In this episode, Eric Dudas of FCH Sourcing Network is joined by Fastenal veteran turned homesteader Sam Aldinger and ERP code-slinger Anders Green of ALOX4 to dive deep into the evolving world of fastener sourcing, exploring why it’s still very much an art—despite rising interest in AI, B2B e-ecommerce and automation. They discuss the complexity of matching vague inquiries to real-world parts, the limitations of current tech solutions, and how tools like FCH SourceFinder and the “Scrubber” are paving the way. It’s a candid, forward-looking conversation on the digital future of industrial sourcing.
Run time: 01:15:18
—
Listen to the podcast here
FTR Unthreaded: The Art Of AI B2B Fastener Sourcing With Sam Aldinger And Anders Green
Fully Threaded Radio: The AI-Powered Sourcing Platform
This episode is a conversation that I’d say it’s about time we had. I’m joined by two very sharp guys, Anders Green with ALOX4, and Sam Aldinger. Many of you know him from his long career at Fastenal. We did an episode with him. He’s operating under the Highland Consulting flag these days. Both of these gentlemen understand fastener sourcing quite well and bring a lot of technical experience to the discussion.
Anders, as he grows his fledgling ERP system, which is 100% cloud-based. It’s being written as we go in the era of AI. Sam, who’s concentrating very heavily on B2B eCommerce, which many people are discovering is quite a bit different from B2C eCommerce. Of course, the fastener industry unto itself is a unique animal.
During this conversation, we talk a lot about what we imagine a truly AI powered sourcing platform would look like and what it would need to do. Along the way, Sam and Anders freestyle a little bit in their various areas of expertise. Drawing from our different backgrounds, it turns into a pretty far-reaching discussion. If you’ve wrestled with some of the problems, some of the challenges of utilizing AI to do what you need to do in your fastener B2B eCommerce journey, you’ll take away a lot from this conversation if only to commiserate on some of the roadblocks that you’re hitting.
On the other hand, maybe you’re way ahead of us and you can laugh down your sleeve at us. I don’t know. We don’t have all the answers here. It was a completely unscripted dialogue that we had. Originally, by the way, it was intended to be an experiment into AI video podcasting, which is something that I’ve put off for a long time, and I’m still really dinking around with the video for this. I decided things are moving so fast in the AI world, I better just get out the audio, which I’m very comfortable doing.
The title sponsors of Fully Threaded Radio are Brighton Best International, Global Fasteners and Star Stainless. Fully Threaded is also sponsored by Buckeye Fasteners, BTM Manufacturing, Eurolink Fastener Supply Service, Fastener Technology International, INxSQL Software, J.Lanfranco, MW Components, Solution Industries, Volt industrial Plastics and Worth Industry USA. This episode with Sam and Anders is probably not for everybody. However, if you have anything to do with driving your business sales team or technical team, you’re going to pick up a few ideas, I’m quite sure. I appreciate you clicking in here. Here is my conversation with Sam Aldinger and Anders Green.
—
I can also understand that those people sometimes don’t realize the depth of the infrastructure stack that they’re relying on.
The infrastructure stack. Yes. Anders Green, I rely on you to bring up such matters when we’re talking about podcasting.
Within the first five seconds. That’s where we’re starting. That’s right.
I think we can say, as much as the world has always run on an infrastructure stack. Traditionally, it has been more like boats, planes, fuel, energy. That has been the infrastructure that we’ve relied on to move everything in the world. Fasteners, media, everything. If you look at the traditional media distribution, it was spools of film being mailed to theaters. Distribution has always been a key to that.
It’s just that now, so much more of our distribution stack is technology, databases, internet connectivity, storage, bits, servers. When you see the truck broken down on the side of the road, it’s easy to be like, “Those guys are having a bad day.” Flat tire. Everything’s hot. I feel bad for those guys. I’m not going to get so angry at the store because they don’t have my stuff.
It's easier to ratchet up the frustration when you can't see why something is broken. Share on XWhen it’s just a blinky red light inside a Google farm somewhere, there’s a lot less empathy and sympathy happening because you’re like, “I don’t know, dude. All this stuff’s right there. You still have air conditioning. Why isn’t all my stuff working?” I think it’s easier to ratchet up the frustration when you can’t see why something is broken. That’s my opening hot take.
The Infrastructure Stack: Why Google Goes Down
Anders Green, you’re with ALOX4. You’re a newer ERP on the block trying to make a toehold for yourself out there in the industry. I’m sure that as you go to the shows and the association meetings and all that, as you insert yourself into this industry, which by the way, you’ve done a pretty good job at these last few years since I first started seeing you show up here and there. You’re probably still talking to guys. There’s not many left anymore, but there’s still running their business on CardEx.
I have seen at least one place that I’ve visited in person that was still literal index cards.
Anyway, there’s a couple of them here in the Cleveland area too. The only reason I bring it up is because those guys, when we have these brownout situations, they get to act all smug and self-righteous. I’m not advocating for that but it’s definitely something I think about a lot. Sam Aldinger, you’re, of course, tucked away up there in the Northern climbs of the country, Winona, Minnesota, homesteading yourself to health and happiness. Good to see both of you guys here. Thanks for jumping on.
Thanks for having us.
Good to be here, Eric.
We tried this last week and I thought it was me, or at least something that was in our tech setup. As it turns out, there was a global wide Google brownout. I guess some of their data centers, I’m not sure if it went over into the Amazon Cloud as well, but Google’s was on its back for a day. With all the geopolitical upheaval, shall we call it nicely, going on, it makes you wonder if that wasn’t a little bit of setting the battleground or something. Anders, what’s your take on that?
I’ve actually had a little bit of research into that stuff. It turns out that there was some new technology that they deployed that had to do essentially with some of their settings and how they handled the load of various things. Normally, when they do that, each piece of that is hidden behind a feature flag. They turn it on like 1%, 5%, 10%, 50%, everybody. This particular feature was not built with that protection, and it also had a problem.
