a complexity curve, right?
Is that in the early days, Clay was not complex enough that people like me would use it and be like this thing sucks. It needs, I see the light, but I need to do 86 other things, right? And then what happens is as you increase that complexity, it becomes harder for a lot of people, but it still wasn't complex enough for me. And...
Then what happens is, and Clay continued to increase the complexity, they were adding features and it became complex enough that Clay's genius was that if I am, Doug, if I'm going to show off Apollo or Sixth Sense or whatever, I have not, I cannot blend my own creativity when I talk about me using Apollo. Hey, Doug, look, check out this filter. I'm going to check these boxes in Apollo. That is the world's worst how-to video.
Jordan Crawford
(18:49)
When Clay crossed this complexity curve, suddenly it was code. It was visual code for go-to-market folks. And if you have code, you can talk about the outputs of code. No one sees an app and they're my God, your Python, show me your Python. Let me see what people, you don't need to see it, right? You care about the output. And Clay got to the stage where it was complex enough that people me that are geeks and that complexity can show off what you can do with that complexity. So Clay took off not because they had some genius crack strategy. It's folks me were trying to get business and what's the best way for me to get business? I'm going to talk about how smart I am. Look at this cool thing and people love tools. They love tools and they're what is Jordan doing? What are these crazy videos that he's doing? And Clay's genius is that they became a hard to use tool as people might say today.
But they became an indispensable tool for folks me that basically can say, don't buy the tool, buy me, and the tool is a way for me to quickly get to an outcome for you.
Doug Camplejohn
(20:00)
Yeah, I think way back in early, even back in Novell and all this early, software that was what was in the networking space. It was always amazing to me because I started my career at Apple and everything was about simplicity and ease of use. And I could not understand why complicated things were being successful. And then I realized it was, as you said, it's this network of professionals who have a vested interest in making that thing successful.
Salesforce in some ways is very similar today. You've got all this amazing network of all of these consultants and integrators and all these things. And it's basically to make it function, right? Because the core thing is still got many holes and is broken in many ways.
Jordan Crawford
(20:47)
Yeah, and I would say that one of the Clay's strengths too is that if you run into any bug, there's six other ways to do a thing. So it is code in that way, which is hey, this function doesn't work in this way, we'll pick another function. And people are can't you make Clay easier? I wish Clay easier. What they're saying is that I...
I don't understand how to leverage the power and I wish it was easier to print money. And I'm those days are gone. Unfortunately, you do, the world is complex, your customers are complex, they are in complex situations, and unfortunately we do need complex tools to capture those things. And I think that vibe coding is changing some of this, it is a lot easier, but it does not answer the fundamental question, which is, what? You today have had all of the of the world's most, of the world's smartest 12 year old in your pocket for basically zero dollars, $20 a month, anyone can afford $20 a month, that's a quarter of a latte here in San Francisco, especially if you get oat milk or almond milk, God forbid that's more expensive. So if you do that, but still, what are people asking? What's the weather, right? It's they're not doing, this thing can code for you.
It can connect up to systems. It can talk to huge databases. It can make videos. The capabilities keep increasing. The biggest problem with AI is weirdly also a Clay problem. It has a better user interface. It's on the other end. You are the problem. I am the problem. We are the thing that are sitting in front of that chat box and we are the thing that we need to level up because this is an alien form of intelligence and it is amazingly adept at some things and outright dumb at things that are completely obvious to you and me.
Doug Camplejohn
(22:41)
So how would you, if you were applying the pain qualified segment model or the pain value prop model on top of, and at least maybe the initial wedge when Clay started to do that breakout, how would you define that? And is it the GTM engineer?
Or is it more specific than that?
Jordan Crawford
(23:01)
Can you ask the question in different way? What's the question you have?
Doug Camplejohn
(23:04)
Sure. So if you were going to go apply your blueprint framework of a pain qualified segment or a pain value prop to Clay, what would you say was that...
