Ep. #937 - Wartime vs. Peacetime for Startups
In this episode of Startup Hustle, let’s talk about the implications between wartime and peacetime for startups. Matt Watson and Ryan Sevey also talk about machine learning and artificial intelligence. Our guest is the CEO and founder of Mantium, a company included our Top Startups in Cincinnati list this year.
Covered In This Episode
It’s no secret that AI and machine learning have seen immense growth throughout the years. But what do these advancements mean for businesses, especially for startups? Will data play an important role in the future?
Matt and Ryan’s conversation answers all these questions. They also discuss how wartime and peacetime can affect startups and what to do during such situations.
What are you waiting for? Tune in to this Startup Hustle episode now!
- Ryan Sevey and his journey (02:02)
- What is machine learning? (04:29)
- Use cases for Mantium tech (07:00)
- On text extract and translating it into data (09:20)
- All about hugging face integration (12:36)
- How does Mantium work? (13:23)
- Using AI as a developer and where to start (15:59)
- On search engine optimization tech (18:37)
- Problem statement in business (20:08)
- The difference between wartime and peacetime (23:50)
- What to do during a recession (26:49)
- CEO transitions from peacetime to wartime (33:34)
- Benefits of a recession (36:23)
- Words of wisdom from Ryan (41:49)
The strategy now is more of a “slow and steady win the race.”– Ryan Sevey
We predict that a recession is actually a pretty good catalyst for all AI companies.– Ryan Sevey
It’s one thing to make it work. It’s another thing to make it work in production. And make it work every day with high availability and all these things.– Matt Watson
Build a team that can help develop and successfully deploy your tech product. Full Scale has the right people and efficient platforms to help you achieve it. Just map out your technical needs and be automatically matched up with a fully vetted team of developers, testers, and leaders. Take advantage of the platform and ease the stress of recruitment now.
On another note, aside from supporting the startup community, our podcast partners also offer services that your business may need.
Following is an auto-generated text transcript of this episode. Apologies for any errors!
Matt Watson 00:00
And we’re back for another episode of the Startup Hustle. Today, I’m excited to talk to Ryan Sevey. And his company’s name is Mantium. And they do a bunch of cool stuff with artificial intelligence, machine learning, and all those things. So we’re gonna learn all about that today. Also excited to have him as part of our series of the top startups in Cincinnati. So big congratulations, Ryan, for being recognized as a top startup in Cincinnati. In our show notes, you can find a link to listen to other episodes and learn more about all the other top startups in Cincinnati. One of my favorite things we do is highlight different cities and cool stuff going on, especially in cities that don’t get a lot of attention. So, before we get started, I do want to remind everybody that today’s episode of Startup Hustle is powered by FullScale.io. Hiring software developers is difficult. Full Scale can help you build a software team quickly and affordably. And has the platform to help you manage that team. Visit FullScale.io to learn more. Ryan, welcome to the show.
Ryan Sevey 00:57
Awesome. Thanks for having me, man. Excited to be here.
Matt Watson 01:02
So, as we get started here, love to first just learn more about your background. You mentioned you’ve founded a couple other companies before. It seems like we all have become serial entrepreneurs. I feel like it’s almost a disease that we get, and we can’t help ourselves. So I’d love to learn more about your background and how you got here.
Ryan Sevey 01:23
Yeah, I think I’ve always just been an entrepreneur. I grew up in a town called German Town. If you know where that is, it’s maybe 25 minutes away from Cincinnati on an apple orchard. And when I was really young, I remember I would go out and harvest these different materials to make birdseed. And that was, I guess, kind of my first business. I got enough money to build my first computer. Then I started building websites when I was younger, then I got really interested in cybersecurity. So I did that for a while. It’s where I met my two-time co-founder, Jason Montgomery. We were both at American Electric Power. And while we were there, our first company was formed called Next Osis. And that was back in 2014-2015. And it was because we had all this data from the power company. And back then, machine learning was being talked about as being really good at finding patterns and datasets. We had tons of data. The problem was all the solutions out there at the time were really geared at data scientists. In our background, we thought it was really intimidating, quite frankly, to try and use technology the way that we wanted. So Next Osis was an auto ML platform geared at software developers. Ultimately, that was acquired by Data Robot in 2018, I think. And then Jason and I and the whole crew at Next Osis went on to work for Data Robot for about three years. And as we were coming up on our time served at Data Robot, Jason and I were really intrigued by what was happening in the large language model space. Companies back then were really open. AI was kind of the main large language model provider. They were getting all the marketing attention and all the news coverage. And we thought that was really, really interesting for a variety of reasons. One of them is you didn’t need a data set to get started. And this fascination with what LLM promised really gave rise to why we started Mantium.