It immediately caused like a boot loop and just tore down everything. The good news is Google was pretty on top of it. They managed to have it diagnosed within about 10 minutes and then had a fix for it in about 40 minutes. It took several hours for that fix to propagate throughout the rest of the cloud and get back to where it was supposed to be.
A little blip. In my experience, to your point about the geopolitical, Eric, the immediate assumption by most individuals is that there’s a bigger cause behind it. In my personal practical experience, I’ve been on the receiving end of, I know exactly what happened on the backside, but then you have people within the company coming and saying, “There’s so much of this stuff going on and we don’t see it all.” It’s like, “No, Jimmy forgot to write the right code.” He screwed up and there’s no ulterior motive, there’s no anything going on.
Jimmy wrote it and then it was right, but then he commented that one piece for testing and he forgot to uncomment that piece out when he deployed it. “Jimmy, we’ve told you before. This is actually the third time, Jimmy. Come on.”
Small Vs. Large Companies: The Impact Of Scale On Business
I’ve been there, believe me. In my application developer days, I’ve seen lots of those, but nothing on the scale of this. When you’re working at Google scale or remember when the whole Cloudflare thing happened, when they propagated that whole Microsoft upgrade, the server upgrade that went bad. 10.5 million servers worldwide were affected in about an hour. These are the kinds of things that, it’s a scale I can’t even wrap my mind around.
Of course, this leads me into a topic that I wanted to talk about with you guys, because I’ve been working with you both individually and we’ve been doing a few team ideas here lately. What strikes me is, of course, Sam, you’re used to working in a large corporate system, shall we say, or an ecosystem where you’ve got people to do this and teams to do that. You’re looking downstream, upstream for everything. There are guys like Anders, who’s really a one man show in a lot of ways, and two opposite ends of the spectrum. When we have these conversations, it’s interesting to see the different perspectives and how that impacts whatever we’re talking about. Whatever project idea.
Absolutely. Scale makes a huge difference. When I worked at IBM and the cafeteria that I ate at served 5,000 people every day and was 1 of 3 cafeterias on site, we did things very differently. Now, much smaller team. Certainly, I can come to work in my shorts. Scale’s very different. Where Sam has been there’s certainly more layers involved in all that management and bringing a more powerful team in at that scale also changes all the capabilities and what you can get done.
I would say, as with anything, there’s good and bad along every step of the maturity curve. When you’re young or when you’re small, let’s just say in this context of let’s just say development or business teams, whatever, you can be very nimble. You can make decisions very quickly. Essentially, you can show progress and make these very small, fast, incremental wins.
Good and bad exist at every step of the maturity curve. Share on XAs you add more teams, as you add more structured processes, you add more tape, but you also, to your point, scale, have the ability to present itself and become a more sophisticated structure in which you have safeguards. Somebody accidentally unhid the comments in the code or whatever you had mentioned. That doesn’t happen the more structured you get because you do allow these processes to come in and govern risks like this from happening.
That can be a really, really good thing, but it becomes a lot more complex. What happens then is this individual over in this team doesn’t fully understand all the way to the other end of the side a spectrum what’s happening over here. You have a lack of appreciation for all the different functions that happen. By nature, you have to work very hard to keep a cohesive unit going in the same direction. It just presents itself different challenges.
Again, you have more resources. You’re able to really master your craft in certain areas. You can focus and have really strong subject matter expertise in specific areas like, for example, introducing parts into your system. I’m going to go into the product data side. Putting streamlined processes for capturing really good mature data upfront within your part activation process that never happened before with a small team. You just were happy to get something filled out and in the system to use.
It’s like a parent with a child. I wish we were still at that age, or I wish that we were in our infant eras, but there’s a lot of things you didn’t like then either that you wish you had. You got to enjoy every step of the way for different reasons. It creates new challenges. It creates your own personal growth, etc. Trying to step back out of that, I learned to appreciate that while not getting caught up in the thick of it and forgetting perspective is always a detriment to people.
There’s nothing more magical than watching your child take their first steps. When you’re trying to walk down the beach and the kid has no speed, and you’re like, “Come on, we need to make it to where the sandwiches are.” You’re like, “I wish he was fifteen and we could just go there.” Every piece of the scale has different strengths.
Some of the greatest humor that there is is smaller companies trying to appear big and vice versa, where some of these massive conglomerates are trying to appear down home and they just completely fail. They miss it. There’s a lot of that with social media where they’re trying to appear like, “Your friends from Monsanto.” It doesn’t fly.
I remember when Brian and I were really trying to present the illusion that we had some bandwidth and some team, and we were doing work for actually some pretty decent sized banks, and they took things very seriously. They weren’t used to dealing with barefoot guys working in a two-room apartment and passing themselves off as a bonafide technology team. That’s what we were doing. I remember my test user always used to be a fictitious test user, Spanky Baez.
I put all these prompts in some of my code, and it was all there just for error handling as I was stepping through some of the code. We had a big release the next day. It was like a beta release. I remember being in the background of this call and the business leader who had hired us, who was presenting the this to his clients, we were supposed to be the men behind the curtain. I remember being on the call and I wasn’t watching, but I was listening, and I remember him saying, “Who’s Spanky Baez?” Just getting that that feeling like that dread and it happens. We got busted many times, which explains how we wound up in the fastener industry, I suppose.
You and Brian have been able to stand up. It impresses me, just in terms of you’ve created a really good platform, essentially, really, on the backs of you and Brian and your small team and what you’ve been able to achieve is pretty impressive.
I appreciate you saying that. There’s a lot of proverbial bailing wire and duct tape back there holding it together, as well as a couple of hamsters on wheels. It gets the job done, and it’s been running for a while. That brings us to a subject that we should probably touch on, Sam and Anders. When I started off the conversation here, I was talking about our collaboration that we have going, and we’ve been talking a lot about inquiries and FCH Source Finder inquiry program that we’ve been thinking about a lot for different reasons. We’re trying to really turn this thing into a much more useful tool for the industry. I look at it, and I think Brian shares this idea, but we look at it as a very late-stage prototype.