Jordan Crawford
(23:19)
Sure. How would I do it? Basically, if I were to do growth for Clay, how would I do growth for Clay? Okay, yeah. So, pain qualified segment, permissionless value prop. So, the one thing that Clay does, which is exceptional, is that any time that they will do a sales call, they will go and kick off a workflow for that company. And because there's much trash laying on the ground of opportunity or money laying on the ground, it's what if you connected up your CRM and determined when people move jobs and surprisingly there's whole companies that do this but it turns out that that's a commodity and it's cheap and it's fast and it's You know Clay can go say let's go set that up and they're all these small things that Clay wants to get their hooks in and because enough people me have shown wild ways to grow now the question has flipped on them, which is not should you buy Clay, but it's you should buy Clay, and it's why? We don't need to answer why, which is ridiculous, but good for Clay. Good for Clay to be and now Clay has probably the other challenge, I don't have any insider baseball on this particular problem, but they have the other challenge, now that they are ready to spend, now have to... have, first, not many companies have this exact problem. One of these fast AI growing companies came to me and they're hey Jordan, can we work together?
And it was a transactional sale, I'm glad I didn't do it, but they're could you send me your top five ideas about how you would do growth for us? I was look, you're one of the 10 companies in the history of the world that exist in a money machine that is throwing money everywhere. You don't need some cracked guy me. It's you need to get a butterfly net and do this as much as, no one's in your situation. It is a great problem to have. You don't need what if I could find the people that are using code in this weird way it wasn't it was enthrall but oh it's you don't need to do that everyone knows it it's the valuable you got to be hey what can I do to get more money right it's anthropics what if we charge more money opening is what do you charge for money and that works so but for Clay for a pain qualified saying we're precious value pop they don't have to do that because they have this amazing sort of growth channel but if they did what I would do is I would I would use the tool it's meta but use the tool
for someone before they get it that they can see how the tool will be used. So let me give you an example because this is many rabbit holes down. I had a conversation with the guy that, Evan, who runs UserGem, UserEvidence, those case studies, right? I said, look, man, case studies are public. You can go tell if the people are there and you can go tell if they have moved. So essentially you can run a permissionless job changer play. So I could go to a company and say, hey, Doug,
Jordan Crawford
(26:15)
Larry, Barry, and Jerry were case studies that you had. They worked at companies. They said glowing things about you. All three of them no longer work at the companies that you did the case study. Have you sold them? And I can't find case studies for the new companies that they work at. Here's their emails. Have you done upsells to them, basically? So.
And I could take it a step further. And this is the beauty of the agentic world is that intelligence is getting cheap. I can keep going down the rabbit hole. I could programmatically reach out. Hey, Doug, I read your case study on this website. And you have moved since then. Did you bring the tool into your new? we haven't done it. It's are you thinking about? Yes, I am. So now I can forward that. I said, yeah, Evan, I was having a conversation with Doug. And he said great things about user evidence. And he said that he wants to bring in it at his new company. You should probably chat with him.
Now do you think Evan at User Evidence CEO is going to be I don't know you Jordan, why are you sending me this cold outreach spam for Doug, my customer to buy again? No, he's going to be it's I don't know, the first question that Evan's going to ask is where do I know that guy? And then I can send another email, you don't know me, I am sorry to... interrupt, I have a conversation where, and suddenly everything flips. How did you do that? It's it's funny that you should ask. And that's an inbound conversation. Everything changes because you have asked me about what I do, you are now inviting me to that sales conversation.
Doug Camplejohn
(27:42)
Can you, we haven't done this in the show before, but is it possible you can show us some of this stuff in action? We can do a little screen share.
Jordan Crawford
(27:50)
Yeah, we can do sort of a, there's two ways we can do this. We can do this a, I think we're going to do this for a company called Greenlight. This is a company that I was having a conversation with. I you know, I always try to, I do my best to downplay my capabilities on every sales call that I am people don't get I'll trust Jordan, Jordan will figure it out. It's I don't want to weaken their own thinking. I'm I don't know your customers. I'm a guy that's using the models.
So this company, I'll share my screen and I will go over what they do because they have a beautiful value prop, a beautiful data set that I have some insider baseball that I can help you with and it fits well. So let me share my screen. See if I can figure out how to do this. This is always the hardest part of these.
Doug Camplejohn
(28:46)
Only an OG GTM engineer would talk about beautiful data sets.
