[00:00:28] Jay: Hi, everyone. Welcome to The First Customer podcast. My name's Jay Aigner. Today I'm lucky enough to be joined by Brad Gardner, who's the founder and CTO of Seven Hills Technology, and the only guy who had more steps than me in Durham.

Hello. How're you doing, Brad?

[00:00:42] Brad: You're pretty good. Yourself?

[00:00:44] Jay: I'm good. I feel like I spent a lot of time with you. It was great. We got some, some quality time, in Durham, which was beautiful, by the way. Lots of history. And, where did we meet first? It was Duke Univer- It was at Duke. Was where we were at? Duke. got the cathedral five cent tour, and it was fantastic.

[00:00:58] Brad: So, Or no, for the ballgame. We met at the ballgame beforehand

[00:01:01] Jay: That's right, that night. See, it was a long drive. I'm not gonna get into the drive though. But, so tell me, Brad, where are you from originally? Where'd you grow up? And did that have any impact on you being an entrepreneur?

[00:01:12] Brad: A little bit, yeah. So, I grew up in Cincinnati, Ohio, where I still am. Kind of a lifer here. And, entrepreneurship is in my family a little bit. My grandpa owned a business, that did well for them, and it's always been on my radar. So the combination of doing something in technology and owning a business is kind of where I was always pointed.

[00:01:29] Jay: What, type of business did your grandfather have?

[00:01:32] Brad: He had an appliance repair business. It was

like a warranty dealer for GE in the '70s, '80s, and, did well. He sold out of it in the '90s.

[00:01:40] Jay: Okay. Did you remember any of it as like a kid of him like being a business owner guy?

[00:01:44] Brad: Yeah, absolutely. We're

running around their shop and the people that work there and all of that.

[00:01:48] Jay: That's cool. Well, so, how long after that did you start your own thing?

[00:01:52] Brad: Quite a while. So we started Seven Hills Technology in, 2015.

So, it's been a little over 10 years now.

[00:01:59] Jay: And what's, where did that come from? Was it just you? Did you have any co-founders? What's the story?

[00:02:04] Brad: Yeah, I founded it and, with a co-founder who I had previously, hired and worked with, at the time. And then, fast-forward to 2020, we ended up kind of shaking up the leadership of the company, and we had already brought on a third partner in Snyder's. There's still two-- There's two of us now. We were briefly three for a while

[00:02:20] Jay: Okay. Where did the idea come from? What, you know, why did you do, what were you doing beforehand? How did you kind of jump out and start this thing?

[00:02:28] Brad: Yeah. So my background now is 20 years in consulting, but I guess at the time it was about, or it was about 10, years in consulting at the time. Everybody seems to have a good, like, founder story and I lack that. For me, it really just seemed like a good idea at the time. I'm impulsive. I had a job with a former client, and I had another former client who wanted me to do work.

I'm like, "Well, I can have a job or I can have two clients." So I, I quit my job and told them that they could be a client, and then I had two clients.

[00:02:55] Jay: That's a great story.

[00:02:57] Brad: That was

[00:02:57] Jay: I think, the romanticized founder story is kinda overplayed at this point. I think, like, a lot of people do what you did. That's what I did. I mean, I just kinda... You go- you went for it. So I think that's a great ... That's honestly what more people should do, I think, you know, is you believe in something and you take a swing.

So how did the first engagements go?

[00:03:16] Brad: Pretty well 'cause it was doing the same work I was doing day by day

so a different pay- just a different payment model. It's

like now I'm sending you an invoice instead of you putting me on payroll.

And then we had to learn how to do that for two clients at the same time. We did it part-time for like six months and then jumped in full-time and then kind of that was that.

It was pretty, pretty easy transition. I was just doing the same thing I was always doing and starting off it's pretty easy when you're doing software consulting 'cause you're just continuing to do the same thing. None of the business challenges had really hit yet.

[00:03:45] Jay: Did you guys, how did you guys, like, parlay that relationship? How did you switch from the, you know, from where you were to kind of being a new vendor with those clients? Was it a pretty easy transition?

[00:03:54] Brad: It was pretty easy. I met with, we gave them, we gave them a really good deal. I think that was probably what made it easy, is I took my salary and I converted it and added back the, the tax burden to it, and then converted that to an hourly rate. So it's like, "Here's your, here's your rate. You're paying me exactly what you were paying me otherwise, and you don't have to cover my insurance anymore."