Source Finder: The Underappreciated Art Of Sourcing
Source Finder’s been running is a system for, how long has it been? We probably released it in 2009, I would say. It’s been out there for a while and we’ve helped out a lot of people with it, but we’re looking to make it better. Anders, you’re running ALOX4 and you’re building a module into your platform that’s a ground up Source Finder collaboration. Sam, you’re doing all kinds of things with B2B eCommerce. You’ve got a very large swath of experience in eCommerce as it applies explicitly for fasteners. We’re trying to see what we can do with this thing.
Interestingly enough, Source Finder, if I’m not mistaken, the intent is that the user comes in and can’t find the specific product they’re working for, but then ultimately says, “I need help finding who can get me this.” In my experience, that’s a sourcing function. I had Source Finder. You are sourcing at this point. You’re not actually creating the acquisition, you’re not creating the transaction. You are trying to find and match a supplier or find a provider of this set item.
In my experience, and as I’ve gotten into B2B eCommerce in general, is that this sourcing function within B2B is a very underappreciated function in that everybody downplays how much time and how frustrating it is for a buyer to go out and find something. The process itself is very manually intensive. It’s one thing if you’re talking about a Milwaukee tool.
If I’m going to buy a cordless drill that I could pick up at Home Depot, Lowe’s, whoever your local hardware store is, whatever, you can go just as easily online and find that same cordless drill. Sourcing that item in and of itself actually isn’t that hard. You do a Google search, or you have a model number or something like that and Google or whoever provides you with results, and you can pick whatever marketplace you want. There’s your digital, eCommerce, more of a B2C platform, but you can buy that tool.
The sourcing exercise there is easy, and that’s what 99% of people in the world, in my opinion, thinks as sourcing. Now, if you were to give them this paperclip that I play with, that’s my little fidget whatever, and you ask them to go source a paperclip, you only have a description. You don’t have a part number, you don’t have anything, you have a description.
Everybody’s attempt to describe this thing can vary unless you really know paperclips. You’ll know the gauge of wire or whatever’s used. You’ll know the style of paperclip and you’ll call you. Very few people in the world are like that. You set somebody like myself to go find this paperclip just like this, “Find this online,” I would spend probably hours if my job was to find this exact paperclip.
I have to be confident in the data that I’m looking at. I have to sift through the myriad of non-existent information and try to piece it together. I don’t have brand information that’s recognizable. I could guess, but I don’t know. I have to start measuring this thing with my ruler, measure it in a different way where it’s millimeters versus inches. I don’t know if whoever made this at millimeters or did they make it at inches or what.
When it comes to commodity, and so all of this across fasteners, in my mind, it adds a whole level of complexity of sourcing that is very underappreciated in the market. In my humble biased opinion, it’s all about presenting good, accurate data. That’s from the supplier side or from the platform side. From the customer side, it’s actually requesting good data because when I’m a customer, I’m going to go on and I’m just going to type paperclip. That doesn’t tell you anything from a data perspective.
If I typed in paperclip, I would just have to peruse through paperclips and try to find manually what paperclip seems to match what’s in my hand right now. While not all B2B business is like that, there is a large piece of maintenance and of scrambling that happens in every day for items that you don’t have a lot of data on, or that the user has to input data on in order for it to match the receiver or the platform or the manufacturer or whatever.
That is a huge art that I’m finding that is almost non-existent because most of the B2C companies are not used to this type of shopping experience. They want to give you their best item, they want you to take it, and they think that your application isn’t critical. They’ll take whatever’s close. You can’t do that in most industrial applications, you just can’t. When suggestions are made, they’re made in a vacuum, not knowing what application you’re using, the criticalness of it. Can we swap a grade 5 for a grade 8? Can we change a material? Nobody knows because you don’t have customer information.
I’m starting to wonder if I had European like A5 sized paper, would I be allowed to use a paperclip that was measured in English units or would I have to get a metric millimeter paperclip? These are questions that might come up. Is it even going to work on that crazy metric paper?
Those are all issues that swirl around. FCH has always been, and I think it’s always going to be a commodity driven platform. In other words, traditionally, you haven’t come to FCH and say, “What does ABC company have?” You’re looking for a quarter 20 grade 8 machine screw with a specific head and other specifications and stuff. We show you who’s got it based on what you said. That’s a simple example. Of course, what you said in your example is a lot more exotic.
What I like in what you just said, though, is that you referred to sourcing as an art. I dig that. It may be an art that there’s a science to it too. It’s another one of these things where you could argue about it forever and that’s not very productive, which is probably why I’m attracted to that topic. I will say that out there in the world now, what we take as an assumption in everything that we do at FCH, and I think what we’re all trying to do independently, and as we collaborate here, we are facing the fact that the generation of professionals that we’re supporting increasingly are less adept at this art. Therefore, they need tools such as we’re building, such as these AI empowered, perhaps, tool sets that we’re envisioning and that we’re beginning to build to assist them.
Exactly. FCH, in my mind, is positioned in a way in which you are defending the requester, or you are working on behalf of the requester, meaning you’re learning how to guide them through that information process of what they’re looking for. You’re trying to give them an experience of which, if you guide them just enough, you can provide them really good results of here are potential options.
To the supplier community, the folks with the products, with the capabilities, whether you’re a manufacturer, whether a distributor, the call is, what are you doing to digitize your data? Do you have your manufacturing capabilities in a way that can actually be read into a digital environment? Do you have your product data clean? Are you thinking about how to present that product data simply to the person looking for it, or are you assuming they’re going to do it?
