[Podcast] Industrial Evolution — A New Era of Manufacturing ERP: Real-Time and User-Friendly with Fulcrum
Chad Perry: I'm your host, Chad Perry, CTO and software engineer. You can find more episodes and contact information on our website, industrial.fm, enjoy the show. And if you need help with custom software, data, integrations, or strategy, be sure to reach out. That's industrial.fm In this episode, we're speaking with Sunny Han, founder and CEO of Fulcrum, a Minneapolis-based cloud provider of ERP software billed as the manufacturing operating system of the future.
We've covered ERP in other episodes, but this is the first time we've seen an ERP tool that represents a truly new generation of software built from scratch within the last decade versus being a continuous upgrade from the software that originated in the 1980s and 1990s.
That's an important distinction because, as Sunny will explain, the new age of digital twins is all about philosophical, behavioral and other changes, not directly related to technology, which can be difficult to get from legacy products that are constrained by decisions made a long time ago.
In fact, Fulcrum was only founded in 2015, a full decade after the first generation of SaaS and cloud really took off. So they've been able to take what works for the cloud, envision the future of purely cloud-based organizations, and start from scratch, building a product with that future in mind. Sunny, great to have you with us.
Sunny Han: Thanks for having me, glad to be here.
Chad: Now, before we get into the ERP stuff, can you give me a quick rundown of your background and how you made your way into manufacturing?
Sunny: Honestly, it was kind of by chance. As a kid, I was a math team guy. I learned to code when I was young and my mom's computer lab when she was in grad school, wrote some software and played a lot of video games.
But, deciding not to go to medical school, I got into consulting just to see what was out there and just by chance, the majority of my customers and clients were manufacturers. So I worked with a few hundred manufacturers over the course of almost eight years. And that's really where I've learned a lot about how things are produced and how manufacturing facilities are run.
After that, I started to implement ERP systems, these legacy systems that you just spoke about, write custom software on top of them, and integrate them with a bunch of different tools. And I started Fulcrum because after a while I realized no one else was going to do anything that I thought was obvious to do, and that's really where we started off in the beginning. It was just writing custom ERP solutions, fully custom ones just to learn what we didn't know and, and how to do it really well.
And then after that, we modularized it and tried to integrate it with legacy systems, but really just got very frustrated at how hard it was to integrate compared to how easy it was to train our customers to use the actual software that we wrote. So then we took all of our custom ERP learnings and all these new cool things that we built and put them together and made Fulcrum and have been polishing it and expanding it ever since.
Chad: Yeah that really touches on the big picture idea here, which is — really the question at hand that I wanted to ask of you — and that is why build another ERP, especially starting from scratch. That is a massive undertaking. And I'm actually curious, you mentioned that you were building custom ERP for other clients. I'm curious if these were actually full-blown ERP or if you're talking about just pieces of tools here and there. So really why start from scratch and how do you even attempt an undertaking like that?
Sunny: Yeah, I think there's just a lot of really cool things that are long overdue for manufacturing. Stuff that isn't even that cool anymore, it's maybe five, 10 years behind where other industries are. And I think one of the biggest reasons why everything's being held back in manufacturing are these underlying ERP systems: Their data structures, the way that they're implemented, it's all very old and archaic.
I think there's a lot of very good value that they delivered in the ‘70s and ‘80s and ‘90s and 2000s when there wasn't really anything else. And that was the most modern technology. Um, but they're fundamentally architected to sit somewhere inside your business to not be updated for a year, two years or three years to be implemented over the course of two, three, four years and cost hundreds of thousands of dollars and be customized for you. And that just makes it not very scalable. It's really difficult to make software really good when fundamentally it has to be kind of semi-custom.
So the model that many, many other really successful business-to-business like SaaS platforms have taken is let's plow a hundred million dollars into making something that no individual customer could ever have the budget to build, make it that good, but offer it to you for a couple of thousand dollars a month or a few hundred dollars a month, if you're a small company.
And that's the model that has kind of swept a lot of other different industries and really just hasn't taken place here in manufacturing. you mentioned other systems that have just been essentially re-skinned or updated from legacy architecture. And that really is what's out there in the marketplace right now: old ideas with a new face.
But you know, you can understand that they're owned by private equity companies they're operating in a way where harvesting cash is their primary goal. That's just the kind of behavior you're going to get.
Chad: There are a couple of points there I want to come back to. The purpose of the company's existence, whether it's to, to harvest cash that you use, as you say, or fundamentally provide value to the end user. But you also said something that I thought was really interesting, and that is that these companies would go and invest a hundred million dollars and build an ERP that is the end-all be-all, and that no single company would ever be able to produce on their own. But in my experience, that's actually been part of the problem because it's really not a one-size-fits-all-solution.
