Latest / Talk Commerce / Business Profitability Through Mathematical Systems with Adam Callinan

Transcript

Business Profitability Through Mathematical Systems with Adam Callinan · Talk Commerce

• Brent Peterson (00:01.518) Welcome to this episode of Talk Commerce today of Adam Callanan. Adam, go ahead, do an introduction to tell us your day-to-day role. Tell us one of your passions in life. Adam Callinan (00:06.094) You Adam Callinan (00:14.269) I mean, my number one passion is my family and kids. You'll see in the, we get into the story, most of my life revolves around that. My operational business passion is a company called Pentane. My last company was a consumer product company called Bottlekeeper. That's way more fun to talk about. So I'm sure we'll spend time down that rabbit hole, but yeah. Brent Peterson (00:34.444) Awesome. So Pentane, Bottlekeeper, tell us a little bit about those two different companies. One was a product company, and tell us about the new one. Adam Callinan (00:48.578) Yeah. Bottle keeper was a consumer product company invented on accident on a beach, drinking a beer out of red party cup in 2012, my cousin. And he invented this product to keep a beer bottle cold. If you can see if you're watching video, it's our prototypes right over my left shoulder. did you just unscrew the bottom stick of your bottle inside, keeps it cold for way longer than is generally necessary, but protects glass in places where it's not a great idea. I had just sold out of a medical tech company. And we came together to do that. And I vowed to never, ever build a big heavy business again. The result of which is we built bottle keeper really lean. put some guard rails in place early on around employees. wanted to build a scalable business that had no team, no internal team. And in 2013, you could do that 2014. And so we did that really successfully. got to 8 million in three years with no employees or investors. And then scaled from there into tens of millions. went on Shark Tank when we were way too big to be on Shark Tank in 2018 and got acquired by a private equity group in 2021. And even then we only had four team members. So we're really lean. We were really profitable. And Pentane comes from that experience. Pentane is sort of the underlying math and systems that we used to execute that outcome. Brent Peterson (02:08.036) That's awesome. Perfect. Adam, before we get started talking about, about Pentane and some of the, maybe the AI topics in there, you have graciously volunteered to be part of the Free Joke Project. And I'm just going to tell you a quick joke. All you have to do is give me a rating, eight through 13. This will be one your kids are going to like. Here we go. What do you goop? What? We'll start over again. Sorry. What do you get when you cross a fish with an elephant swimming trunks? Adam Callinan (02:18.819) you Adam Callinan (02:27.278) Perfect. Adam Callinan (02:37.614) Nice. I would give that an 11. My six-year-old daughter will deeply appreciate that. Brent Peterson (02:38.744) Yeah, was, it wasn't, it wasn't. Brent Peterson (02:44.676) Yeah, absolutely. All right. number one, that's an amazing story. I'm in an entrepreneur's organization. I just got done with GLC and Honolulu. And I love to hear those kind of stories. I think in today's world, it's much easier to be lean, right? Especially with leveraging some of the AI things. In your show notes, you had authentic Adam Callinan (03:06.508) Yeah. Brent Peterson (03:13.4) listed a number of times. Tell us about how authentic plays into the role that you're doing in Pentane. Adam Callinan (03:22.326) I think that my bet, if we back up, is that real human and authentic is going to become a lot more important in the very near future than it has historically been. But that being said, much of the... The success that we had at Bottlekeeper, was sort of like part of it was luck and timing, which is always going to be the case 100 % of the time. Luck and timing is always going to be a thing in there. Part of it was we were really good at math and solving problems, building puzzles, and creating solutions to things that we needed as the operators of a business that had no employees. And another part of that was that we were extremely personal with how we interacted with our customers. And that was my, when Matt and I started that company, we split the company in half. He was responsible for the backend, which was our financial stuff, our inventory, our shipping and fulfillment dealing with that manufacturing and all that. And I was responsible for everything on the front end. So that was anything a customer would see the website, the copy, the content, the creative, the visuals, the videos, the customer service, the email marketing, the paid media, all the stuff. I wanted to. communicate in a way with the customer, knowing that it was automated. That was the whole thing. If we were going to have no employees, we had to automate the heck out of everything. But you can do that in a way that is still personal and extremely authentic. And it was important then, and this is way before AI. This is like 2013 through 2021. Now, as we look at AI and you see this, I if you spend more than 10 minutes on LinkedIn, you quickly notice that half the comments and half the content are clearly not written by a person. And you're starting to see the knee-jerk reaction to that type of content with that, what I will continue to say is it will be a knee-jerk