Matt Watson 03:34
Well, for those who aren’t really familiar with AI and machine learning and all that stuff, maybe it’d be good to help give them the basic one-on-one on how that stuff is used and common use cases of it.
Ryan Sevey 03:50
Yeah, so machine learning has lots of different use cases. It can be used to predict the price of a house. Like if you look at Zestimates as a good example for you on Netflix, and they give you recommendations. That’s another example of machine learning. But we’re specifically focused on large language models. You can think of stuff like content generation. So you might give the model a couple sentences. And then, based on the sentences that you gave it, it’ll write the next sentence. It can write ad copy as an example or marketing material, and it can assist with that type of stuff. The other uses are around sentiment analysis. It can do some classification type of task, right? Like is this sentence happy, sad, or neutral? We have a customer that’s using us today to look at whether or not marketing material might be interpreted as racist, okay, by different groups or, you know, is this overly leaning towards like a male and you’re ignoring females perspective on this. So that’s a pretty interesting one. But then we also have customers that are using it to look at documents such as invoices, and then they can use the AI to ask questions about those invoices. So, for example, you could say, Hey, show me all the invoices that were over $10,000. And you’re doing this in a very natural language type of way, versus creating queries and things of that nature. So that’s a high-level overview. Maybe the other important part talked about why MLMs are so fascinating is because they’re trained on massive datasets. We’re talking terabytes of data. And then that gets translated into something called parameters. Not to get too technical, but GBT three, which is open AI, that’s a 175 billion parameter model. And there’s a debate on whether or not the more parameters you have get better performance. We’ll see what happens with GPT. Four, there are some rumors that it’s actually the same size, but they figured out how to make it perform better performance. We’ll see. That’s right around the corner, or so I’ve been told from the rumor mill, but hopefully, that helps give a little bit of flavor. When we do train these models, it’s just vast datasets typically from the internet.
Matt Watson 06:15
So for Mantium, unless I’m listening today, I’m like, Oh, I have some software. And we could add some AI to it from a lot of what you just said. There sounds like there’s a very wide array of use cases. And you and your platform can help with all of those use cases, or they’re just very specific ones that your company can help with.
Ryan Sevey 06:36
Yeah, so Mantium is really aimed up? How do you get these models into production? So you might start playing around with something like an open API’s playground, but then you’re gonna run into a challenge of, well, I’ve made something interesting here. But how do I take what I use and put it into my organization or into my business, and that’s where Mantium really shines is, let’s say that you made a content generation type of prompt or AI, where you can put that in Mantium, then we actually spin up a whole application behind that, or you can come and interact with it. On the enterprise side, most of our customers are doing things with invoice processing or document intelligence more broadly. So they might want to run PDFs as an example, like 100-page PDFs, these presentations, and they’re trying to extract out maybe four or five key insights. So without us, you’re having staff spend however many hours going through 100 slides, verse, you can just come to us, you say, hey, here are the four things that I want. As an example, you can say, I want to know what the growth rate is. I want to know what the ARR is. I want to know what they’re spending on it, insert whatever else, and then the AI will read those 100 pages, and then it will just tell you the answer.
Matt Watson 07:58
So I’m a developer myself. I’ve been involved for 20 years, right? And so I think one of the biggest challenges with doing this kind of stuff is it does not necessarily have the raw data. It’s building all the stuff on top of the raw data, right? And so it sounds like to me, if I’m following you correctly, it’s like one thing to dump a bunch of PDFs, you know, basically in like a database or data warehouse, like I have all this data. But now, what do I do with it? And it sounds like that’s part of the problem that you’re trying to solve is like, Okay, well, that’s great. You can run this machine learning model. But it’s like, you need to run it like a million times. And you need to have a database that almost sits on top of that so that when you want to ask questions of it, you have a way to actually ask questions of it. Right, instead of building all that infrastructure yourself. Exactly. Is that one way to look at it?
Ryan Sevey 08:45
So we’ve partnered very, very closely with AWS. And maybe to go off this example. AWS has something called extract. And if you’re not familiar with extract, is that basically you give it a PDF, and then it will give you the insights out of that PDF, like here was the invoice total, as an example. So that sounds great. Much To your point, when you actually look at how you spin up a text track is pretty complicated. Like, right?
Matt Watson 09:11
I mean, what do you do with the data that comes out of it? Well, exactly. So how do you get the files into it? How do you get the files out of it?