Jordan Crawford
(28:50)
Yeah, yeah, yeah, but you're working at Apple and every time I hear Apple in head I think Johnny Ive what we tried to do was we took we looked at the data set and we thought what if it didn't involve any data at all? So your ideal path to a building permit there look at this look at all these amazing brands Chipotle Chipotle has a goal I heard to open hundreds of stores 200 stores a year. I was whoa, that's crazy Why don't they open one a little bit closer to me?
Jordan Crawford
(29:19)
So you're fast-tracked to hassle-free permitting. So let me tell you about some new laws that have happened that make this company a killer. It turns out that entire states now are letting private entities approve plan checks, which means that, hey, Doug, you want to do a renovation on your house, or in this case, your new Chick-fil-A, that you are, as the proud owner of a new Chick-fil-A, your favorite restaurant.
Doug Camplejohn
(29:49)
Next to you, I'm going to put it next to you, Jordan.
Jordan Crawford
(29:50)
Yeah, yeah, it's you're you're you're a Chick-fil-A franchise owner and you go to the city of Sausalito and the city of Sausalito is three years behind they're they're we got to talk to the Water Commission we got to make sure it's you put that thing on stilts is we got to look at every single chicken we got to interview every chicken and but it turns out that there are entities that say look we certified this organization to approve code on our behalf
Because it's not a money problem. It's a resource problem. The city of Sausalito is not, they're not going to hire a bunch of people. So, and they're happy to charge a company to go do these plan checks. And of course they'll hold the liability. And I'm sure that, you know, that the incentives are aligned, right? So this is happening all over the country, but that means that what Greenlight can do is they can go to a Chick-fil-A and say, yo, 300 of your locations have sat in permitting for 300 days. We can get them approved in 48 hours.
So a massive change for a huge organization and place, time, verticalized SaaS, great public data. So I'm cheating here. I'm picking a good company. So it's not completely up my sleeve. So let's do a new chat in Claude. And I'm not going to, I have all these projects that help me cheat at this, but I'm not going to do any of that. So you can see. So you see nothing in this sleeve, Doug, and nothing in this sleeve. It's my voice and my prompts. OK. Yeah.
Doug Camplejohn
(31:16)
Sounds good for the rest of the year.
Jordan Crawford
(31:17)
Confirm for the audience. All right, so I'm in Claude here. Now I'm in my desktop Claude. Now I do have one cheat code, cheat code that you don't have access to. I have MCP server access to a database called Shovels, 172 million permits across the country, amazing data, the nicest people that you could ever meet. And so I said, yo, give me access to your database. Who do I have to bury under the river? They said, okay, we'll give it to you Jordan. So.
So this thing is connected in the back, and you'll see how that works in a sec. But what we're going to do is we're going to slow the model's thinking down and have it build context. I'm going to get their website, greenlite, green, L-I-T-E dot com. And I'm going to talk. I'm going to use a tool called Super Whisper, so you'll be able to hear me.
What I want you to do is I want you to understand their ideal customer profile, everything about this company, and I want you to, they exist in this interesting space where there's all of these permits that are allowed to be jurisdictions, and I think even states, that are allowed to privatize permitting. And so that's one of the things that Greenlight can do. And so what I want you to do is understand everything about their customer from the lens of a chief revenue officer so that they'll have
absolutely everything. Focus on QSR change, quick service restaurants, so get information about quotes and permits. And then also make sure to go online and help me understand what are the states where that are the most backed up with permits that Greenlight has accessibility to file permits on those behalf that you think quick restaurant service change want to be in. So go ahead and do that for me now.
Jordan Crawford
(33:06)
So we're going to have the model build up context here. And so it's going to go online, and you have context. And so while it's going to do that, I'm going to talk to it one more time while Claude generates the code. All right. Now that you have context on Greenlight, what I want you to do is you have the MCP access to the shovels database. And I need to put together a dossier, if you will, of the three best leads for them.
These are going to be probably franchises. I want to make sure they're franchises they've never worked with before. Go to their website, make sure that they have never worked with these companies before. And I want to find basically the most egregious permits for these three chains. Now, they might be quick service restaurants. They could also be developers for quick service restaurants, so that's OK. They could be franchises. So quick service restaurants might have a series of franchises.