reaction back to being really human and really authentic, which is why I now do a podcast, which is why now I put myself out there far more than I historically have as a very private person, because I think it's important. Adam Callinan (05:36.286) Now with AI and the direction that all those incredible things are going, that being human and being real and making mistakes and breaking things publicly is important, increasingly important. Brent Peterson (05:51.074) Yeah, I'm making a similar bet. We just bought a company that has human writers, 40 human writers. And so it's not a lean operation, but it's certainly making that bet. I do feel like humans are going to be a factor, like you said earlier. People identify that content as being AI written, and it's going to turn people off. Tell us about how you think that's going to change in Adam Callinan (05:57.304) Amazing. Brent Peterson (06:20.226) what you're doing in maybe in the back end of the business, but also just from a marketing side. Adam Callinan (06:27.956) In the backend of Pentane, despite the incredible advances that have taken place with coding tools and systems and the really amazing blocking and tackling you can accomplish there, Pentane is built on a lot of math and a lot of really complex math. with the purpose being that it simplifies that for the user. It takes in all the math in a business. It takes in all the information in a business, applies all the math, and then tells you what to do to be profitable, to get to breakeven, to get to a 10 % margin, whatever it is for you. Using AI systems to code math is treacherous because AI is not that good at math right now. It will get really good at math, probably. It's not that great at math right now. It operates at about an eighth grade math level. you know, that's changing on a daily basis. So that's probably an outdated statement in a month, but the reality is today it's not. And the last thing that we need is an AI to go into a code base and make an adjustment to something unintentionally. The result of which is that it's giving our user incorrect guidance on decisions to make about their business. Because it's really, again, it's giving really tactical, like go do this specific thing, increase your ad budget to this at this specific return on ad spend. That can't be wrong. So. Very, we have to be very careful in reticent about using AI from a code based standpoint. I do think that there is really significant and interesting use case for looking into data that is coming in via API's across, example, marketing systems with meta and tech talk and Google ads and these things, and being able to use AI to pull interesting insights from those. but we are, we are primarily focused today, although I use AI a lot personally and professionally, we are the system we are primarily focused on building and underlying durability on with conditional logic and math that we will then later be able to AI things when it makes sense and not just bolt an AI thing on to call it an AI thing, which is what I feel like most people are doing right now. Brent Peterson (08:32.1) So Pentane, I'm familiar with EOS, Entrepreneurs Operating System. Is this in conjunction? I'm assuming this doesn't replace like a framework, a business framework that for operating this, would this enhance it or tell me a little bit about how Pentane works? Adam Callinan (08:36.75) Hmm? Adam Callinan (08:48.43) It would definitely enhance it. And I'm also familiar with EOS and actually the first company in Pentane had gone through EOS, the longer story there, and they've been using Pentane for two years very, very successfully. Pentane is, it is not an accounting system. It is not a replacement for QuickBooks. There may be a play there at some much later date. That's not an interesting, helpful thing to me. Pentane is the, how do you operate your business? system. It is the we're considering hiring a new person and that person's going to cost $10,000 a month. How does our revenue need to change to pay for that? And how are we going to change our revenue? Do we need to increase our ad budget? Do we need to get more efficient? Do we need to increase our pricing? Do we need to do some website optimization and increase our conversion rates? Those are all things that Pentain tells you exactly what to do. So it's much more of a how to make money system. Like let's use math. to make decisions and remove emotion from the equation. Because at the end of the day, profitability is a math problem. Like it really is just that simple. Now there are lots of artistic components in driving profitability when you get into the creatives and whatnot, but profitability is a math problem. Brent Peterson (10:06.274) Yeah, I think I experienced that in a company we sold where we had cash flow projections of what you think you need in three, five, six months or whatever that is. And trying to match that, read the case study that you had in your website for the business that wanted to grow and they struggled on their P &L, right? And making sure that there was enough there. And I certainly have spent time with owners who would wake up and look at their QuickBook balance and saying thinking okay I'm okay today right and they don't think about what's what's it to be like in in three months so it sounds like Pentane helps to solve that problem right not just what it's good what it's going to be in a month or two from a sales standpoint and a profitability standpoint Adam Callinan (10:41.55) you Exactly. Adam Callinan (10:57.024) It does, it solves it in the context of are the things that you're doing today and this week and this month, are they adding to the business or are they taking from the business? Are they adding fuel to the engine or are they burning fuel in the engine? Now at the end of the day, that fuel is profit. have to, unless you're just gonna raise money forever, which some people do, unless you're gonna do that, you have to have profit in the business. Coming back into the business to add fuel to the engine to move forward. Now getting into Cashflow projecting is a whole other thing. I mean, frankly, that we're building because that's an important tool as inside of Pentium, can literally take a lever and ratchet up your ad spend and see how it changes the entire business. And that clearly has an impact on inventories. If you're selling a physical product, it has an impact on cashflow as, know, generally when you make money, there's costs associated to that make money. There's sales related expenses called variable expenses from an accounting standpoint, but. It's like your cost of goods and your shipping and fulfillment and the credit card fees and all that stuff that goes with selling product. So getting a full picture of what's happening in the business is the next big end goal until we kick the goalpost down the road perpetually and just keep adding and adding and doing things that our customers are telling us they want. Brent Peterson (12:17.07) Talk a little bit about using an LLM for your data as opposed to using an algorithm that you've written for your data. And I kind of go back to those mistakes that are made typically by LLMs. Adam Callinan (12:30.828) Yeah, I mean it. Right now, and all the AI super fans are gonna hate me for saying this, right now I don't need AI to solve really, really important problems for a $10 million a year e-com brand or a $1 million a year e-com brand. We can answer critically important questions with 100 % accuracy using algebra. and a little bit of calculus. Now where the LLM component, I do think it's really interesting is when we look at massive data sets because of the way that data gets into Pentane via APIs. So it's like you click the buttons and it connects to your Shopify and your Amazon and your Amazon ads and your Google and your Meta and all these things. Those APIs can have a really broad reach and we're not using the full breadth of their reach right now, but getting down into being able to pull insights and data. at an ad level, know, goes like campaign ad set ad level across the base of a business, across all of their advertising campaigns, using AIs and LLMs to be able to pull interesting things from that and to provide some guidance. think that's where things get really interesting. But that's, I don't know that that's necessarily saying like, like we can with math say like, here's the answer to the question, go do this exact thing. Like here's the number to the hundredth of a point, like go do this one exact thing. which we can do a lot more safely with math today. Brent Peterson (14:07.534) I'd like to kind of continue on this conversation about, my experience with LLMs has been a lot of math errors and people would argue, no, I would argue yes, it does give you math errors. And it's very good at apologizing too when you point out the error, but it doesn't help that it's made a mistake. The other side. Adam Callinan (14:21.582) Yeah, I know. with high certainty. It answers the question with high levels of certainty and then apologizes when you tell it it's like it literally can't do addition, correct? Brent Peterson (14:30.5) Yeah. Right, exactly. But then the other idea is that keeping things on track, right? So I think, I'm assuming Pentane keeps you on track in what you're trying to do. And I can say that I've used Claude code to help me to develop an app. And if you don't, if you're not on top of it at all times to monitor the progress, Adam Callinan (14:44.803) Mm-hmm. Brent Peterson (14:57.12) It is like having a group of really talented school children code your app to suddenly have it add another phase to it that you didn't even ask for. tell us a little bit about the guidance and how important that is. Adam Callinan (15:14.072) I mean, the example that you just gave is why we can't use, I argue in a very loving, positive, productive way with our development team constantly. Because look, my, clearly from my history, I'm a very efficient operator. Like I want it as lean as humanly possible. I want to squeeze as much water from the rock as physics allows. AI coding development seems like a great way to do that. Like we have a team of developers. If we can get... three teams of developers worth of code from one team of developers, why wouldn't we want to do that? The problem, which you just perfectly exemplified, is that the AI doesn't just go in and do the one thing that you're trying to get it to do. It kind of does that. Maybe it's correct. Maybe it's not. have to double and triple and quadruple check it. And then it maybe goes and messes with this thing over here that's completely unrelated. And again, like this is, we're layering math on math and doing complex things. That can't happen. So it's just not to a point yet. mean, I think in a silo, there are incredible use cases for it that will exist and continue to develop inside of Pentane. One of them that we're working on and have been playing with