So they got a deal. And then went finding my own insurance, and that was that.

[00:04:15] Jay: So what did you do back then compared to what you do now as far as a product or service offering?

[00:04:22] Brad: Wow, a lot's actually changed in the 10 years. So at that time it was me just writing software for clients. And so now we've branched into a whole lot of, I call it like higher degree, higher level consulting. We've launched a bunch of what we call productized services to how we go to market.

I've got a sales guy now. We've got a bigger team. So I do a lot more oversight, process engineering, things like that for the company now, and a lot less client work.

[00:04:50] Jay: Mm-hmm.

How did you guys survive the great kind of axing of generic custom software development shops in the '21, '22, '23 timeframe where it seemed like they all just started to get eaten alive if they didn't focus on something very specific, which a lot of them tried to do, or they picked three verticals and they weren't really great at any of them, and a lot of those, you know, had to lay off 75% of their staff over the last, you know, couple years.

How did you guys survive that timeframe?

[00:05:22] Brad: Good question. I'm really stubborn about being a generalist and, 'cause I'm-- And this is an internal conflict that we have with our, like, sales and marketing and my current business partner is, like, do we niche down? Do we focus on something on a specific industry and a specific expertise? And I'm kind of the opposite.

Like, I don't see myself personally ever, fitting into that mold. I think, most businesses think they're unique and they're not, and you end up doing the same thing. Like, the same kind of work that we do for one business generally transfers pretty well to a different type of business in a different industry.

There's a little bit of a uphill climb to, like, learn the lingo and learn how to talk and learn the things that are important to them. But when you're building business software, you're building the same kinds of things. So it's like building inventory management systems. You're building, like, integrations with accounting systems.

It's kind of that every business has those same things. We typically don't do as much work at the, like, bleeding edge, like, industry-defining kind of work. It's like we're focused a lot more on the operational software that gives you a-- keeps you, like, lean and mean and operating well.

[00:06:31] Jay: Well, I mean, has that been commoditized at all recently? The operational stuff, the less creative though?

[00:06:39] Brad: Commoditizing it. Like, that to me, that's with AI and how everything is easier to build now, what I'm really enjoying is seeing companies enter the market that wouldn't have gotten in the market before, where like, "You know what? Custom software isn't really for us. There's an off-the-shelf solution that gets us 80% of the way there, and we're-- that's good enough," because going fully custom is like total cost of ownership is hard, building it is hard. And now that math has changed, so now we're the ones like picking up those kind of companies and saying, "Hey, we can take that thing you're paying, you know, a big ticket for monthly, and we can rebuild it however you want for a one-time cost, and then switch into a really low like support mode that's significantly cheaper than what they were paying before, and we can do it extremely fast."

And we're seeing a lot more traction in that kind of space.

[00:07:28] Jay: Now, the only question I always have there is like, if it's slowly becoming more commoditized, does that eventually become a race to the bottom? That sort of model. Like, do more and more people able to do that for cheaper and it just keeps the cost just keeps going down. Is that a chase down or is it like level out at some point?

[00:07:47] Brad: I think in some areas it's absolutely a race to the bottom if you're looking kind of like at the edges of things. But I the-- for more operating towards the center of the bell curve, like you've got a bunch of business owners or like department heads who have budget authority and things like that, that they just want a problem solved, and that's why I sign up for SaaS software.

That's why they have something custom built. They don't-- The idea that we see online currently of like, well, if it's that easy to build, then everybody's gonna just build own software. Anybody can just install plot code and just build software, and it's true, but it's gonna be a bad experience.

And so to me, you're still like you're solving a problem for money and that has value. I don't think that changed materially. Like the window of like project and ticket sizes that we're operating in has changed. Like we're doing smaller projects for smaller companies, we're doing more of them, and value is still there.

'Cause like if you're a moderately sized regional like construction company, you don't want to hire a software developer 'cause you don't have enough to keep a full-time developer busy. That's a big cost. You don't want to hire a contracting company necessarily, or you don't want to go to a big expensive SaaS company that's gonna charge you based on revenue, especially in construction where like revenue is like, inflated.