So you've got two sides to that problem. The first is that these giant systems are not really appropriate for smaller manufacturing or maybe really highly specialized manufacturing operations. And the second is that even the larger ones that could use a lot of that functionality, it is a massive undertaking to incorporate all of that into the business at once and it essentially becomes a linchpin.
And you probably see this out there as well, where these companies that are essentially stuck on software and legacy software, this doesn't just happen in manufacturing, it's across the board. So I'm curious how you feel like Fulcrum approaches that differently.
Sunny: If you boil it down to its essence, it really has to do with trying to change the way that the product is developed, not just from a software development standpoint, but from a whole business standpoint.
Historically speaking, most ERP systems are sold from the top down. A group of smart individuals that are customizing software and putting together a presentation for owners and executives, relying a lot on how much power that can be delivered by certain reports or dashboards or things like that. And then all the things that feed that data that support the dashboard, those things are kind of second place or third place or fourth place or last place. And we, we start from the top and we start building down.
And I think the realization that we've had across the entire software industry is that the data that's generated — the end user, the person that's actually doing the work, the operator, the machinist, whoever it is on the shop floor — those people are actually far more important.
It's weird for a lot of people who have bought software that was primarily designed for executives to switch into this mindset where, this product, 90% of its development is in looking at how the end user uses the product. I think that's really where the big pivot is, is that we don't want to go to market by making huge presentations and selling big contracts. We want to go to market because the actual users of the software say “this is the best thing that I've ever used, and this is the most useful thing that I've ever used.”
So, you can put a hundred million dollars into creating a great sales engine that sells something that is very attractive to an owner. Or you can put a hundred million dollars in thinking about how manufacturing actually works, the first principles of adding energy to material to make things better and more useful for the person that's actually using that manufactured part. And then you can build your entire structures up from bottoms up.
So, I would say for us, that's really the big thing is that our distribution method, the way that we price things, the way that we sell things, the way that we market, the way that we implement, all of that is informed not by how do we make the most money, but by how do we deliver the most value and what is the natural way to support letting the end users kind of lead the way?
So that I think is something that is really valuable, it's harder to do, right? It's harder to do because there's far more voices when you're not just talking to owners, there's a lot of different kinds of companies. Like you said, it isn't a one size fits all problem, but we have to compress everything into a really nice, simple, intuitive, easy-to-learn product that does fit everybody. That just means it's harder for us, right? We have to play a lot of mental Jenga to try to slide little pieces out and put other pieces in to try to assemble something that's really good.
Chad: Yeah, I'd like to go a little bit deeper on this, but I want to rewind the clock and look at why it is this way. Because it's not like there weren't innovative companies in the ‘70s, ‘80s, and ‘90s, and it's not like they didn't have the opportunity to go and talk to operators on the shop floor.
So, I remember in our intro — one of the things in our intro call — one of the things that you said that stuck with me was that over half of the software developers in the world in the 1980s were machinists who were programming CNC machines. And I just thought that was such an interesting factoid. So, why do you think we are where we are right now as such that you can take this approach and it's this novel approach or this novel pivot?
Sunny: I think we've seen it in just about every industry, right? Um, in the ‘60s and ‘70s being a pilot was really, really, really glamorous. It was a really hard thing to do. But by and large, nowadays, most of the work of a pilot is compliance-based and babysitting an autopilot algorithm. Now, obviously we've seen some issues with that, but that has formed a very similar pivot.
And I think that what happened in manufacturing is that business software was always, historically, innovated before consumer software. So you got a lot of people that didn't have desktop computers — that had no idea what a computer was — that went to work and had to work on a computer. And so there was a lot of training that needed to happen for the everyday person to know how to operate these systems. So they had to make it implementable in a way that was customizable. That was driven by what would make it work for this particular business, because it was really difficult to have any sort of intuitive understanding of computing inside that person.
Things have changed and you can look on Crunchbase, or on other data sets, and see the amount of money that has been poured into consumer software versus business software. And you can kind of see this leapfrogging where social media, games, other utilities — even like Mint that helps you manage your budget at home — those things have gotten a lot more attention, a lot more developer time. And we've had this effect where people are actually more intuitively engaging with their personal electronics and personal software than they are with their business software. So we can tap into that now.