Ryan Sevey 09:15
What do you do with that, in this case, is typically a JSON response. Right? So that is a big, big, challenging question. With us, we basically abstract all that away. So you get to select, hey, my files are in SharePoint. So we have a SharePoint integration as an example. So then we start looking at just your existing infrastructure, in this case, SharePoint. When you drop in a PDF, we basically make it like magic to you. We consume that PDF, and we send it through the extract. And then we return a JSON string, but then that gets translated into either a CSV file, so maybe you don’t want it to go anywhere else. You just want a CSV file. Hey, here are all the invoices that have been extracted with the Total invoice number, the date, etc. Now you get this nice CSV file, or we have other customers that are using more like Intuit think there’s a company called SPS, but enterprise resource planning and employees resource management type tools, we can send that data directly into those existing platforms. So, in the end, you’re actually now using AI. It’s not just a model, and people are seeing real value. It’s no longer this perceived. Okay, well, what do we do? We built this model. How do we implement it? That’s really what we’re solving with Mantium?
Matt Watson 10:36
Well, it sounds like to me, I would guess your target audience is a company that maybe they don’t have a lot of software developers themselves. And you guys make it easy to say you send us all the data. And basically, we have the platform that does all this stuff. But May had a lot of software developers themselves, they could do the same thing, it’d be a lot of work, and then don’t have to maintain it, where you guys provide kind of the easy button to like, hey, we have a platform that just does all this. You don’t have to build it.
Ryan Sevey 11:03
Yeah. And I think what’s really interesting for us is we have customers on both ends of the spectrum. We have customers that have robust data science teams, and robust engineering teams. And they’re still using us because setting all this up is extremely infrastructure behind it. Yeah, it’s just like, so we hear from data science teams all the time, like one of the other big use cases, and this is a little bit more technical is let’s say that you build a model. In your data scientists, well, now you need some kind of interface. Typically, to interact with that model, you don’t want to sit there on a notebook. And that gets a little bit cumbersome and not super awesome in user experience. But going to your software developer team and saying, Hey, I need you to build this front end interface for this model. They’re gonna say, okay, cool. Here’s our backlog, throw it, and maybe in six months, we’ll get to it. Yep. But with Mantium, you could take that model, we just completed the hugging, face integration. So if you register the model on hugging face, it will show up and Mantium, you literally click one button, and we create a spa or a single page application. So you get your friend interface for that model that you might have created or fine tuned. And now you’re no longer waiting for your development team to figure out where in the backlog, they’re going to figure out how to make you a front end interface. So we really span that gambit.
Matt Watson 12:18
I totally get it right. Because it’s like, even if you make all this stuff work, it’s another thing to then make it work in production and give it the care and feeding and, and scalability and all the other crap to keep it running, right? Like, it’s one thing to make it work. It’s another thing to like, make it work in production and make it work like every day with high availability and all these things, right, and you guys just provide all the tools, it sounds like to just make this shit work.
Ryan Sevey 12:42
We do one thing that we found in the market, which I do, so I’ve been doing this for a little bit in the AI space. And traditionally, all AI companies asterisk most AI companies want access to the data, right? Like, if you look at pretty much name an AI company, they want you to use their platform, they want you to take your data, integrate it with their platform, we take a different approach. Mantium is a control plane. But if you have your own infrastructure, you can have your own data plane. So we actually never see your proprietary data, especially if you’re an AWS customer, all your data could just live in your ATS account. And then Mantium sits above it as that control plane but your data never leaves your trust boundary, your data never actually comes into Mantium. We did something interesting in order to control the plane which is really resonating with the enterprise customers.
Matt Watson 13:41
I can see how that would be really important for people that care about security, right? Like, like you said earlier example of scanning invoices and all these things. And, you know, most big companies are very secure and security sensitive these days. And
Ryan Sevey 13:54
I think it goes even farther than that, though, to a more macro strategy of if you believe that every company to be competitive has to figure out how to become an AI company, which I believe is very similar to maybe 1020 years ago, where every company was trying to become a software company. Right? Like Amazon’s a great example, though, Amazon sure they’re a retailer. But they were really good at figuring out how to use software to make things more efficient. Now, we look at, you know, 2022. I think that same trend is going to happen, but instead of becoming software it’s going to be how do we become an AI company. And those that figure that out are going to be the winners. I think that’s really how you’re gonna start seeing disruption in the broader macro sense. And those that don’t figure it out are probably going to suffer some pretty bad fate.