[00:08:59] Jay: Mm-hmm. Sure

[00:09:03] Brad: But instead you go to a company like us and like we can help you operationalize your software, stick around for a little bit of support. Like it's a much better value play, in that space.

[00:09:12] Jay: Yeah, it's really interesting. Yeah, and, what about implementation, of those AI systems? Like I know everybody, you know, wanted to just add AI generally to their platform or whatever when it first, you know, became the hot ticket. But what is that translating to today? Like what are real products that are like kinda starting to shake out as opposed to, you know, everybody...

Or do you see a bunch of asks still for that like, "Hey, we just need like a chatbot that is AI for us right now"? Like what level of like, actual AI implementation are you getting requests from the general kind of masses right now?

[00:09:49] Brad: Interestingly, it's about what you'd expect but in different ways. So I'm seeing a lot more, people who are, like, paying attention to AI, but for doing more in the, like, traditional machine learning space where, because things like ChatGPT made everything kind of ubiquitous where, like, everybody has an AI subscription, everybody has talked to a chatbot, and, like, most of my clients have, like, a project in one of these where it has their entire, like, company charter and they're helping it to help-- using it to help run their company and they're like, "This is amazing," and it's just text generation.

But are like, "Well, can we use AI to do this or that or that?" I'm like, "Well, yeah, we actually could have done that five years ago, didn't, you didn't understand it then."

Because

machine learning and working with, like, neural networks and things like that have always been not really approachable to the layperson, but now they think they have this magic AI that can do all these things.

We're like, "Well, sure, we can do exactly what you're looking for, but it's not, it's not ChatGPT doing it. We're gonna train a model to do these things now because now you understand, like, the application." So we're seeing an uptick in that, which I think is really fascinating.

[00:10:57] Jay: It is really interesting, and it's, you know, hitting the... it's like trying to find a nail 'cause you got a hammer, you know? It's like, see that all, especially even with agency owners, right? That are like gonna-- they're trying to build their own products or doing all these things, and it's... I was just talking to a couple of them earlier, and I was like, "This may be the ultimate trap for ADHD people."

Like, you can do anything that you've ever wanted to do at percent. And it's like, that's an ADHD person if you wrote it down on paper, right? It's like, do everything, do a million things, but get it done to 80% and then wander off and do the next thing. That's kind of what it feels like we're all subscribing to a lot of right now.

We're trying to do too much at the same time. But, so what about the price? We talked about that at the conference. I am curious, like, as an agency that's passing through this stuff, like, what does happen? Like, if it's, you know, we talked about a crazy number. I think it was like 1,000 times what it was, you know, now, or 100 times what it was.

But, I mean, what is the future of that pass-through, do you think, if it just continues to climb and it is 10 to 20 to 30 times in the next year or two?

[00:11:56] Brad: Yeah, the pricing's a moving target. So with the rapid efficiency gain for people who are competent with AI, it changes our pricing model where we're looking at either higher time and materials rates or, pushing more towards fixed bid and retainer models. And that's one thing, that's one way to kind of like help push that cost through a little bit because there's a cost for us to use the AI, but, on a lot of projects it's so far because it's so heavily subsidized, it's been relatively immaterial.

But that subsidization picture changes, like it's just gonna be a moving target where, I think there's a lot of pressure for them to stop subsidizing as much. But they do that, it threatens their business model because people are using the token counts that they are because it's subsidized. If that tanks their user, like their usage, how's that impact, like investors and things like that? So there's competing priorities there, and I'm hoping, I'm hopeful that local models continue to become good enough where we can install a local model here that's as good as like, call it Opus 4.6, which we can be extremely productive on.

And

we're getting close with like some of the Qwen models are there or close to it. And so, you know, fast-forward six months, we might be installing a really competent local model on a server in our office and cutting our cloud code usage way down.

[00:13:17] Jay: Yeah. Yeah. Well, and you were-- I mean, we saw, again at the conference,we're still kind of doing AI the dumbest way possible. We're just brute forcing this thing. We're burning down the world, like to get 1% more because

[00:13:30] Brad: What a transformer is

[00:13:32] Jay: we have to iterate so many times, and it's like obviously there's a to- I mean, the efficiency gains just seem to be almost infinite when you look at how ridiculous the problem is that we're facing, 'cause it's...