We can tap into the understanding that people know how to interface with a touch device. People know how to interface with something that is real time. They expect that. And so it's not as foreign to them anymore. There isn't this resistance or pushback when you say, “Hey, you put this information here and somebody else can say, see it elsewhere.” In fact, I think more and more people are expecting that and not getting it from their business software.
So I think it's fundamentally the curse of early adoption in an entire industry where if you spent $4 million implementing an ERP in the 1980s and 1990s, you're going to try to avoid that experience for as long as possible and not maybe realize that the way that modern software is implemented is just not the same.
Chad: Do you think that is part of this pivot that has happened, where we've gone from management-centric software to everybody-centric software, or are those kind of two different threads?
Sunny: I think it's the same. I think that people understand now that work in general is changing. And I'm not just talking about having zoom calls and sort of meeting in person, but there's just a labor pool for manufacturing that's getting sparser and sparser. It's getting harder and harder to excite younger people to work at manufacturing facilities when they're primarily using pencils and erasers on paper travelers and not really getting a sense that they're participating in the future instead of, you know, hoisting up and shouldering the burden of supporting a business that's in the past. So I think things are changing in a lot of different ways, but both of these things are along the same thread, in my opinion.
Chad: You mentioned the paper travelers and what comes to mind is these intermediary steps where you've got a legacy or an established business with legacy tools, legacy digital tools, or maybe they're implementing new digital tools, but they're doing it in such a way where you've still got this legacy mindset. So paper becomes an intermediary, and so it's not a fully optimized process.
And I remember in our intro call, you were talking about this idea of wanting to actually launch into the future, not just say “okay, what can we do and then upgrade,” but actually say, “let's create the future.” So tell me about that future.
Sunny: If you think about it from a usefulness standpoint, like what would we — regardless of what economic structure we have or societal structure we have or political structure we have — what would be useful for how we live in general? And one thing that'd be really useful is increasing the ability for us to build things better and faster and help us innovate more. There are things that are holding us back. Like you have to order a million units of something to get any sort of economy of scale, and you have to figure out which certain manufacturers you have to work with to get something built.
So for me, one of the things that I believe will be generationally valuable is a way to network all these businesses together. And it's obvious to me, right? I grew up in a world where, when I learned how to write simple scripts and code, when I was a little kid, the internet didn't really exist. It did, but it was, you know, uh, dialing up to servers that were hosted on some university.
And we networked all these devices — people don't really realize that in your pocket at home, there's a dozen, two dozen different devices that are all connected to each other. And you can find your phone by pushing a button, and you can find what restaurant to go to, and you can find these things — that doesn't exist in manufacturing, which is one of the core industries and actions of being human in a civilization.
We would always want to have more and better useful things. So imagining what that future looks like, where the ideas that we have, that when we're designing a product naturally fluidly flow out into an entire network of manufacturing producers that know how to operate a CNC lathe, that know how to do rubber extrusion, that know how to new plastic injection molding, that have this expertise, and being able to harness that in a way that's automatic, that is definitely useful and will be useful for hundreds of years to come.
So how do we create that? How do we allow this data to marinate through this entire network? One of the biggest bottlenecks is that no computer system actually sees very much into the shop floor. We wait for people to write things on paper, then enter it into a system and then write a report on it. And we maybe kind of understand what's going on on the shop floor.
I think over time, we're going to get more and more data closer and closer to real time. And that's going to help us understand what's going on and make real-time decisions. And that compression of the timeline between event and decision is going to make a lot of people, a lot more profit in our current economic scenario, and also allow for us to actually get this connection to happen. Because if we're working with data that's two, three, four, five weeks old, that's going to be an impossibility.
So one thing we really, really focus on is: How do we give people tools that are as close to real-time as possible, and how does it benefit them and how do we align the long-term interests of this network with each individual manufacturer?
Chad: I keep thinking this is almost a new take on our redefinition of this idea of just-in-time manufacturing, because you mentioned being able to, or right now where you have to place an order for say a million units, just to be able for it to be economical. And we're already seeing a shift in this. I mean, I hear a lot of people talking about this, I read about this. We know that where we want to be is to be able to order either essentially mass custom anything or low-volume unique units. And so there are a bunch of different things here that are kind of floating around in my head. And you mentioned really the big innovation being this decision to pursue real-time knowledge to compress that timeline.
Before we get into that though, I want to come back to something. And that is, if you imagine that you've got the ERP and it's in place and it's working smoothly, that's great. That's a good future to think about, but there is the implementation and the implementation has always been a challenge. I know that you guys are really striving to allow customers to do this on their own. So how much progress have you actually been able to make given that is such a substantial challenge?