I look at it as, again, you have the requester inputting data, and you have the re the receiver providing data and the goal is to match those two. The responsibility of the provider is to digitize their data. That is all about capabilities. If you’ve got a CNC lathe that goes from quarter inch up to 2-inch chuck with 28-inch length max or whatever, how do you actually turn that to data? What does that mean in terms of a fastener in features of a fastener or a component, diameter wise, material wise, whatever.
All of those are data points. It’s all of that requires that information to be now connected to the requester. The requester says, “I need a metal cylinder two inches by whatever, with a hole in it.” That very vague request needs to be somehow mapped into capabilities. That’s a very farfetched example. The reality is there are a lot of components out there that, again, the buyer doesn’t care about. They just know they’re being asked to buy it. How do you help that process along to be very quick to be very easy and to be painless. Sorry, Anders. You were going to say something.
I was thinking about how that connects with what Eric was saying earlier about the art of it. I think for some of the wider audience the idea of, like Eric was saying, if you have a part number of a Makita drill, you’re like, I can Google that. I can find that, and there’s no art or there, but also no demonstration of expertise because you are looking for a particular thing. You’re not looking at the end goal, which is what you’re talking about, Sam.
You’re saying that not only do I have to understand what it is I’m looking for, I need to know what am I trying to accomplish? Certain things I might need to bring my experience in and say, it’s not only going to be the size of this thing, I also need to know what is your manufacturing capability? What’s your timeline? Potentially, what is your material sources?
The only reason I know that is because I overheard that my buddy Tommy was saying, the general needs this, and so it has to be us domestic. All of that stuff, you’re not looking for the Makita M235-B and that bringing that experience into it, I think is not obvious to someone that is outside procurement. I was thinking that maybe for many people, if they’re outside this, the more down-to-earth example of the complexities of procurement. Disclaimer, I’m not any a foodie. I like eating food just as a human being, and I’m not a huge cook either, but I have one grocery store.
I’ve seen you eat Anders.
That brisket was delicious. I’m not going to deny that. It was amazing. If I want to go get some good roast beef or fresh tomatoes or apples, I go to this one grocery store. If I just want like rice and eggs, I find that this other grocery store actually has a better selection. This is the most basic thing because I could say, “Jonathan, go get rice, eggs, roast beef and apples,” and you would think that would be easy.
Yet it turns out I’m driving four miles in two different directions just to accomplish that. That is because of I’m valuing essentially the supply chain of the 2 grocery stores and how good they are at bringing in those 2 items that you would think would be commodity. You’d think anyone can get apples, but it turns out, one of these guys just sucks at it.
You bring up such a great point, Anders. Everybody has the ingredients for the pie, but who makes the best pie? In the case of the sourcing of buying the right ingredients of who’s got the best apple, there’s actually a lot of experience that goes into finding the right apple for the best apple pie. Not every provider is the same.
Actually, this now speaks largely to my first point about we’re not talking about the Makita or the Milwaukee tool. We’re talking about a product that can be provided by a myriad of suppliers. While they all may be able to provide you that finished product, the steps along the way are what make them the right supplier. It could be quality driven, it could be price driven, it could be service driven. These are all factors that the buyer, the procurement person has to know as it relates to their specific business, which all of those sacrifices along the way of learning about product data.
This is why you have eight different kinds of low carbon grade 2 bolts out there because somebody was sacrificing maybe the quality of the material because they just needed a price point, and it was for a very low application. It wasn’t a non-critical application. Somebody who has a very engineered specific application may not need a grade 5, but they’ve engineered this grade 2 per spec into this application. They need it to work exactly like that.
That just cut out 15 of the 20 suppliers that provide this thing because they can’t really make it to that spec. Only the buyers who are in that situation have understood the pain that’s gone through the learning curve of knowing that. They’ve been burned a few times, and so they’re looking for control within their processes to make sure that that never happens again. That usually comes back in the form of data.
AI In Fastener Sourcing: An Unmet Frontier?
They’ve literally ended up with bad apples. I think a lot of the nuances of the sourcing process are known to you guys for sure, and to a lot of people who possibly would be reading this conversation. I think that it’s also pretty self-explanatory that the world is looking to AIs or technology that’s calling itself AI, let’s put it that way right now, to address some of these issues.
Given what we said earlier, that our crop of digital native sourcing professionals are relying more and more on this tool. That puts the challenge on guys like us to figure out how do we actually go about training this technology to achieve what we’re describing here? This is what everybody’s standing around looking at each other, waiting for somebody to make a breakthrough. In the realm of fastening, I’m talking about you hear amazing stories every day about applications everywhere. I haven’t really heard one that’s impressed me so far in this area in fastening. Have you, guys?
I’ve been looking.
I’ve been looking so hard. Admittedly, my focus has been looking for AI success stories within the ERP space and within the software and technology space. To find applications that are actually like, you can buy this now and it exists now and you can use it tomorrow, and it actually does something, that list is super small. It is not nearly as widespread and functional as the press releases would have you believe.
There’s some invoice processing, there’s some marketing stuff, there’s a little bit of that. The stuff that I’ve seen that is like, “We are fixing business processes,” very long on announcements, very short on me seeing those show up as results in anyone’s financials. Sam, who’s knocking on the door? Come on, you got someone.
I haven’t seen one yet. I’ve seen a lot of promises and I’ve politely just asked, “Show me your tool.” When I start asking the hard questions, I get crickets and a response because every sourcing solution that I’ve seen come out has essentially required one thing ahead of time. That is a catalog. Meaning, in a B2B relationship, you’ve got a customer and a supplier or a customer and a vendor. There are terms within that relationship where you’ve got a prebuilt cross-reference catalog. You’ve got a preset catalog.
Every sourcing solution essentially requires one thing: a catalog. Share on XAny of the sourcing efforts that have come in has already matched points A to B, i.e., points supplier, part number to customer part number. There’s assumption that if I take a picture of this item, I want to take my phone and snap a sourcing request. It presumes that there’s already an existing catalog in place between those two companies, essentially between the manufacturer and the supplier relationship.