They're doing it just about the worst way. I mean, it's the way to do it 'cause it works, but like, it's a wild way when we've got a a little glance under the hood like, "Oh, this is actually not,

[00:13:52] Brad: It's linear algebra. It's linear algebra matrix math at scale in a big loop. It can be computationally difficult, and so GPUs are suited for it. But like I think there's some cool innovation there. If you look at Google's like TPUs, things like that, that are looking at really higher efficiency and that helps the power draw, things like that.

But even the way that they've structured some of the models where like to really address different levels of computation, like even playing with like Qwen, there's-- install a bunch of different versions of the models, have varying capabilities, but also varying computational loads.

[00:14:25] Jay: Sure. Yeah.

[00:14:26] Brad: There's a future where people can really pick and choose the things that have the right load profiles that they need, it helps to solve some of these challenges 'cause, my Duke energy bill has gone through the roof, 'cause there's data centers popping up near me, and it's not pretty.

[00:14:41] Jay: Yeah. No, it does feel like that's the way we're moving, and it also feels like one of those things that the industry has been established, so now as the, you know, titans begin to be overpriced or whatever, the next crop will come up and it'll be cheaper for a while, and then it just will just keep looping forever, right?

It's just like you, you just keep getting the one that's good enough for now, which it seems to be. But I like the local model idea, and I think, yeah, if you get to Opus 4-6-4-7, like we think that's probably good enough for right now.

[00:15:11] Brad: The economists would tell you that it's, you can't escape supply and demand. It's an course, right? And like, as the-- if you push that price curve up, like you're gonna impact demand, and that's just the way it is right now. Supply is inflated because price is held down artificially.

[00:15:28] Jay: Yep. Yeah, it's, it'll be very interesting to see how it plays out, and I love how, like, everybody's gonna be wrong. We're all gonna be wrong. We have no idea. We have zero idea of what's actually gonna happen, and they're just like, "It will... We'll see. We'll see." Oh, do you care about personal brand at all?

Do you put any stock in it? Like, I know you're kinda like the face of the company, the CTO. You're like the, you know, the brains of the operation when it comes to, like, the tech side and understanding all this technology. Do you put any stock to, like, kinda your own personal image outside of Seven Hills, or is it all just kinda the same stuff?

How do you view personal brand?

[00:16:00] Brad: It's always been one and the same to me, for better or for worse. I think, I put less stock into it than I probably should, and it's something that's been a little bit more on my mind in recent years is like I should probably, invest a little bit more time and energy into that. And it's not that I've made time for yet.

[00:16:18] Jay: Yeah. Me, it seems like this is not true, but me either. I hate... It took me forever to even, like, understand what it was, but, yeah, you're kind of tied to your business until you decide specifically not to be, you know? It's like everybody sees you for that 'cause that's your baby, that's who you are.

All right. So, if you had to start over today, kind of everything, lessons learned, et cetera, et cetera, no infrastructure, just the knowledge that you have, what would be step one to starting Seven Hills over today?

[00:16:48] Brad: Find a salesperson. We under-invested in sales from day one. And that, hard 

[00:16:55] Jay: What type of salesperson? A young person or an old person with experience with a book of business? What would like describe the ideal salesperson?

[00:17:01] Brad: For us, probably young, energetic. We've got a guy here now, his name's Charlie. He's fantastic.

[00:17:09] Jay: I know Charlie. Charlie is great. I can vouch for that. Okay.

[00:17:13] Brad: Yep, he's absolutely great. And it's just somebody that needs to be able to be out there and be the things that I'm-- like, we talked about personal branding. I'm not the guy who just knows everybody and can remember everybody's names and remember who I talked to and work a room at a conference.

And they're all the things that, our business spreads so incredibly by word of mouth and by networking and relationships. Like, we're not having, we're not having success picking up the phone and doing 100 dials a day, cold calling people. Like, that's just not realistic in a service business. Maybe in a SaaS product company that works well, but that's not us.

So we need people... We went years and we're kind of sitting around like, "You know what? We're, we've been in Cincinnati for years now and nobody knows who we are except for the people who've worked with us." And that's what Charlie's changed for us. Now I go to an event and I tell them who I'm with and like, "Oh, I know Charlie."

[00:18:02] Jay: That's it

[00:18:04] Brad: That's success. He doesn't have to close every deal or anything like that, but he has to get me to the table more often.