Sunny: Yeah, when we first started, it took two years, two-and-a-half years to implement the first projects. Over time that's decreased down to six months and, with some of our newer customers, as fast as four to six weeks. So that's a huge difference already and primarily differences just in making the application intuitive enough that you're not having to train and retrain and retrain and retrain because people forget about things. I feel anxious about things.
One of the things we observed when we were deploying our first-generation software and looking at other legacy software is that there was this feeling of anxiety. Like, “I don't want to touch something; I don't want to submit this data because what if it's wrong?” And it takes that away, giving them some safety, letting them undo stuff, giving them instantaneous feedback on how it's changing other data, all of those things just relax people and allow them to learn far more and in a much shorter amount of time.
We've also done some automation internally on how we scrubbed data, how we integrate different systems, how we upload things, we'll continue to work on that, but we really believe that a big, $50 million manufacturer should be able to implement Fulcrum in less than a month within the next year or two.
So that's really our goal. And I think that's the boundary where it transformatively changes the way that you even think about implementing software in this space. So we're really trying to push for that goal.
Chad: So that's kind of your first category of innovation is you go from implementation being something that is a deterrent to now where you can do it yourself in less time. It's less complicated. The factors that allow you to do that also have downstream consequences in terms of just everyday usage and adoption. And then you have, of course, the future that you talked about, where you just get the benefits of this near real-time system.
And then I'm thinking of kind of a third bucket, which is that now that you don't have this system that you have to feed and maintain, this is the advantage of the cloud and SaaS, right? Is that now instead of having a legacy device that's — sorry, not a device — a legacy system that is updated say every three years, or maybe a lot more, a lot longer than that, because you're afraid to touch it. And now where you have constant updates, maybe hourly, maybe every few minutes even.
Sunny: Technology is changing faster and faster, right? You can imagine in the 1990s, if you updated your software every five years, you wouldn't miss very much. But imagine if your phone didn't update every five years right now, think about how much you would miss. So that is the future. We have to be closer to real time ourselves in order to deliver good security for our customers and all the most recent technologies from an optimization standpoint and performance standpoint. So things are changing fast enough, where I think that if our deployment cycles are any slower than once a week or once a day even, in the future, it will be too slow.
But yeah, that's, that's exactly the big change and big pivot that's happened in the world is that we've released a lot of these bottlenecks, which has allowed us to develop faster, implement faster, and push faster, which means that as a company, our philosophy can't be to build something, own the market, then harvest cash.
I think there will be other competitors that will participate in this market. Other things that will come about, we have to constantly deliver our customers this feeling that you don't ever have to worry if Fulcrum is the best thing for your company; we’ll always ensure that it is that. That's not a sprint, right? That's a marathon.
Chad: One thing I have not heard you mention is anything about “well, we came up with the latest, greatest scheduling algorithm” or whatever it may be. I just used that as an example. I remember you saying before that a lot of these algorithms have been around since the 1960s. I actually think about algorithms — even though I'm a software engineer and I'm a computer guy — I think about algorithms in terms of just steps and structure. That's actually something that predates the digital era. Tell me about where you guys kind of fit into the landscape of algorithms. Are you just making something more accessible that already works really well or is there anything new there?
Sunny: So with any data set that you're running a software program against, the features or the shape of that data set really determines how you write that algorithm. If you have a lot of small data points that are very related to each other, you might write an algorithm in one way, or if you have very few data points that are unrelated to each other, you might write it in a different way. I think what is happening is related to what we just talked about a few moments ago: The shape of the data that underlies manufacturing is changing at a really, really fast pace.
If you write an algorithm that really fits this particular customer, this particular niche, or this particular industry, the likelihood that the underlying data isn't going to change in some way is very low. You might get a bunch of new orders all at once because we're coming out of COVID and maybe in a year, the PC board manufacturing market will actually get chips from Taiwan and from China to be able to make cars or wherever else, we're going to see this huge uptick in demand for all these components that go into everything else in the car.
That really changes the shape of your data. What was otherwise a very consistent order stream is now a drop and then a huge spike. Whatever scheduling algorithm that you had that was written for a very consistent “here's my blanket order, here's how I fulfill it” is really not going to work when you lose a lot of revenue and gain it back again.
And so writing something that is for every single instance is going to become more and more and more impossible. I think what you're seeing is that these algorithms were written by smart people, like you said, but they were written in the ‘80s or ‘90s, and any algorithm that you have still has to be adapted to that particular data set in the data shape. Really what we're trying to do is what everybody else is trying to do in the tech world, which is creating these learning algorithms, systems that are looking at the shape of the data and retuning themselves to be able to be performant, no matter what the environment is.