By nature of the scenario, the reason that that buyer’s taken a picture is because it’s probably not part of their catalog. It’s something that they need on the fly, and they need help finding the best source. Now, usually in those cases, price is the least amount of their concern. They need the product and they need it quickly, and they just need to find the right supplier.
The sourcing tools themselves are not helping what a human would do of the hours and hours of scrubbing through websites to find that picture match to the best likely source and then provide a solution. I’ve not seen that. Again, within the context of this, I’m going to keep it to fasteners or commodities in that regard.
You could include electrical components, you can include some of the HVAC in an industrial type OEM component that are out there. I think they’re fair game in this space as well. I’m not seeing that that’s being done well. There’s a lot of presumptions, like I said, that are required in order for that to actually be successful. This is a call if somebody’s got a tool, DM me on LinkedIn, Sam Aldinger.
Industrial Marketplaces: Data Quality & Merchandising
You did a lot of work for Fastenal building out their eCommerce platform, and I believe it’s really more of what we’d call conventional technology foundation on that. It was never touted to be an AI platform. It may be, these days, growing into that. I have no doubt about that. Fastenal is a giant and they are a progressive company. I guess that’s really side to my question. You must have been looking over your shoulder at what Grainger’s doing with Zoro. Have those guys been letting any AIs go that you’ve seen that, that are doing some good things?
I haven’t seen any within those spaces. Most of what I’ve seen is a fairly arithmatic, “You’re interested in item X. Other customers have bought YZ and I and J and so we’re assuming that those are related because they tend to get bundled together.” I haven’t seen super AI within those platforms yet.
Sam, have you come across any?
I have not. What I’ve seen across the industrial marketplaces in general is just a focus on data quality and actually more on what I would call product display or merchandising. McMaster has historically been a dominant player in fastener product merchandising. Anybody sourcing for OEM components usually likes the way McMaster lays out their website.
Grainger’s made pretty significant milestones or progress in their fastener data. Somebody’s clearly focused on that. In terms of tools for sourcing, again, I haven’t seen it. Usually, when you go onto the folks like Zoro or whatever, now it’s a product search game, but you’ve got a captured audience in that case. You are only buying from one company, i.e., Zoro and all of their suppliers, obviously, but you’re buying through their marketplace. You’ve got a captured audience.
You’re also dealing with a user community that expects to have elastic search and they expect to have really solid product recommendations and things like that. That’s all out there in really the high-octane eCommerce platforms already. I think what we agreed to here in this conversation, and we stumbled into the fact that the AI enabled sourcing, or at least just let’s say full featured sourcing is still a unmet frontier.
I think you’re correct because if you say, “I want a fully AI sourcing,” you’d be able to say, “I want a bolt that’s about the size of my pinky and needs to go through like two super strut channels, the regular size.” It should be able to figure out, “He is talking about one and a half inch square and his pinky’s probably a three eights because that’s normal size of the slots on super strut,” and figure out all that stuff for you. That would be the equivalent.
If we’re saying artificial intelligence, I should be able to have that conversation with the guy behind the counter and he’d be like, “You want a three and a half inch three eighths bolt for that,” because I had described it. There’s nothing in AI that is close to that level of comprehension or bringing that amount of experience to the sourcing effort.
Until we get a large amount of fastener people with Neuralink where we could just tap into the global brain that where all this stuff is already just out there, until we get to that point. What I’m fishing for is how do we train our AIs to do this kind of work? This would be the secret sauce. I don’t think we’re going to give any anything away here because I think that a lot of people have probably struck the same question, and if it was easy, they’d be doing it.
Has anyone gone the route of taking some of the spec manuals and dumping them into language models?
Yeah, I’m sure that’s been done a lot of ways by a lot of teams, but again, we don’t see the fruits of that anywhere, really.
If somebody’s doing it, I’m not seeing it’s proprietary. Somebody’s keeping it for an in internal use or something. I’m not seeing it as a public domain or a public service.
Somebody’s got to be taking a stab at that, even if it’s just for more of a vision of a more conventional a marketplace idea.
Even the spec sheets, all that does is create a massive data library. It does not tie it to product. There’s a search connectivity piece to this. I’m going to have a 0.25 by 1.25 hex head cap screw. You have to turn that digital inquiry to something that is in that library. You have to do a lot of transformation to that data to make it tangible to the buyer.
There’s just a lot of work there that almost needs to be built and pre-structured. Again, AI is actually not that great with structured data. AI is very good at unstructured data. The training of AI actually has to train the user, or anyone successful in here would’ve to train AI on not looking at the whole, but actually following the logic of the structure in these little pieces.
This isn’t an AI thing, it’s actually more of a search thing. There’s a lot of conversation out there on the AI and search and the differences between them. Again, you then have to also match that to who’s got this product. Not only do you have this large library of spec information, you have to actually create physical product and you have to actually tie it to somebody who provides the product. There are multiple databases that need to be created in order to take on this feat.
If you were to look at this in your technical mind, Anders, this is a high complexity, high value proposition. This is a high complexity, low value proposition. This is a low complexity, high value. You put it in the quadrant of this is high complexity, low value, at least as it’s perceived in the industry. I’m saying the broader industry. I’m talking all of B2B, you’ve got not only industrial distribution, you’ve got the construction market, you’ve got the medical distribution, you’ve got many of these different industries out there. This type of a challenge seems to be pretty unique to fasteners or to commodities. It’s low value, high cost or high complexity.
The Low Value, High Complexity Of Fastener Sourcing
I don’t think it would be common for you to hear we took the best guy in the company and we put him on finding someone to sell us bolts. Even if it may be like, “It is going onto a literal rocket. We need the right bolt there. I think you’re right that the perceived value is low because someone up top is like, “You just buy it. We just buy it from somebody. Isn’t that what we do?” Even when you find it, like you were saying, Sam, all three of us are customers of McMaster on at some point.