[00:18:10] Jay: I love that. I mean, I tell this story all the time. There's a guy that lives in Rochester, New York, and his name's Ben Albert, and he said, "I wanted to be the business connections guy of Rochester, New York." And I thought, "That's the dumbest thing I've ever heard." And 

[00:18:25] Brad: It's so 

[00:18:26] Jay: then it sat with me for, like, months, and I'm like, "That's the smartest thing ever."

I'm like, "Why do I not own my backyard in Philadelphia?" And I took the same approach, and I was like, if people... Like, another guy told me one time, he said, "I'm okay if people don't want to work with me because they don't like me or they don't like my work." He's like, "I'm not okay with it if they don't know who I am."

I was like, "That, I like that." Like, you just... If people know who you are, they know you do good work, You know, I love that approach, especially kind of in your area, right? Own your backyard is one of my favorite things. So, I have one more thing for you. This is non-business related.

[00:18:57] Brad: Cool.

[00:18:58] Jay: If you could do anything on Earth and you knew you wouldn't fail, what would it be?

[00:19:02] Brad: Ah, man.

Yourself

available? Hmm

[00:19:07] Jay: That's why I saved it for last.

[00:19:09] Brad: Yeah, that's a hard one. Just launch a bunch more businesses for fair answer.

[00:19:13] Jay: No, no. Boo. That's not Brad, maybe it is Brad but that's not the Brad I'm talking about. What would Brad, what would Brad really wanna do but, you know, maybe fear is holding him back of just some ridiculous physical feat or, something, I don't know, solving world whatever, you know? It doesn't have to be altruistic.

I've gotten all sorts of answers.

Yeah. I don't even give people examples. I just gave you help too. I don't even-- I never even give a I just give you phone friend.

[00:19:38] Brad: This is a really hard one for me 'cause, like, my thought process doesn't really leave too far beyond, like, Seven Hills and additional ventures. And I can give a little bit of background to, like, justify that answer a little bit, is that I find myself in a situation where I have, three kids that have special needs, and so my, like, driving purpose is largely around, being able to sustain not only myself but also potentially, one or more of my children for the duration of their life.

So I'm a very coin-operated, if you will, or just, like, money-driven. And so, like, the idea of operating different ventures, so, like, it would be things, like what we've talked about, and the reason that that comes to mind is, like my oldest is, the most impacted by autism and is probably going to struggle, just being independent in life.

But, you know, he wants to do things. He wants to work. He wants to be independent. He's already fiercely independent, to a difficult challenge sometimes. And, like, I would launch the places that can employ him and give him a place to be, and that applies to all of them.

[00:20:42] Jay: That's a great... Now that's a great answer. That's a great answer. All right. I will... 

[00:20:46] Brad: Okay. We'll accept that

one. 

[00:20:47] Jay: You won me over. All right. I thought you were just gonna be like, "I wanna be cool business guy and

[00:20:52] Brad: So there's like some self-interest in there, there's some altruism in

there's 

[00:20:56] Jay: You painted it with enough, you know, care and altruism that it counts.

That was good. I like that. All right. Brad, if people want to reach out to you directly about something they heard today, how do they reach out to Brad himself?

[00:21:07] Brad: brad@sevenhillstechnology.com probably the easiest way. I'm pretty findable on LinkedIn, despite my disdain for it. And, so there goes that personal branding, right?

[00:21:17] Jay: Hey, we're gonna get you. I-- We'll work on it. We'll get you there. All right, man. Brad, you're awesome. I'm glad we got to spend some time together. Genuinely a great dude, and I love, like, your passion for all the stuff you do and, you know, thanks for being on. I'm gonna keep hanging out and bugging you and asking you AI stuff 'cause you're my, three in the group.

No- only people at 7 will know what that means but,I appreciate you, brother.

[00:21:37] Brad: in the room, the whole staff's here, so

[00:21:38] Jay: Yeah, I got... I got my ha- I got my hats, my kids the Durham Bulls gear. We're all good. So Brad, appreciate you, brother. Be good and, talk to you soon, all right? Thanks, Brad.

[00:21:45] Brad: Sounds good. Talk to you

[00:21:46] Jay: See you, man. Later.

[00:21:47] Brad: Bye 

​