We still have a lot of work to do there. We have a budding data division that's just starting out, but I don't really see anybody else doing it yet. So it's not as if we're doing it for competitive reasons. I just think — I have a really deep belief that this is the future of how data should be analyzed and executed on, for manufacturing and for all sorts of other industries, too.
Chad: That's really the revolutionary aspect of the fourth industrial revolution, from my perspective, one of the core pillars of Industry 4.0 is this idea of the digital twin. Well, the digital twin is not really that useful if it's outdated by — in this case, it could be by minutes — but, imagine something that's outdated by hours, weeks, or possibly even months. So am I getting that right? That really the core of this opportunity here that Fulcrum is trying to capitalize on is that we're now entering a world of real-time decision-making, the ability of decisions to be made more currently in the business.
Sunny: Imagine that your system is really rigid, whatever system that you're using. You're faced with a choice between continuing to operate your business the way that your ERP is structured or making changes to your business process, because the world is changing and continuously creating this tension and seam between the real world and the digital world that you have.
That seam is almost always filled with Excel spreadsheets and paper and whiteboards and things like that, right? So you can — every single manufacturer can look around in their office — and they can quantify pretty quickly like how different their digital system is from reality, just by how many clipboards and whiteboards and pieces of paper and spreadsheets there are.
I think that the goal then is to have a really adaptable business so that if you are a CNC shop, but you want to start doing powder coating, you don't have to buy two different systems to make it work.
You can do everything still on the same system and you can, you can make that happen. The same thing that's going to create a network for Fulcrum is the same thing that allows every one of our customers to do whatever different process they need to all within one system. So it's not just even about creating this digital twin, it's like a digital shadow you can move and it will move with you. It doesn't require any effort. It just is automatically a cast shadow against what's actually happening in reality. So that's kind of how I think about it mentally.
Chad: I really like that concept of digital shadow. You mentioned this idea that you can look around, if you're a manufacturing business, you can look around and actually just take note of your environment and see how many Excel spreadsheets and clipboards. I think that's such a powerful way to self-assess. You don't really need to do anything else, just really sit down and — not even sit down — just take a look around you.
Now, I would say you've mentioned competitors multiple times. I'm sitting here thinking, well, obviously the digital twin and tools and ERPs and stuff like that, that supports this concept. There's not one only one way to do this, right? So when you describe a shop floor, I'm thinking that there are two things going on: There are machines and there are human operators, and they work together. And some of them, you know, a machine may be able to work by itself, maybe with minimal or no intervention or whatever it is.
And we want information about what is going on, on that shop floor from both the humans and the machines. Well, one aspect of industry 4.0 is we have all these cheap sensors now that can feed data back into the system. But the data coming out of human interaction seems, on the surface, like it would be much more difficult to capture, especially when you think about things like error conditions, whatever that might be — maybe a human forgot. So how does something like Fulcrum, or just this vision for the future, how do you accommodate that?
Sunny: We internally talk about Fulcrum as actually three things. One is this platform that handles the same things that everybody needs — invoices and sales orders and purchase orders and inventory and things like that — and handles that minutiae in a way that is compliant with accounting rules. That is the way the business is done. These are kind of cultural things, right? Business has done a little bit differently in every single culture and their kind of cultural rules and what the decorum is on how to do business.
But I think the biggest innovation in the Fulcrum platform is a way for other companies like MachineMetrics or like Paperless Parts or other people that are working kind of on a specific problem to integrate with our ERP, without having to pay an implementer or have someone on staff that their job is just to learn your unique custom implementation of Epicor and link it together. Something that is a modern API, where we can automatically link this data.
I think some scalability problems that exist already with these other companies, that are doing really good work in understanding what's going on in the real world, we can help solve by just making a really fluid interaction between them. Whether it's CMMS data for maintenance or upkeep, or MachineMetrics, or Paperless Parts for quoting, being able to be that platform and allow them to scale faster will help reinforce a flywheel of innovation between different companies.
Secondarily, internally we talk about the need to create one or two or three killer apps, just like VisiCalc was for the Macintosh way back in the day. For us it's a digital paper, that job tracker, that in-process tracker that replaces the drawing, that replaces the work instructions, that's all digital, that's live, that gives the user the ability to stop and start a timer to interface with a bunch of different things. There's a few other things that we have that are like that too, essentially demos, if you will, like “look at this cool thing you can now do with a future platform” just to open people's eyes and have that be the hook that brings them into this platform.