With McMaster, the idea that I can be extremely sure that what I believe I am buying is what’s going to show up, the categorization, the detail, the individual pictures, the grids, but also I’m like, “Four bolts for $38?” I get that they were especially a brass or like, whatever, something. If I wanted to buy 40,000, yes, McMaster has them. I would never go to McMaster for that.
They wouldn’t take you. They would tell you to go someplace else. Everything I know of McMaster is they don’t want that type of order. They’ll actually direct you to their supplier.
That’s exactly what you’re saying, the difference between the search part of it and then supplying as far as matching up all the other pieces of that business need, the quantity, the timeline, the delivery, the supplier materials and everything.
On one side, I say that, but then I’ll also think about the countless individuals out there in this industry that are having to do this on a daily basis. While it may seem small in the broad industry, there’s a massive amount of labor hours that are being consumed in the consternation of this process of making and matching their company’s needs to the right supplier at a reasonable price in the right time.
Nobody wants to pay the $38 for the bolt, but when the machine’s down, that $38 has nothing compared to the machine being down. It doesn’t matter. How do you help that individual find that to keep the machine up? We’re not talking about the tangible peace price anymore. We’re talking about general machine time or loss of revenue at a company because of what they couldn’t have in the right time.
Good companies will monitor this and track this, but do they put it on a sourcing budget or do they tie it to a sourcing KPI? No, they really don’t. Not that I’ve seen. They look at it as just the cost of doing business and they’re not directly attributing it to maybe these sub parameters that would cause that machine to go down or whatever. That’s just one case or that production line to goes down.
That would be a pretty interesting report to read. Let’s say you had something like a power plant, power company or maybe like an aluminum smelter that like we deal with so much electricity and every 10 seconds of downtime is $200,000 lost in productivity. I have no idea where the numbers really are. If we spent an extra $100,000 on procurement talent each year, how much would we recoup on reduced downtime? It’d be a fascinating analysis. You’re right, Sam. I’ve never seen someone say, “Our downtime is X and we’re going to fix it by doing this in the procurement department and inventories and stocking and inspections,” or whatever. They’re mostly just looking at the overnight FedEx bill.
Reminds me of the old the old joke about the boardroom meeting where the team is developing a very complex, very high risk, but very high possible reward project. The senior guy says to the junior guy before they break for the day, “I think we’ve thought of everything, but why don’t you come up with a list of all unforeseeable circumstances and then we’ll present that tomorrow morning before we go ahead?”
That’s the thing that you’re basically asking these tools to factor in. I think that’s why everybody’s looking at the ceiling when they actually think about how you do what we’re envisioning doing here. The more you peel back the question as we’ve been doing here during this conversation, the more you appreciate how multifaceted that fine art of fastener sourcing is. The practitioners are quite valuable team members, aren’t they?
The Power Of The Fastener Industry Network
I think so. I think that this discussion of nuts bolts and commodities probably speaks to why the fastener industry is so collegial in that everybody gets together and you talk about, “What can you do? What can you build?” It is not digitized. It is such an unusually shaped set of variables that where we are now, the current solution is I need to know that I’ve got these seven suppliers that do these different things, and I know that I can count on them because they’ve done this stuff for me before.
I know Bobby over here, and I know Davey down there, and Johnny always helps me out with the stainless stuff because he’s a whiz on that stuff. There’s so much legacy in this industry because we have machines that have been running for 80 years still churning up bolts with absolutely super quality. There’s not yet existing a fully digitized, fully digital twin of all these industries and factories and capabilities. We’re just down to like, “I know the humans, and I’m going to call him up.”
I’m going to play devil’s advocate because, first of all, I agree with that. The devil’s advocate perspective here is, while that is the case, the customer’s demands don’t change. Industrial distributor, I need cost neutral on tariffs. While that may be unreasonable, any good customer worth their salt smart customer is going to do that. They’re going to expect that you as a distributor are servicing them by ensuring resourcing activities, by ensuring cost downs, by leveraging your procurement arm and essentially mitigating industry events like tariffs. Tariffs isn’t the only thing, but it’s just hot in our minds.
The practice of the industrial distributor is to ensure that they’ve got mechanisms in place to respond to these events in a way that is efficient and effective. That requires resourcing. Not only are we looking for this most simplistic perspective of I just need this paperclip. That’s easy. In the grand scheme of things, when you’re talking production fasteners, when you’re talking OEM components, these are critical.
While they may be a smaller cost basis of the overall system, they still are a cost of the system. When they’re going up 15%, 25%, whatever the customer’s asking, “How do you mitigate that for me?” Resourcing is actually, again, an art of how much drain in your labor are you going to spend on resourcing? That is a very labor-intensive process to create bid events, to go out and find new sources, to qualify those sources and even to change your buying over to a new source. There’s a lot of things along the way that are very disruptive.
A lot of friction in booting up a new supplier. Sure.
There’s no tools out there to help in this, in my mind. The best tools you’ve got in this case are buying groups essentially that say, “We’re going to try to do this on your behalf for you. Just give us the easy button.” Are they really giving you the easy button or are they just controlling the purchasing and leveraging it? That is something you could ultimately do as you grow. Anybody who thinks through vertical integration knows that those buying groups make money. They’re not passing this service on for free.
There is the room on the table now they have much larger volumes, but how long are you going to stay with that buying group then before you realize you can do some of that yourself? Don’t assist in all of the sourcing functions. They just try to give you a means to say, “Give us all your data and we’ll do it for you and cut us a check, and we’re going to take 15% off the top,” or whatever that is.
I just see this as I’m still yet to see tools that come in and help procurement in sourcing, in leveraging their labor, in optimizing their needs, which are cost downs, quality improvements, supply chain disruption. These are all sourcing reasons, reasons why you would actually go and source to the benefit of the buyer, to the benefit of that organization in a way that that helps them.