Matt Watson 14:45
Well, I think this is a great topic. And it’s something I think of every day actually because I’m basically the chief technology officer of a company that’s a digital marketing agency. And so we do a lot with Google AdWords and all this stuff. And, you know, our goal is to help our customers, you know, optimize their spend to get more leads, you know, all that kind of stuff? And it’s like, yeah, how do we use AI to optimize this stuff? How do we use AI to optimize their budget and, and moving the budget from, you know, one spin to another spin and all those things. But the challenge for somebody like me is trying to figure out like, Well, how do I even do that? How do I, I don’t know shit about AI. Are what you know, all the other machine learning languages, like I don’t even know what all of them are, like, I have no idea. I didn’t know where to start. So for somebody like me, he’s like, look, I have a business. And you’re telling me I need to be an AI? Business. I need to use AI, where the hell do I start? I know what to do.
Ryan Sevey 15:37
Yeah, and I think that’s been a common thing that a lot of companies are thinking about, as recently as a couple years ago. But what we have noticed, and I actually love this story is one of our customers. They’re a bottling company, a manufacturing type of organization. And their controller was attending a continuing education class, just you know, to be maintained, or CPA license. And in this class, they were talking about how AI can help with accounts payable. And she was thinking, wow, you know, we spend six hours a week for people, manually reading and reviewing these invoices. And now she’s hearing Oh, wait, I can automate this part of my business. AI could potentially save me six hours a week. So then they reach out to us, and then we show them how this works. And then they become customers. And I think that’s a really, really important story, right? Because that’s not a technical person. Right? This is now a data entry.
Matt Watson 16:38
Was that really the primary use case there?
Ryan Sevey 16:43
Yeah, it’s like data entry, right? It’s cool. We look at these invoices, their old process was the people that circle and read like, here’s the total amount, but now instead, the AI just looks at the document, takes it and puts it into their system.
Matt Watson 16:57
And when they’re do with it, we’re . . .
Ryan Sevey 17:00
Yeah, and honestly, one of the important things with them was the human in the loop process. So that’s another big feature of ours. So we can configure it that at the API comes back with a confidence interval, it’s they can define it, let’s say it’s like 96% or below, well, they then get a notification, they can come into our system, and they can look and see, oh, this came back below 96% confidence, they can see the bounding box in this example. Or maybe the total, the AI says 1000, maybe it’s 1001, they could just edit it. And I think that’s really important for the future. It’s how humans work with AI, it’s not a 100% replacement. How do we work together? Also, for Go ahead, I was gonna say on the Google topics, some of the stuff that they’re talking about with arch language models and AI, at some of their developer conferences, it seems like search as we do it today might be fundamentally changed. Like when you go to google.com, I think in the future, how you get results is going to be way different. And they’re already starting to make changes to AdWords, where it’s more media, not just text based. So I think there’s going to be tons of just the rate at which this is going to explode. It’s going to be very, very fast. And kind of to your point is like, Well, how do we keep up? Right? Like, how you do search engine optimization today is probably going to be wildly different in two years because of AI.
Matt Watson 18:31
Well, I love machine learning and I love AI. But I think I have the same struggle that everybody else would have. I also don’t know, I don’t know where to get started, like, how would I incorporate it into my software? And you? Do you have any tips about that? Like, Oh, I know, I could use machine learning models to help in some way. I’m not really sure how to do that, like, how do we get started?
Ryan Sevey 18:54
I think the first step is just understanding what problem you are trying to solve. Right? Like, once you understand the problem that you’re trying to solve, whether it’s data entry, or maybe it’s writing better marketing, content, whatever it might be, figure that out first, then once you have a good problem in mind, you know, it could be hey, I want to help. I want to make better copy or better headlines for my advertisements. Well, then you can start looking into well, what solutions are out there for that? Right.
Matt Watson 19:24
Do I need to go find a data scientist to help me do that and figure that out?
Ryan Sevey 19:29
The problem statement? I don’t think so. I think the problem statement is primarily driven by the business. Yeah. I don’t think you need a data scientist. yet. I think where we’re at today with AI, and just how approachable it’s becoming for some of these problems like content generation. You should be able to sign up for something like either Mantium or if you go to eight, open AI and play with their playground and get started like the barrier to entry now is so low and you’re not gonna break anything. That’s the other beautiful thing you can just start playing. And you’re gonna start seeing what is capable with AI. Now, the third step would be okay, so you’ve built something, let’s say, that might be good enough for you. You could say, Yeah, this has given me exactly what I was giving me really good headlines. Or you might now say, hey, I want to go talk to a data scientist to make this even better. So data scientists still have a role. But I think it’s going to shift to having a prototype being built by the business, the business now going to the data scientists saying, hey, how do I make this better? Shredder can be how do I make it faster? How do I make it more accurate? How do I make it more creative, whatever, then ultimately, a platform like yours is used to bring it to life and happily actually operationalize it, make it run?