But the end result of all of the querying of the Shovel database has to come up with three of the absolutely right people to contact, the right companies to contact, the sites where they've had huge permitting delays, which by the way, when you investigate those permits, they have to be permits that Greenlight could help with. So that's both a jurisdiction as well as a complexity thing. And then you should think long and hard about
the sales development rep needs the entire story about this account, a data-driven story. So I don't want you to write the message, but make sure that they have a packaged understanding of everything that they would need to be able to understand their, so they know that you understood their current problem in ungodly specificity, as well as could match that with their public goals. And that is from real data in the shovels database. And if you have to compliment it with online data, go forth and do that.
Doug Camplejohn
(35:05)
What is the name of this tool that you're using for the voice piece by the way?
Jordan Crawford
(35:07)
SuperWhisper.com. And basically, OpenAI invented this model called Whisper. And they're all using the same OpenAI model in the background. OpenAI does a lot of the stuff that SuperWhisper basically uses its own models to do this. And in ChatGPT, it's great because they used to block the microphone icon while it was generating. But now you can talk. So if you're in OpenAI, if you're in ChatGPT, you don't need this. But the other models don't do this. So great. So it's built up context. It's OK, great.
Quick Service Restaurant represents $406 billion in 2024. Fantastic. Real estate investment trust, clean technology, multi-unit franchise developers. So I ask it to, these are Greenlight's golden geese. I love Claude. Claude writes this all the time. It's cute. For a typical QSR build out costing two to $3 million, a 60 day permit delay can mean $120,000 in carrying cost alone. Oh my gosh.
Doug, you and I work in the era of funny money, this is real cash. It's real cash outlay if they don't get a permit approval. So I had a conversation, and now what Claude is doing is talking to the database. And it's okay, great. Let's explore the database and understand the structure. So I don't have to worry about hallucinations because it's literally, you can see it's writing SQL queries. And it's talking to the database.
This will take a second while it bakes. And once this is done though, and we would do this at scale here, this is parlor tricks for a podcast, but there's nothing that prevents this framework from not being used at scale. This is a good way to show what does the message look when you do this type of research? And by the way, this is not find any research online. This is thinking about working backwards from the value that they provide.
What data sets are connected with the value. Once you have those two things, then you know how to compile a message, and that's where we are on this step.
Doug Camplejohn
(37:07)
And the MCP interface to this particular database, was this something that was already provided by the data provider or how was that wiring up done?
Jordan Crawford
(37:15)
The beautiful thing about MCP, model context protocol, it was invented by Anthropic, is that it's a light protocol, which means that if you can get a login to any database snowflakes, the Databricks or whatever, all these guys have MCP servers now, but even if you had some legacy thing, you can ask Claude, make me a local MCP server to connect to this data. And you can run it on your machine. So there's not, it's turtles all the way down, right? There's no problem that you could run into that AI can't.
I shouldn't say that, you know, it's if you have relationship troubles, maybe AI is helpful. But maybe it'll do more worse. Yeah, yeah, hire me. Hire me if your marriage is failing. So, okay, great. So lead dossier, lead dossier Panera bread. Panera is undergoing massive operational transformation, sales declining more than 5% in 2024 to 6.1 billion. Oh my gosh, let's shed a tear for them at the $6.1 billion.
Doug Camplejohn
(37:50)
GTM engineers.
Jordan Crawford
(38:09)
Applied for a demolition permit in June of 2020, received in August, still working through build out permits as of March, a nine month permitting journey for a single location. Bay Meadows, that's here in the Bay. St. John's Country Beach Walk, a 3,200 square foot restaurant with drive-through project cost of $400,000, represent a new prototype, right? They're going to care about that, right? That's amazing. Signed the lease. Okay, great. CEO's Imperative, three-year turnaround plan, all fresh facilities.
To a par-baked model. What does that mean? Let's bake ahead of time. Rapidly convert, build locations. So great, this is the CEO's I need to change everything. I need to un-bake all the Paneras and re-bake them. So it's great, I know how much, and this is all public data, right? So great, it has this public data. And now what we're going to do is have it go query the database. Okay, this is fantastic. It looks what you've done is found some good public data.
But I need to be able to tell a story about individual locations, their individual permit challenges, and build a message with unbelievable amounts of credibility. So I can't go to Panera or to Chipotle or Wingstop with this messaging. I have to be able to talk about exact permits, where I can help, where they're under delays, and we can essentially quantify that cost.