I’m talking a little idealistically here. You don’t do that overnight. You can’t boil the ocean. You have to take it in small steps. That’s where the industry, the comradery or I can’t remember what word you used, Anders, but collectively, manufacturers and distributors have to work together on this process. Maybe it takes a middleman. Maybe it takes an FCH to help make sure that the best interests of both sides are in play.
Not that it’s all manufacturer biased and you’re just trying to steer towards 1 or 2 companies that are paying you all the money or you’re not all buyer focused where you’re just trying to squeeze every price penny out of the manufacturer. You’re here to make tangible connections in real life industry needs around sourcing and can you do that in a way that benefits the industry at large, not just comes and sucks one entity?
It wouldn’t be hard to imagine a massive vacuum trying to suck all the value out of that because we’ve seen that in other industries, whether that’s conglomerations or buying other businesses out. I think when we look at FCH, what we see that I think is certainly the seed of the solution here, it is an ecosystem and it has the network effect.
Similar to any social media thing where you look at Facebook or any of them that a big part of what’s going on is it works because there are a billion people on there. You look at FCH, and I don’t know what the number are, there are thousands of different companies that are already in that ecosystem. I think that’s the only way that you’re going to get the required volume of network effect is to have that.
It would be easy to imagine some perfect database, but you’re like, “There’s three companies using it.” It’s not enough. All using some universal spec of fasteners, which I’m sure Mr. Brian Musker would tell us does not exist and possibly cannot exist. Until then, I think a collaborative effort like FCH has going where everyone is contributing data but also benefiting from that data existing, I think that is the route toward having, having some solutions. At least I feel good about it, the direction it’s going.
Having been part of manufacture distributor relationships for majority of my career, there’s definitely value in data that the manufacturers could glean. There’s definitely value in data that the distributors can glean, both from sourcing activities as well as selling activities and in understanding complementary aspects of your business. There’s so much information that can be used to help develop a sales runway, a more robust supplier community, supply base for risk mitigation. It’s a two-way street for data and data sharing. There’s real opportunity in leveraging that and benefiting that, to your point, as an industry, as a group of folks that can benefit from this. Teamwork.
Fasteners Clearinghouse And ERP Integration: The Future Of Sourcing
I’m right on board with the basic idea that you’re discussing on this part of the conversation with the network and the power of it and the various derived benefits that it offers. I think as the technology-enabled world evolves, the way we think that it’s going to over the next 5, 10 years and beyond, I can’t even imagine what it’s going to be like after it really takes off.
A lot of people who are thinking about these things are wondering how they’re going to differentiate themselves. Even the old standbys like engineering expertise and things like that seem like they’re going to be in danger. How do you dig a moat around something like this like we’re facing? The network, if it’s a unique, organic entity unto itself that could fit the bill, one of the aspects of the network, at least the one that we’ve built, is that it has these integrations with other networks.
I’m speaking specifically right now of the networks of ERP platform users. We’ve integrated with different ERP platforms. Some of them are the standards for certain segments of the industry, mid-size distributors is what I’m talking about at this point. Anders, you’re really making a name for yourself in that niche, and you’re building a module to support Source Finder from the ground up. It’ll be native to your platform. What do you see as the benefit for your platform in doing this? How do you envision building a thing like this?
There are a couple pieces to that. We’re building that, we’re building it in, we’re working with some of the new technology that FCH is bringing forward. I don’t know if you’d call it a version three or version next or whatever of the next round of technology improvements.
We just call it Luigi.
We’re targeting the Luigi platform. For us, what we see is that by bringing this in as a little more than, “I just want to find this bolt,” to do some slightly deeper integrations to say, “If I have this set of tags that exist within ALOX4, maybe I’d match those tags up to various categories within FCH to say, ‘I’m going to allow some mapping back and forth so that when I am looking at my internal tags of stainless 316 hex cap, I already know that that’s going to map over to the FCH categories that they have in those areas.’” beyond that, we are looking at connecting to slightly more than just the sourcing to enable some deeper RFQ integrations because usually, it’s a rarity that someone’s looking for just one thing.
You don’t need just the bolt. You usually also need the nut that goes with it, and maybe a variety of lengths or the washers or whatever else. It’s rare that just one piece is the only thing that you’re needing. We want to work on the idea that we can find the connected pieces. Once we have one piece, show me the related pieces, build up a multi-item RFQ for that, and then tie that into the vendors that we have within ALOX4 itself.
Also, if you’re finding new parts from a supplier that maybe haven’t used before, can we pull the data from FCH and say, “Let’s create a new vendor. Let’s get some of that paperwork done automatically based on what FCH already has. Let’s pull in those part numbers that we’re trying to RFQ for, match them up to our internal part numbers, build those connections automatically. Get all of that stuff done so that we can send the PO over without all the manual work of doing all of that stuff piece by piece.”
As Sam says, there’s a bit of a grind to churning out that paperwork. We want to integrate that a little more deeply. I’d say we have an advantage in this area just because our stuff is fully cloud-based anyway. All the servers live in Sky and having that full-time connection to FCH all the time would let us do that slightly more easy, just because this goes all the way back to the very beginning of the conversation. The technology stack, the infrastructure stack is built for connecting all that information. That’s what we’re working toward, to take that next step from just the search, from just the finding and then do an assist in the sourcing procedure. That’s what we’re working for.
This is actually interesting. Are you connecting to PIMs at all?
Have not yet connected to PIMs.
In general, any distributor that’s going to have a massive product library will likely store most of that. They’re going to store the product data itself. The characteristics, the nature of the product, what it is, the pictures, all that stuff is going to be in the PIM in general. Now, some erps have basic structure, taxonomy structure, but most don’t carry imagery or anything like that. I’m just curious if you’ve had any discussions in that regard, because now you’re getting another system involved. That’s one, and then I’ve got another point after you answer that.