Matt Watson 20:44
Exactly. You have to go from idea to we have this model, we know what we need to do. But now we need a platform to actually run this thing on and actually integrate with and actually execute.
Ryan Sevey 21:02
Yeah, and in most cases, it’s always when you think about Mantium, think, how do we get input to the model. So that’s what we accomplished, we also have the model up and running in production, and then output, right? So we take care of that whole thing, you actually get an end to end application up and running. And sometimes those applications could just be running in the background, right? It could be ingested from SharePoint, run it through some kind of OCR tech, enrich it with a large language model, and then send a JSON response to an accounting software solution.
Matt Watson 21:38
Well on for like what we do, it could be like looking at the weather and looking at what your current budgets are for advertising, and then making suggestions around like, maybe you don’t need to advertise for fixing air conditioners, because the temperature is going to be 105. Next week anyways, and people are going to be knocking on your door. Right? Like, yeah, and some of that’s common sense stuff. But the thing is, humans still have to manually go make those decisions today, right? But if you can use AI and these algorithms to automate all of that, that’s where it becomes really powerful. Just automating away those kinds of things.
Ryan Sevey 22:11
Yeah, exactly. And I think even in that case, it’s just giving more information sometimes to the human that’s making the decision, right? Like, yeah, why do you need to go look all this stuff up, but the AI is going to come back and say, Hey, I’m predicting this. And here’s why I’m predicting this. Yeah. Well, I agree with it. Or you can say, Wait, I don’t agree with it, because you’re factoring, like you’re giving too much weight to this particular variable, or whatever the case might be.
Matt Watson 22:38
Well, take a break for a second and remind everybody that finding experts, and software developers doesn’t have to be difficult, especially when you visit FullScale.io, where you can build a software team quickly and affordably. Use the Full Scale platform to define your technical needs, and then see what available developers, testers, and leaders are ready to join your team. It’s sort of like a little bit of machine learning there to match new up to developers. Visit FullScale.io to learn more. Well, I think we’ve talked a lot about machine learning and what your company does, but I’d love to switch gears now and talk about the difference between being in wartime and being in peacetime. And I think it’s an interesting topic right now, as you know, we’re all trying to figure out are we going into a recession, they’re trying to redefine what the word recession means. But we have things like unemployment that are still very low. So it doesn’t seem like a lot of people are losing their job, at least they’re not yet. You know, and it’s interesting as a founder and entrepreneur to kind of figure out like, are we in peacetime? Are we in wartime? And what does that even mean? And love to get your opinion on this?
Ryan Sevey 23:43
Yeah, I think that when we first JCI made our first company, we were starting, we’re in kind of the infancy of pretty easy money from VCs. And that got a little bit crazy, especially in probably 2000, where it just seemed everybody was getting these crazy valuations money was pretty much just available to everybody. And the side effect of that is, well, how do you retain and attract amazing talent? And that I think correlates to you, you almost have to operate as if you’re in peacetime and have these more peacetime programs because that’s what employees are going to want to see from their employer. We just got crazy, like a bunch of interesting side effects happened because of how readily available funds were sure. And a lot of entrepreneurs are being told, Hey, grow at all cost, right? Like you’re gonna be able to keep raising money. You know, you raise a $50 million Series A, like a $500 billion pre money valuation and your mindset and everybody’s telling you to grow, grow, grow. Don’t worry, you’re gonna get 100 or million dollar Series B and whatever. And then suddenly all that stopped. Right? Like, it really was just like that. And I think people have to shift and understand, whoa, wait a minute. Number one, the public markets are coming back into what I would consider reality with how they value companies, right? You’re not seeing these crazy multiples anymore. And that translates to startups too, right? Because it’s like, if you’re a VC, you’re looking for an IPO, what kind of multiples are realistic, and it’s no longer you know, 5060 70x. So that’s going to impact valuation.
Matt Watson 25:40
And then this looks like you guys raised quite a bit of money last year, right? We did. Yeah. And I would say you guys kind of raised it right in the frenzy of the market where there was a lot of money running around, thanks to the partially thanks to low interest rates, the federal government’s printing money, there’s a lot of money run around, which made to what you’re describing, right, the valuations were really high. And it sounds like you guys raised your seed round, you raised a lot of money, probably at a very high valuation. So you guys, were at the right place at the right time to some degree for that, right?