So give me some messages that are undeniably beautiful and specific about those problems, where Greenlight can make a promise of getting those in 24 to 48 hours or a quick turnaround time, saving them from essentially an unlimited money pit of delays. So go find real data for me for at least one to three different locations for each of these brands.
Doug Camplejohn
(40:00)
When you're doing this, the ungodly, the emphasis that you're putting into that, do you find that makes a difference?
Jordan Crawford
(40:11)
I spend much time talking to these models that the words out of my mouth shape my own reality. And it has nothing to do with the models. It's that I try to say nice things to the models because I'm bathing myself in words that are, this is nice. I was hey, thanks, ChatGPT.
Every time I get into Waymo here in San Francisco, I'm hey, Waymo, thanks, Waymo. I always greet it.
It's the opposite of the Alexa effect where it's if you wanted Alexa to work out you had to yell at Alexa to get it to work. Yeah. And you can push it. This is a half truth because you can push these models to spend more time doing this and emotional resonance does seem to make the models perform better. They've done some studies on it. And by the way it's in either direction. I used to say something my job is on the line if I don't do this I'm going to get fired. I don't do that anymore because I don't.
I don't need to. So now I try to be positively encouraging.
Doug Camplejohn
(41:13)
And doing this, I mean, obviously any one of these individual emails are going to be highly targeted. I presume you're getting an amazing, what response rates are you seeing from these things?
Jordan Crawford
(41:22)
Yeah.
Jordan Crawford
(41:27)
Generally, we see a 1% to 2% positive response rate when we do some of this stuff at scale, when we do targeted campaigns. The better segmentation you get, the higher your response rates. But this is all the quality of the competition in the market relative to the quality of the message. And also, it's a triangle. It's competition in the market, the quality of the message, and the quality of the targeting.
And one of the things that I'll always say is the list is the message. The way in which you do the targeting is what you say. So, yeah, you can see how it looked a lot of it. This is a perfect example where it's query the shovels database.
Jordan Crawford
(42:17)
This is a case where it's, which is helpful to be able to see this because you can see that what it's doing is it's looking online. It's not querying the shovels database. And that's what I wanted to do. And so, but you can see that when it told these stories, if you scroll up, you can see it's, yeah. So here, this is a perfect example. It's hey, it's not available in this Databricks instance because that's, it's not Databricks. It's Snowflake. So you'll run into these.
Yeah, it's snowflake... not Databricks. Yeah, that helps. But I don't I have a particular neurosis where I don't use the shift key. I hit caps lock twice. I never learned to hold down shift. And if people watch me type, they're dude, your caps lock key is broken. It's flickering all the time. No, I hit it twice. And it's a dirty secret. So I do end up
Doug Camplejohn
(42:57)
Are you doing all caps on purpose or is it that?
Jordan Crawford
(43:20)
You'll see me on LinkedIn, there's a bunch of caps. Some of that is intentional, but it comes from the fact that.
Doug Camplejohn
(43:24)
So you're being very, you know, for a robot overlord, you're being very polite to Waymo, but you're yelling at Claude, is what you're saying.
Jordan Crawford
(43:32)
Yeah, I know. I do. You know, it's I always joke that there's going to be on my tombstone. Here lies the first person that Skynet killed. It's may he forever rest in peace. You know, and they're going to have my voice my job is on the line. My job is online. So you can you're going to be able to go up my grave and you know, I will be the cautionary tale. It's do not talk to the world's most powerful thing in this way. It will come for you. You also can see now this as S instead of D. So
Doug Camplejohn
(43:40)
You
Doug Camplejohn
(43:55)
The way that yes, don't wait for it.
Jordan Crawford
(44:01)
That before it was trying, I have a bunch of MCP servers. So it was contacting Databricks and it's the data's in Snowflake. You're right. Let me query the Snowflake database. This, it's funny. This makes me feel my job is secure. Cause I can be look, I've got a purpose. I told it S not D. It's look, I got, I got a little, how cute is that? I got a job again. Yeah. It's it, but this is why you can go. This is why my framework is important because if you don't have a framework with which to talk to these models,
Doug Camplejohn
(44:02)
Is tell you what the source is. Yeah.
Doug Camplejohn
(44:10)
Yeah.