is when a company comes into Pentane, they connect their QuickBooks Online account and it pulls the necessary expense data out of their QuickBooks Online to create this financial understanding of the business. Like, how does it make money? How does it spend money? Because that is... the basis for answering questions around how to be profitable and intentional. And when it does that, there are line items in that the system, our system may not have because this company calls this specific expense a certain thing that no other company that we have, and maybe it's because there's a typo in it. So it's like, those are really interesting use cases to use an AI to start helping categorize that thing that the user still has to click, yes, that's correct. So, I mean, there are a lot of interesting little use cases like that. We just have to be extremely careful that it doesn't, it can't even get into the room that has math and calculation inside. Brent Peterson (17:20.388) Tell us your typical endpoints. You mentioned QuickBooks, you mentioned probably Google Ads. Are those your kind of metrics you're working within to help clients build and grow? Adam Callinan (17:31.598) Yeah, the typical e-comm business, you know, is on QuickBooks online. You don't have to be, but that's one of them. Shopify, Amazon, they're advertising on Amazon ads, TikTok, you know, maybe Pinterest, Bing ads, Google ads, Meta. They're using Google Analytics, which is awful, but it's what we have right now. so those are all things that when they go through the seven minutes of onboarding or however long it takes, it's fast. they click all those things to OAuth that data into the system. it's already the second they're doing that, it's already doing math in the background. So by the time they see their system, it's a fully functional system. It's already answering questions. They input, we call it a mission objective, like what their net goal is, if it's to get to breakeven or generate a 12 % margin or whatever. And then it starts telling them exactly what to do and gives them the means of tracking and make sure as they go and execute that the things are happening the way they think they are. Brent Peterson (18:28.03) so I like to ask people kind of what they where they think that e-commerce is going this year and what is your predictions on where you're going, where e-commerce is going this year. Adam Callinan (18:41.482) Well, I would be remiss if I didn't start that off with tariffs. Love them or hate them. don't think anybody loves them. the setting aside political components there, tariffs are starting to wreak a bit of havoc largely, you know, for because of the unknowns, I think right now. But businesses are trying to figure out what their cost basis is going to be and maybe shift around manufacturing into places that are where they can. positively impact that cost basis. mean, we're building a part of the system right now to be able to make adjustments to your cost basis and see how it impacts the business. You can literally go in and add, you know, a 10 % tariff or a 22 % tariff for, you know, in the case of China, maybe it's higher and see where your marketing need all those things need to change from an e-commerce overall business standpoint. mean, what AI and we are obviously at the very beginning of it. What AI has the capacity to do from an from a efficiency standpoint is wild. mean, I think that AI is not gonna replace someone building a physical product, but AI, I mean, probably within this year, will be at a place where it can run agentically with an AI manager running a of, boatload of agents, can run most of the business without having a human directly involved. And that is incredible. I mean, it's nuts to think of that. Brent Peterson (20:16.196) Yeah, that's such a good way to look at it. Agentic, I've been at a couple of these big tech conferences, shop talk, and agentic was the buzzword. I went into that thinking, what the heck is agentic? And I was like, oh yeah, duh. So Adam, we have a couple of minutes left here. As I close out the podcast, I gave everybody a chance to do a shameless plug about anything they'd like. What would you like to plug today? Adam Callinan (20:23.714) Yeah. Yeah, of course. Yeah. It's the keyword right now. Adam Callinan (20:31.286) Ahem. Adam Callinan (20:42.808) Well, you can definitely find us at pentain.com. That's easy. That's straightforward. I will shamelessly plug my podcast, Growth Mavericks, which is about, it's part about operating consumer brands and businesses. It's more about the things that we need to do to become mentally and physically resilient and anti-fragile to survive the wild ups and downs that come with entrepreneurship. Because we are often like, just around the corner from that big break. It never happens the way that you think it's going to, so you can't plan for it. And companies fail when people quit. So how do we keep like our brains and bodies in the game for long enough to be able to survive until we get that big. Brent Peterson (21:28.884) Yeah, that's perfect. And people quit, meaning either the owner quits mentally or their key employee quits, right? Adam Callinan (21:36.706) Yeah, sure. And there's a lot of reasons for that, but that's generally what happens. Brent Peterson (21:41.964) All right, Adam, it's been a great conversation. Adam Kalanand, Kalanand from Pantane. Thank you so much for being here today. Adam Callinan (21:46.766) Thanks, Brent. Appreciate it.