Then lastly, what we talked about earlier, the grease as we call it, to get things to go faster, to self-implement and things like that. So I think that there will always be a lot of unimaginable cool things that I can't even imagine that will be able to be built if we build this platform right and disseminate it across the market correctly.
And you're already seeing it, like the way that MachineMetrics does their data collection — and some of these other newer startups that are doing it with really cheap Raspberry Pi sensor arrays and things like that. I think there's going to be a lot of different ways to do it from vision systems, to seismic systems, to electric systems, to all different ways to collect information, as long as we can aggregate it in one place and serve it to you to make decisions with that's really what our goal is. So our goal is not to be the owners of all of these different types of technologies, it's to create that platform from which other people progressed on top of.
Chad: That’s actually one of the reasons that I really wanted to talk to you guys: Your focus was, from the beginning, on UX/UI usability. This is a gripe that I have just personally with a lot of business software and more so now with, actually surprisingly, consumer software.
I feel like we've gone through these cycles where in the early days, let's say the early 2000s where we were finally starting to have a lot of interactive capabilities online, UIs were still well behind the desktop. Then we went through this transformation, and I would say maybe it feels like we kind of peaked at maybe 2015. And then I feel like I've personally been seeing a lot more bugs and just there, there's more of this focus on “how can we optimize the data points?” such as usage versus the actual usability. So I feel like companies are kind of getting sidetracked by these other things and actually want to talk about that in relation to Fulcrum as well.
But coming back to this original topic. So you guys are really focused on UI/UX and usability. Do you feel like that's paying off or is that just kind of a nice to have?
Sunny: I think it really helps with greasing the wheels like we talked about earlier. Having a really good user experience means that when we watch our customers using the software, especially in implementation, they get it. We have an hour long call and, five minutes in, they're like, “we don't need the rest of this time. We got it.” I think people come to these meetings expecting to take all one hour going through one specific module, and yet in 10 minutes we're able to transfer the knowledge because the interface is super intuitive.
And as we rolled out more features, we went from two-and-a-half engineers working on the product to over 13, our pace of product development skyrocketed. We, internally, have some bloat that we've been taking care of and spending some time to refactor and redesign to make it better. So we've been redesigning the entire product over the last three months, we're halfway through, and we have another half to go and then we'll have the excuse for ourselves to add more features and more bigger features instead of enhancing the current one.
I think it's discipline on our end to resist the urge to build more features and get more customers, and instead to continue to make it better. But I think the primary reason why the UI and the UX is so important is that you need to establish a non-adversarial relationship between you and the software that you use.
Otherwise, the primary motivation for putting data in isn't enjoyment or value, it's anxiety. And that's really what all these spreadsheets and pieces of paper are: It's anxiety driven. I'm trying to make sure that I'm not missing something and miss a due date and get a customer to be upset.
We want to take away that anxiety because making cool stuff is more and more difficult. Every single year, as new technologies come out and new specifications come out, we need those people that have the knowledge about materials and how to work with them to be focused on that instead of on the operating system that they're running, right?
So for me, every ounce of anxiety and annoyance that we can take out of their usage of Fulcrum is an ounce of willpower and effort they can put into making good things for the world. And I don't know how to do all that, right? I know a lot about manufacturing, but I couldn't operate a wire EDM shop — I could learn it over some time, but there are other people that have this rich knowledge. Why are they spending 50% of our time figuring out how to pull a report?
We should be spending all that time trying to improve our processes and trying to do other things that will actually help civilization, right? I think that's really where the value comes in: We want to shift that effort away from using the system.
And there are times — we use a system called Pendo to track a lot of data in our products — there are times where we're like, “whoa, people are actually spending 80% less time on the invoice grid!” but then we realized they're just able to get their job done a lot faster. So yeah, if we were trying to optimize for utilization and usage of the system, then we would aim for the wrong thing.
Chad: That's kind of what I was alluding to earlier, but what you're talking about, making the experience enjoyable or something that is desired behavior, this is the behavioral change that I was referencing in the introduction. By contrast, I know that eventually we will be in a world where we've got lights-off factories, we've got AI, we’ve got machines, but we're probably still pretty far off from that, at least in terms of the stuff that needs to be done to do that.
It's extremely difficult to replicate an intelligent process with AI without huge amounts of training data. So it sounds like Fulcrum bypasses that by simply having a really smooth fluid interface between the human operator and the machine.