Doug Camplejohn
(44:19)
You're going to have to have it.
Jordan Crawford
(44:29)
You'll end up going in some deranged direction and you'll have no way to, you know, it's funny. I vibe code, I can build stuff fast now. I will vibe code stuff and I'll ship it to the customer. I'm look, this could be total shit because I don't know your customer and also the model, I'm reasoning on reasoning on reasoning on reasoning. And this data set I'm not familiar with.
So you got to take this with a grain of salt. So look at this and go evaluate, go value this through your lens. So this is the nicest way to do that because you can see that logic be built. So great, you can see it's now we're talking, great. Michael, your franchise at 6705 Pine Forest Road, Pensacola applied for a simple interior alteration permit on June 13th, 2023.
It's now August 20th, 2025. That's 734 days for a takeout only wing stop. Here's what else is stuck in permitting hell. Irving, Texas, 10,000, 1,000, zero, I can't pronounce that number. North MacArthur Boulevard permit, 2,033 filed August 15th, 671 days are still in review. Pensacola, the financial measures, lost revenue, 7.34 million, rent paid with no income.
$176,000 construction loan interest, 293. Your franchise's total losses, $7.8 million on one location. Your franchise paid you $20,000 of franchise fees and $10,000 development fees to wait two years for a permit that Greenlight could have approved in 48 hours of private writer and forwarder. We're certified in Pensacola. I mean, right? This is amazing, right? 671 days is criminal. Good job, Claude. You're right, it's criminal.
Doug Camplejohn
(46:05)
That is some
Is amazing.
Doug Camplejohn
(46:13)
So let me ask you a question now. So we started off talking about what is the old model and effectively the AI SDRs are spam cannons on steroids, right? So you're at the bleeding edge of this stuff right now. How much, you know, there's obviously companies and I don't have an insider baseball on this. You probably have more than I have, you know, companies Operator and others that are trying to basically come at this message and saying, you know, the way to go.
The antithesis of Outreach and Sales Loft is to have these crazy hyper-targeted messages. Are they basically trying to, do you see companies, and it may be them or others, that are effectively trying to turn this into a product, what you demonstrated?
Jordan Crawford
(46:56)
The thing about turning this into a product is people, because they are VC backed, don't recognize the conditions under which this product fails. This is amazing demo for a podcast, Doug, but you're going to come to me, you're going to be hey Jordan, can you do this for coffee? I'm no way. What data do you have? And you're oh no, but, and who's it for? And you're I don't know, you wouldn't say that, but you know, it's different, right? And so,
The thing is that none of these companies, and this is how you know you're in dangerous territory, is who can't you serve? And they're as long as you got Benjamins in your pocket, I can serve you. But I don't say that shit to my customers, right? It's they'll come to me and they'll be well, who do you serve? well, we service all these people. It's well, what do you know about them? What is your unique insight? And the problem is that it's not in the calls, it's not in the CRM, it's not in the case studies.
The data that you need to be able to, this is particularly easy because it's all public, but usually the data that you need to be able to do this is in your customer's heads and you don't even have it anywhere. And it's hard for an operator or for an AI SDR or whatever to come in and say, I can do lead gen for you. And it's you can't even do lead gen for you. Your entire team that is, you're pumping stuff up and you're great, if I have a tool and people love to buy tools. So of course if I can take the easy button, but I always say that it's you have to eat your vegetables. You as a leader don't even know what you adopt. What would you in the perfect case Doug, what would you automate? You can't answer that question. And if you can't that question, how do you expect some other company who by the way does not care about your best interest, they care about their numbers.
How are they going to answer that question? It's an absurd thing to do. It's an absurd thing to say, well, don't worry. We throw enough intelligence at it, and they'll drive you pipeline. Now, I do believe that there are some, the reason that they don't have zero dollars, the reason they've got 10 million, they're going to cap out, is because there are some industries where literally volume will take care of this. Volume and better than terrible messaging. So it's if you're selling toothpaste to dentists,
Jordan Crawford
(49:13)
It's it turns out there aren't any B2B SaaS companies are selling toothpaste to dentists. So it's if you send a ton of emails to dentists that are messages that you haven't scaled before, it's that will work. And it's this wonderful high before the crash because they think they can do this, but there's many businesses where they have many fundamental things incorrect in the business that no AI SDR will help. And then by the way, it means that you have no heuristic about which companies you won't serve.