Ryan Sevey 26:10
Yeah, and in a lot of ways, we started to predict what was going to happen. And we were pretty early on in deciding, hey, growth at all cost isn’t going to be sustainable. So internally, we started looking and saying, Hey, we need to keep burn reasonable. We needed to keep our runway to be way longer than we wanted. Slash maybe what traditional wisdom back then was, I think people were saying, like, oh, you only need eight months of runway, right? Like you’ll be able to raise. Luckily, we didn’t, like we were starting to scale up, we were starting to get to that burn rate that would require us to raise, and that was going to be kind of that trajectory that I think a lot of startups started off in 2020. But we kind of put the brakes on that. And we said, Wait, SFX is slowing down. Let’s keep everything in check. And really strategy now is more of a slow and steady win the race. Sure. The other side effect is discounting booked revenue. Right? So if you’re going to have a recession, a lot of customers aren’t going to be able to pay, especially startups. Sure. Everyone’s way riskier. Yeah, that’s part of our own thinking is well, okay, when we’re signing this customer. Like, we used to sign up startups all the time, right? It’s like, cool, they’re gonna pay the bills, everything’s great. But now it’s more scrutiny on our part, a little bit of saying, is this quality revenue? Right? Are they going to be able to keep paying the bill?
Matt Watson 27:42
Enterprise accounts are more likely to pay no problem. And then you’re going to have some that are going to have a lot of issues? Like, we went through this with COVID at Full Scale. So yeah, when COVID gets announced, you know, it’s March 2020. I think in April, I mean, we lost about 20% of our revenue in one month, just boom, overnight. And I mean, we’re flexible for our customers, you know, we don’t have long term contracts with them. So we can scale up and scale down. That’s one of the benefits. And so, you know, a lot of them scaled down a little bit, a couple of them said, Hey, we were trying to get this thing going. And we just said, we just give up or like, whatever, you know, so you get a mixture of all that right, just to again, your point about quality of revenue. And then we had a lot of clients with no problem. They’re like, You know what, maybe we don’t hire as many, maybe we let one person go. But you know what, we’re in this for the long term, and we’ll get through it, whatever.
Ryan Sevey 28:27
Yeah, and I think the earlier stage companies are probably if you adjust course, you’re going to be fine. I think where this is really going to hurt us, those later stage companies that we’re raising at crazy valuations.
Matt Watson 28:42
Are high in burn rate.
Ryan Sevey 28:44
Like they have to cut costs, they probably have to raise more money, and they probably have to raise that money on a down round. Yeah. And that really sucks for all the people that they hired, because the right price of those options is going to be reflected probably on the old valuation. Now they got this new valuation. So now your employees are underwater, like it’s a very, very bad scenario. I think for a lot of Yeah, late stage companies, and you kind of know, you also don’t know if the recession is going to help or hurt you.
Matt Watson 29:07
Right. So, you know, my, you take COVID, for example, like COVID happens. And so, you know, our business loses, you know, 20% overnight, and eventually we, you know, we’ve doubled in size since then, and life is good. And but industry by industry and company by company, you don’t know what’s gonna happen, right? And you don’t know that if you sell trampolines, you’re about to be completely out of trampolines for the next two years straight. Because everybody’s at home now and they want a trampoline in their backyard. And never in a million years, you’re going to find a trampoline. They’re all sold out. Right? So if you can manufacture trampolines, or a whole bunch of other house household items, right that nobody would have expected. And the point is like sometimes you just don’t know. You don’t know how this recession is going to help you or hurt you. And this happened to my first company and I was coming up with VinSolutions in 2008-2009 through that downturn. We were selling to car dealerships, which were going bankrupt and GM and Ford and all this going to all these problems that closed dealerships. You know what they all needed to save money, and we had a way to help them save money. So next thing, you know, we were growing like gangbusters, right? And just like you go through all the mortgage foreclosures you got, if you’re the guy that’s got to go clean up a house that’s been foreclosed, you’re really busy, right? So the point is, when these things happen, there are always people on the positive sides of the negative side. And sometimes you don’t know which side you’re going to land on. Right. And so you have to take a more conservative approach until you figure out which way the trend is going to be with or against you.
Ryan Sevey 30:38
Yeah, and on that topic, I think when we look at AI, typically the story is We have tremendous ROI savings for these companies. Right, right. And so we predict that a recession is actually a pretty good catalyst for all AI companies. Sure. You save labor cost. Yeah, exactly. If you go back to what I was saying earlier, if we can save you six hours a week, you now have a choice of, well, do we need the four people that we’re doing this, that we’re spending six hours a week with? Can we basically get rid of one of them, or in their case, it was, hey, we can have them do away with more valuable tasks for those errors than then looking at these invoices and putting them into our accounting software. And I think that story is something that we’re seeing time and time again, right now from our customers is, wow, this is saving us a bunch of money. And every company is trying to figure out how to cut costs? Now there’s others like, I think I was seeing that the financial industry like FinTech companies are probably the ones that are most impacted right now. They sound the fear, which makes sense, I think like the funding for FinTech companies is that an all time low? So that’s an example of probably a space that could be a little bit terrifying to be in, but they’re probably still be winners in that space as well. Right? Time will tell.