Sunny: Yeah. Traditionally, you know, this experience that you're talking about — Intel, famous American manufacturer, they famously had this tick-tock method where on the tick, they enhanced the way that they were making these processors, and on the tock, they changed the architecture of how these different processors and the different components were interacting with each other.
I think the same thing is happening in manufacturing. We've had this big tick where we have robotic arms, we have artificial intelligence that can work on data, we have conveyor belts that can be really smart, we have warehouses that can be automated, we have vision systems, but we haven't had that tock where the infrastructure is rearchitected in a way that supports all these new innovations.
So I think it's become apparent to me that this transition from where we are now in 20, 40, 60 years later into where we can be, it can't happen quantum leap, we can't start from here and then instantly get to there. And really that's kind of the purpose of Fulcrum is to provide the infrastructure, to make this transition possible in the best way possible for each individual manufacturer, not just for some unknown future state, mega factory that you can imagine.
That, I think, would be catastrophic for us as a civilization. So really trying to create the infrastructure and the environment to allow for that decentralized smaller production, localized to where we need the goods. That's really the vision of the future there.
Chad: Speaking of humans, another really key aspect of this transition is knowledge transfer. When we're talking about human operators being able to do things better, more efficiently; they are essentially large stores of information. So especially somebody who has been with the company maybe, or in that particular role for maybe 30 years, how do we get that knowledge transfer and where does software like fulcrum or really anything else fit into facilitating that?
Sunny: Here we're copying again from — standing on top of the shoulders of giants. I think if you distilled down the problem of knowledge transfer, really, it's a problem of indexing and search. So Google, obviously their mission is to categorize and know all the data in the world or something like that. I'm misquoting that for sure. But part of that effort has allowed you and me to go online and find information.
We're essentially using Google as a staging ground for all the knowledge in the world. And then we can go get it whenever we want to: We can find the right Wikipedia article; we can find some article about which TV to buy; we can filter for only results that are within the last year so we don't get old TVs, but if we want to look at old TVs can as well; all that is just out there.
So similarly that same thing applies at Fulcrum. If we give people a really easy way to put information to the system, to put notes against the jobs that they're working on, to put check points against their operations and store all that, and then we also index really well, use the right data structures, and allow people to search organically for that information.
That is essentially the long-term self-sustaining plan for knowledge transfer, so that you can search for stuff a hundred years back in the future or whatever it may be, and you can say, “oh yeah, the last time we used the salt tank at this temperature to cure the rubber, these were the environmental conditions.”
It's the lab notebook that you're supposed to keep in college chemistry that has the temperature outside so you can detect patterns. And then with all this data, all this knowledge, then we can sic the AI on it and say “look at all this data and tell me patterns that I can't see.” So it starts with collecting data, and then it's searching data, and then it's finding patterns within the data after that.
Chad: And the big advantage there is scalability. So if you have all of that data preserved, that data is not only immutable, you can also copy it as many times as you need to mix it with other data and interpret patterns as you said. By the flip side of that, I think about something like a human-to-human apprenticeship. And I'm a big fan of this idea of apprenticeships. I think that it, it seems clear to me that they are making a comeback and that they have a very powerful role just in general in our society. But that kind of interaction is extremely not scalable. It's the opposite of scalable. And it has the downside of emotions, or if you have somebody who's really busy, you don't want them necessarily distracted by this apprenticeship, but it's necessary.
This is something that won't happen overnight though, so how do you see all these things playing off of one another? Or what do you, what do you see as the ultimate plan 10 to 20 years down the road? What this is going to look like in terms of a mix of all these things?
Sunny: So internally at Fulcrum, when we onboard new employees, they have access to this library of videos that we've recorded of demos of the product and things like that. And we struggle with things — like the product changes so fast that those videos get stale really quickly and documentation gets stale really quickly — but I think this concept of having a library of knowledge, that you can go to, to find information, will take this apprentice and fill their brain up in some way, but it’ll be broken. You're going to be missing some context. Things will be different. Your interpretation of it is going to be different. You have a different background than the person that recorded the video and the way that we can maybe not make it infinitely scalable, but increase the scalability, is that every single apprentice should be able to autonomously and asynchronously get to some level of knowledge.
And that one person that has an incredible amount of knowledge, or the 10 people at a company or whatever it may be, can just work on the finishing touches of filling in the gaps and bringing that knowledge really into cohesion. So if we can take somebody from 80% to 100% instead of from 0% to 100%, we've instantly made it five times more scalable, right? So there are things that can help ease this transition and that knowledge base, if you will, that data store can be very human in the way that we're putting the data in it. It can be a message instead of a computer program or CNC program or something like that. It can be something that has notes that has something very human, has pictures, has things that are visual in the way that we learn things.