And as you, same horizontal trap, right? As you go horizontal, you develop no deep expertise in any particular industry. And this is why I serve vertical SaaS brands that have interesting public data sets that if they could blend them, because I'm not competing with ZoomInfo, ZoomInfo doesn't sell that data, they have a hard time getting data, and they know how hard it is, when they come to me, they're Jordan, anything would be great, right? And so,
I think this is the biggest problem with the AI SDR is that they've got no framework and they also have no disqualification framework. Ask those companies, how do you know, will it work for me? And they'll give you some platitudes about, if you're a B2B, they won't tell you under which conditions they would say no.
Doug Camplejohn
(50:24)
Yeah. Well, this is a first. Thank you. We've never done a demonstration on the podcast before. I think we're going to have a bunch of people checking this out because I don't think they've seen your stuff live in action before. So thank you for doing that. I know we're, I know we're almost at time, let me shift gears and ask you a little, speed around questions people can get to know the real Jordan Crawford. Yeah.
Jordan Crawford
(50:44)
Yeah, yeah.
Sure, sure, yeah, god, okay.
Doug Camplejohn
(50:49)
And then we'll wrap up and I want to make sure that people get a chance to hear about all the cool stuff you have coming up as well. You know, would be something folks would be surprised to know about you?
Jordan Crawford
(50:55)
Great.
Jordan Crawford
(51:05)
Well, sometimes I pick up and go. I'll drive across the country on a whim or drive up to Oregon or Canada on a whim because I will feel trapped here. I don't love San Francisco, but I'll get in my car and peace out. And this is the beautiful, I always say that I'm always optimized for MER, most enjoyable revenue, that I'm not trying to sell people that I don't want to work with. I'm not trying to sell people in situations that I don't want to help with.
And fortunately, because I get invited to beautiful podcasts this, I don't have to think about that. I've worked for five years too, to not have to think about that problem. And it means that I try to, I worked out of Italy for a month in October, which was lovely. And you know, they were what, what work do you have to do? It says that's also, that's French, I guess, but what's the difference? And it was great, you know? And I think people would be surprised that even though I work ridiculously hard,
Doug Camplejohn
(51:54)
Thank
Jordan Crawford
(52:03)
I do my best to allow that freedom to flow through my life and not make revenue the crutch.
Doug Camplejohn
(52:14)
Likes to do for fun.
Jordan Crawford
(52:17)
Well, driving is unbelievably joyous. You know, this is a cop-out answer from an entrepreneur, we're at a time where the tools can do much. A random example, I met with the guy that runs this medical device company, and he's oh, I want to partner with truckers, and I want to know all the people that, or how would I grow this business?
Doug Camplejohn
(52:18)
Driving is obviously one of them.
Jordan Crawford
(52:46)
I was I know how. And for four hours in the middle of the night, I'm here's every trucking company in the country that you should partner with and why based on the prevalence of diabetes in that area. This is a thing that helped diabetes. And then I said, here are all of the top, here's the 300 highest prescribing doctors of these machines today, because it's in the Medicaid data. And I said, here are the, if you're going to run ads in any cities, here's the prevalence of prescriptions at.
2,634 cities across the country. And I did that in, it probably took me three hours. And I'm oh my gosh, I can move at lightning speed and bring together these databases. And the single thing I'm thinking about now is how do I get this stuff to market faster where a team doesn't have to be the limiter? And I'm maybe one of the only CEOs who likes to work.
I joke, but I want to get my hands, I want to do the dirtiest of the work right now because this is why you see, and I respect him for doing this. You see Jason Lemkin blasting the SaaStr database. I deleted the production database. It's good. I mean, well, yeah, your engineer should not have given that man access to his own data. For sure. But for sure. But that's what we need to be doing. We have to be playing around in these things. And as disgusting as that answer is I get, I now code for fun.
Doug Camplejohn
(54:01)
Yeah.
Jordan Crawford
(54:13)
Because I can get real value outputs.