Matt Watson 32:03
I think the one of the challenges we’re going to have now is you have people and CEOs that have been growing at a rapid pace over the last couple years through this, you know, quantitative easing and a lot of money in the market, low interest rates, and all of a sudden, it’s gonna be like a punch in the face them all of a sudden, if you know, the recession really kicks in. And it’s hard for people who have that momentum. And then like, Hey, we’ve been growing, we were winning, we’re doing all these things the right way. And they keep spinning, spinning, spinning, and all of a sudden, they’re like, oh, shit, we just smashed into a brick wall. And like, I don’t, I don’t know what to do. Like, I don’t know how to manage the company the right way. Right? Like, I’m the guy that’s out front leading the charge. And all of a sudden, I need to be hiding in a bunker somewhere. And potentially, they’re different personalities, and the person can’t even do it. And I can’t accept defeat.
Ryan Sevey 32:56
Yeah, I mean, that kind of goes to can a CEO transition from peacetime to wartime. Yeah. And I know CEOs, I’m friends with them. And I’m not going to name any names. But I’ve seen they’re not transitioning to wartime, they’re still spending crazy amounts of money on projects and programs that don’t impact the bottom line. And that’s, you know, I think they’re betting that this is all just overblown hype. But that’s a pretty big bet, right? And I think from our perspective, though, we look at this and say, Okay, if we’re wrong, and there is no recession, then whatever we can ramp up, it’s not that big of a deal. But if you flip that, and he doesn’t take cost cutting measures, and he doesn’t take a hard look, you’re kind of setting yourself up for failure, right? Like, it’s easier to conserve cash, the sooner you do it. Alright, if you have six months of runway today, you make some hard cuts, and now you have 12 months of runway, versus waiting till you get down to a month of runway. And then you make the cuts or what you may be by yourself for a month. And like there’s some, I think, basic math here that people should really look at. But you’ve got to make the hard call, you have to transition to being a wartime leader. There’s a great book called the hard things about hard things that has a whole chapter on wartime to peacetime. That would be a great reference for people to check out. But, you know, in my opinion, you have to transition to being a wartime CEO.
Matt Watson 34:30
I heard a great quote or saying the other day, and it actually was related more to trading like trading stocks or crypto, but I think it relates directly to this. It’s like whenever you go to make that trade, you have to consider how much money you could lose, not just how much money you could make, you have to understand you have to understand how much money are you risking. How much could you lose by doing this? And I think that I think that is perfect for this now, right? You talked about the burn rate and stuff and it’s like if I Don’t make these changes. Now, if we continue to do XY and Z, what risk Am I putting us at? Versus like, I can just totally I can eliminate the risk, I don’t have to take this risk, I can eliminate the risk if I go a different direction. But the problem is, so many people are going to continue the same habits they have. And it can be little things like, Hey, I don’t buy lunch for all our employees anymore, the right kind of budget for this right kind of budget for that can be little things at first. And that’s what we had to go through during COVID to the downturn of COVID. Like it was brutal. It was hard, right? Like you had to make those cuts. And at that time, it was much more severe. Because like, you really don’t know what’s going to happen with the world and this pandemic, like literally, is everyone on the planet going to die? Like what is going on? Right? Like, you don’t know, now this recession, like, you know, I don’t think it’s quite as doom and gloom as that. But it’s the same mentality, right is trying to figure out how much risk we take? versus, you know, do we eliminate some risks?
Ryan Sevey 35:59
Or I think also, if you look, historically, some of the best companies, some of the most iconic companies were formed during economic downturns. Yeah. And I think that says a lot about whether you can operate in more time. Like, one of the benefits to a recession is the labor market becomes less competitive. Ideally, although employment is currently very, very high. But yeah, like, there are companies, especially tech companies, that are constantly laying people off. That is an opportunity.
Matt Watson 36:32
Yes, for every other year ago, you couldn’t hire anyone.
Ryan Sevey 36:34
So if you think about what we did, we did cost-cutting early. And now when companies do their riffs in, you know, 400 500 800, however, many people are laid off. Now, most companies post publicly, like, hey, here, I was impacted. We can go and recruit the best of the best. There you go. We bring them in like that’s, that’s like the mindset people need to get into. If you’re running a company as well, wait, where’s the opportunity? With an economy down because there are tons of opportunities. It’s not all doom and gloom. And I think you could just look at the situation that we’re in a little bit differently. And think, Wait, how can I capitalize on this?