Even just our job tracker, being able to take a picture of the final good and store it with the job forever. Everybody else in the future can immediately see, “Oh, well, this thing doesn't look like the thing that is in the picture, that was how it's supposed to look.” Just passive things like that are how humans learn. And the human thing that can't be replicated by computers, at least not yet — and I don't know how long it'll be until we can — let's reserve that for the people that actually exist out there, so that we're efficiently using as much time as we can of theirs.
Chad: One last point. I want to bring this conversation back around to something that you mentioned in the beginning, companies that exist for the purpose of just harvesting cash. And I, and I'm not saying there's anything wrong with that, every business has their model and investors. But in terms of what works for creating downstream value, one of the things that it sounds like you guys are doing differently coming out of the last previous few decades is the pricing model. How is it different?
Sunny: One of the things that we reworked from the ground up is that data is really important, it's very obvious. But if we price everything by the user, then every single time you have a new employee, one way to save money is just to share user accounts and not get people on the platform, or have them work in a spreadsheet elsewhere. In order for us to be philosophically high-integrity, we have to make sure that every facet of our business model fits what we're trying to achieve.
So for us, every one of our customers pays a flat monthly fee by some metric of their gross margin or whatever sort of economic value that they're producing. And no matter how many employees they have, they'll have the same cost. And it's important because not only does it flex up when they're growing, we don't reassess that on a daily or monthly or annual basis for as long as they sign an initial service term contract-wise, they're locked in that price.
They can feel free to grow, and invest their capital in equipment, and hire more people, and expand their shops, and our costs will be the same. Now three years from now, when they're triple the size, obviously our price will go up — won't go up triple — but we know it's a pain in the ass. We know that it's harder to manage something like this than just by the user, but it philosophically represents an alignment with what we're trying to achieve overall with the product.
So, very soon, we're going to have a very public pricing page that you can just put in your shop and what you have in terms of workstations and things like that, and it'll spit you out a price. Our implementation fees will go quickly from, you know, a million dollars for a really big company or $50,000 for a small company down to $20,000 for a big company and $0 for a smaller company.
We really want to make sure that we're aligning the economics of how we charge our customers with what we're trying to do: We're trying to create this path where we are pushing your business forward passively. And that means that we need to take the pain of not getting paid upfront and instead load our economics to where our investments are going to be, which is in the long-term. And I think that manufacturers have been used to making long-term decisions for a long time. You don't buy a $500,000 five-axis CNC machine for three months of utilization. You're thinking about it in terms of five years or seven years or 10 years, so I don't think it's going to be a very foreign concept. And I think that once we give them this ‘aha’ moment where every month, every week, sometimes every day, new and better stuff — that isn't inconveniencing them, that isn't just a color change — that's fundamentally adding new concepts and new ideas into how they run their business. That I think will help us to kind of close the loop here on what we're trying to do.
Chad: I'd like to ask you one last question and that is: What would be the key takeaway that you feel like is most relevant, given everything we've discussed,
Sunny: As things are changing, as owners of manufacturing companies are getting older and next-generation talent is taking over, as supply chains are getting shorter, as environmental concerns are taking over distribution decisions and things like that, as goods become more personalized and order quantities become lower, it doesn't have to be scary. It doesn't have to be a horror show to manage this transition.
Really the fundamental difference is believing that the future will be different instead of expecting it to be the same as it always has been.
I think that there is some psychological comfort in assuming that the future will be the same as what you're really good at, where your intuition comes from, and where your experience comes from. But really the only way to grow is to feel discomfort, right? So being willing to feel some upfront discomfort in order to gain access to the future of how work is going to be done is a small price to pay. But it's one that everybody has to grapple with themselves.
So, I think that's the biggest human thing that I see that is a big takeaway is that things are changing. It just doesn't have to be as hard as you may assume.
Chad: Well, on that note, Sunny, if somebody wants to get in touch with you or Fulcrum, or they want to see what you guys are up to, what is the best way for them to do that?
Sunny: On our website, fulcrumpro.com. You can book a demo and talk to any one of our people. You can email me directly Sunny, like the weather S U N N Y@fullcrumpro.com. If you have any questions directly from me, feel free to reach out. And I'm not the most responsive, but I'll try to respond as timely as possible, but would always be happy to chat with anybody that's involved in manufacturing anytime.
Chad: All right, Sunny, thank you so much. Really appreciate it.
Sunny: Thanks, Chad.