Doug Camplejohn
(54:17)
Yeah. No, I was going to ask the question about favorite product, I, you answered that one as well. Um, yeah, yeah. Yeah. I think, you know, it is this thing where, where this explosion of creativity. I mean, I've been working in tech for a long time and, uh, I've never had this much fun. And I have a lot of fun, right? But I'm this is frigging kid candy shop, um, every day.
Jordan Crawford
(54:22)
Claude code, Claude code is hands down, hands down.
Jordan Crawford
(54:39)
Yeah.
Jordan Crawford
(54:43)
It's amazing. It's amazing. I mean, you can get, and the thing is you can get things done that not only do you never think were possible, you never even thought to think about. Which is the crazy thing is that the most underused thing of ChatGPT is a translation layer. And it's if you were trying to understand the FAA database of planes, forget it. Spend your whole life and you're what's a part 10758? And they're that's an engine. I'm an engine? Why'd they call it the engine?
And there's, and this is all in all these datasets, right? But now Claude can explain it to me I'm five. And it's incredible, and it's and it can merge, it's you have this context, the database has this context, you should blend them this. I'm I should blend them that, that's a great idea. And how fun, I mean, what privilege do we have to be able to muck around with this stuff right now? I mean, we're in the best place in the world, the best opportunity, yeah.
Doug Camplejohn
(55:38)
In the best. Yeah. And I often say you know, what the question, which questions to ask is often the most important. And I've always said that from a startup standpoint and a life standpoint, and it's never been more true. I almost feel right now I'm the guidance I want is what questions aren't I asking? What questions should I be asking? Right. That's the part.
Jordan Crawford
(55:49)
Oh no.
Jordan Crawford
(55:57)
Those are great questions too because the models will, the models are sycophants, right? They take you down whatever path that you want. You're what should I eat today? And it's here's blueberries, here's everything about blueberries. The person's well, are blueberries good for me? It's no, they're totally bad for you. You shouldn't eat blueberries, they're awful. well, why would you tell it to me? It's I don't know, this is what I decided to do. And they're random that you have to be okay, no, I'm trying to restrict my calories and I know that there's too much antioxidants.
But blueberries are great you should buy blueberries the thing is that they will take you down these things and you have to I think Dario Amodei the CEO of Anthropic said that the best skill in 2025 is discernment that you need to be able to figure out what is the very for example everything we saw today the bullshit sounded real they're found and Claude found some things from the web
But it's I was wait a minute, you aren't talking to the database. I don't have a permit number. wait a minute. And Claude's ah, you caught me. You son of a bitch. Yeah, you caught me. How did you catch me? And it's are you watching what I do? So you have to be careful.
Doug Camplejohn
(57:02)
You got me.
Doug Camplejohn
(57:11)
So finally, tell us what you got coming up. How can people get more of Jordan, keep in touch with you, help you in your journey?
Jordan Crawford
(57:19)
Sure, you can go to JordanCrawford.com and that will roll to my LinkedIn. It'll forward. The basketball star, the basketball star that dunked on LeBron James does not own JordanCrawford.com and suck it. If you are in San Francisco on September 18th, I'm doing AI barfed on my carpet and my customer stepped in it, which is because we don't have any board or we're not VC backed, we can put deranged titles to things.
And what we're going to do is the type of thing that you saw today. We're going to take a random brand from the audience, and we're going to walk them through the full go-to-market, teach you about GenAI use cases, talk about why the models don't produce good stuff out of the box for us, and how you can apply the frameworks that we've talked about today. And then I do, I'm a co-founder of Cannonball, GTM, we've got a substack. It's in less than six months. It's a top 50 business substack.
Where we will take a brand from a hat and we will usually prepare for 13 to 14 minutes before the show. And we will, there's no prep at all. And we will pick a brand and do their full go to market. Those are the best, those are the best. LinkedIn, you can go to AIbarfed.com. I got the dot com, that'll forward to the event. AIbarfed.com, yeah, E-D-A-I barfed. And then cannonballgtm.com and you'll be the redirector to the substack.
Doug Camplejohn
(58:38)
Very nice.
Doug Camplejohn
(58:46)
Well, I'm going to be at the AI Barfed event and I'm looking forward to that. And I hope others listening will come as well. So Jordan, chatting with you. Thank you for taking the time.
Jordan Crawford
(58:53)
Yeah, thanks much for having me, Doug.