Matt Watson 37:14
It takes out a lot of your competition, like a good example is this friend of mine here in Kansas City who owns a company that wholesales concert and sports tickets. You can imagine how terrible that industry was during COVID when there were no events. It is wiped out, like the vast majority of all their competition is gone. Because they couldn’t hang through the downturn, right? And that same thing happens in a lot of industries. And even for Full Scale, like in our business, because of the way some of our competitors were handling things or not allowing people to work remotely or different stuff, we were able to swoop in, right? Like there are a lot of companies right now that will hire remote people. So the recruiting away from all these companies that won’t allow remote, it’s, you know, being able to take advantage of those times and those opportunities when they do exist.
Ryan Sevey 38:01
Yeah, that’s another great point. We started the company right in the middle of the pandemic, then remote first, since the beginning. And that’s hard. You know, I would much rather prefer to have everybody in the office, especially in the early formative days. But now that’s such a huge advantage for us is being able to hire anywhere. Yeah, that we want. And that just you know, that’s another prime example, you just got to look at the opportunity.
Matt Watson 38:30
Yeah, it’s always, I mean, Full Scale changed our business model because, you know, we have two stories of a skyscraper now that are empty. So eventually, we’re like, you know, what, we don’t need to pay for that anymore. If everybody can work remotely. You know, it’s interesting how these things change, you know, so for your business, do you see you staying remote long term? Or do you foresee having some little regional offices and co-working space? Or, like, how do you foresee that playing out over time?
Ryan Sevey 38:58
I think long term, we do want to start opening up regional offices. We do want a place where we can bring the whole company together. And there is still a lot of value in having people come into an environment where you’re all interacting face to face.
Matt Watson 39:15
It’s just hard when you’ve hired people in every state.
Ryan Sevey 39:18
It is so so I think, Okay, I think this would be a great business opportunity for somebody like a great awesome startup would be a great, amazing space, kind of like we were but a little bit different, where companies can kind of have the whole floor for a week or two at a time. Because what I like to do is fly in my whole company, maybe once a quarter, or even bi-annually, but have our kind of there do big kickoff suit like quarterly planning and then go home, and if you look at that, yeah, the flights and all that it’s actually still cheaper than having, you know, a dedicated around dedicated office.
Matt Watson 39:58
Okay, so we’re gonna do this, and we’re gonna call him. I don’t know what to call it, but we’re gonna do it as a cruise. How about that? Cruise?
Ryan Sevey 40:09
But yeah, I think long term, that’s something that we do want to bring people together. Last year, we brought everybody down to Florida. But it wasn’t our space. And that was actually pretty expensive. So yeah.
Matt Watson 40:23
Well, as we wrap up the show today, I do want to remind everybody that if you need to hire software engineers, testers, or leaders, Full Scale can help. We have 300 people and the platform to help you build and manage a team of experts. When you visit FullScale.io. All you need to do is answer a few questions and let our amazing machine learning match you with the right vetted, highly experienced team of software engineers, QA, and other engineering talents at Full Scale. We specialize in building long-term teams that work only for you. You can learn more at FullScale.io. So as we end the show today, what other suggestions or, you know, nuggets? Can you share what other entrepreneurs can learn from?
Ryan Sevey 41:07
That’s a good question.
Matt Watson 41:08
What is any final wisdom for those that could just be about machine learning and how we can use machine learning or about entrepreneurship?
Ryan Sevey 41:17
Yeah, I think just never give up on your dreams. And I know a lot of people say that tenacity, at least from my experience, always won out in the long term. Even if you look at what Mantium is doing, we’re just now being able to realize a mission that I’ve been on for years and years and years. It started kind of percolating back in the Next Osias days. The tech just wasn’t there for what I really wanted to build. And we’re finally at a place where transformers have come online, and it’s just never giving up. Right? Like, eventually, stuff will happen if you just keep grinding. At least, that’s been my experience.
Matt Watson 42:10
All right. Well, thank you so much for having you on the show. So once again, this was Ryan Sevey and his company Mantium. And he is one of the top startups in Cincinnati. So congratulations on that. Be sure to check out the show notes and can learn more about other top startups in Cincinnati. So Ryan, thank you so much for being on the show today. And again, I should mention that Mantiumai.com; check on their website. So, yeah, thanks. Thank you so much.
Ryan Sevey 42:45
It was awesome. Thank you.
Matt Watson 42:47
All right. Thank you. Thanks, everybody.