We haven’t supported pictures yet. This comes back to scale. What we found is that most of the distributors that we deal with like under $30 million, they’re small enough that what they end up doing when you give them the ability to have pictures is they take one picture of a bolt and they copy it 10,000 times, and it’s the same picture. It doesn’t end up moving the needle on making the them more productive.
For the taxonomy of that is that you can go through the tagging, some of our customers do very minimal tagging, and it’s like five categories. Some of our customers are super deep into it, and their tag tree is about 200 levels deep. I don’t know, maybe that’s overkill, but that’s what they want to do. Yeah, they are tagging the material, the spec, the length, is it metric, is it M6, is it 30 mil? Is it inch? They’re all the way down through that. They can do very detailed multi dropdown searches internally.
Coincidentally, this is where I think FCH stands significant value in that they can help validate some of that data or normalize that data for product matching, for making sure that descriptions are consistent. There’s a lot of things I think, Eric, on your side that can bring pretty big value to that, especially if a company doesn’t have a PIM. Even if they do, it would only compliment the PIM because it would help make sure that it translates their PIM data to somebody else’s potentially for finding the right supplier or deduplication. There are many different things that services that could be done around your product catalog.
This is a project that like, just even thinking about, I’m like, this is somewhere between a 10 to 50-year project of the idea?
We’re full of them, aren’t we?
I know. There are certainly many ways that you could describe any particular fastener, and many people have their own favorites. Could you get the industry on board with saying, “This is how we want to do bolts. We are going to put the length first, or we’re going to put the head first,” whatever it was. This is why I’m saying it’s 50 years, because you’d have to churn through so much data over that time to normalize even just the presentation structure of the description of the fastener.
I think it could be done. I think there’d actually be an appetite for it, at least within your integrations. A lot of that can be done behind the scenes. If a distributor or a manufacturer wants to keep their product the way it is and not change the flavor or the design or the marketing aspect of however they want to manage their data, there could be a key on the background that essentially could be done the way you’re describing that would allow people to connect and standardize their products across the industry.
The reluctance you’re going to have there is people want to keep their cards close to the, their chest and not give away their standards so that they can be unique and not give anything away to their competitors things like that. As with anything, the more standard you get, the more commoditized it is, but what’s more commodity than a piece of metal, right?
The Scrubber: Cleansing And Normalizing Fastener Data
I can’t believe it. We’ve been talking for over an hour and we haven’t really even brought up the scrubber yet, because the scrubber definitely plays a role in one aspect of this bigger picture. We’ve had some success in retooling our scrubber system, which has been in development since 2007. It’s just a giant group of libraries and programs that function to identify fastener inventory, cleanse the data, normalize it, categorize it, that stuff.
Brian and I have used that to run FCH and to do custom data jobs for all kinds of companies for many years. This is widely known, but over the last year or so, we’ve done a lot to try and refit it or update it to fit into the LLM idea to bring it into the AI world, and so that it has a machine learning component to it. I view that as a very vital piece of what we’re doing, and it’s coming along really well. It’s got lots of its own challenges, and it’s just a subset of what we’re talking about, but we’re bumping into it here so I’m just throwing it in.
That’s your transformation conduit. Essentially, if the scrubber allows companies to connect without having to standardize per se or it could it help it because it could assist companies in standardizing that want to use it.
It takes a big bite out of the potential universe that we encounter. As we know, there’s still this giant pile out there that it’s made to print. It’s just stuff that nobody’s seen and it’s not readily identified using a standard. That’s the stuff that people are looking to some of these new technologies, some of the optical recognition transferred into a vector drawing that then it is converted into a database that we can map to. That’s the Jetsons technology that we all envision that doesn’t quite work yet.
I had an idea just while we were talking here, Eric. This sounds like more work for me.
I need the light bulb to flash above your head.
This idea of standardization, what if there was a way to go bidirectionally with the scrubber to say, I have an ALOX4 customer and I’m going to submit my inventory to you and say, “FCH, please tell me how I should actually redo my descriptions to an industry standard better ingestible norm.”
What's more commodity than a piece of metal? Share on XYou’re right on. We do that in a piecemeal way now. It needs to be automated, it needs to be put into a much cleaner process. The idea that you’re talking about exists in nascent form already.
I’ll give you the benefit that have probably thought through, but essentially we are super close with that functionality is essentially to do complete data audits and data scoring across your catalog. That would be very helpful. There’s some challenges with that, but I think overall, though, company ABC wants to upload their catalog, and Eric could spit right back to them here’s your data quality, here’s your gaps, here’s your grade. Here’s actually an action plan for you to go up and clean that. Potentially, here’s even suggestions on how to clean it. There’s a lot of different things you could do there.
Maybe that’s the round three of the integrations that we build with ALOX. That’d be super interesting.
We’ve put that off to the side because we took a big bite out of it already with what we have on the table but it’s sitting there. It’s just sitting there. There’s a lot of different ways that you could take a conversation like this, obviously, we just said turn on the recorder and see what happens. Not a lot of laughs in this discussion so far.
Sam, so we were doing our first experiment with video. This is going back a couple of months, remember? Out of nowhere, I think we were on a Teams meeting and you made some gesture and an AI picked up on it, remember? It embellished what you were doing, or it put like a thought cloud over your head. Remember that?
There was something like a thumbs up. I think Teams does that. You can enable gestures or something like that.
Yeah, but this was just spontaneous. To Anders, when you had that idea a minute ago with the light bulb thing, I was surprised that it didn’t just show up. Given that we had a Mexican last night, I’m not sitting here with my own icons floating around right now. Stuff like that, anything’s possible these days, but so far, none of that’s appeared on this particular video chat.
I don’t know how that plays into fasteners other than maybe if we enable video chat in the sourcing piece of it, it’ll just eavesdrop into the conversation and just start flashing visuals of exactly what’s being discussed as a visual aid. “Yeah, that right there. What’s over your left shoulder? That’s what I need. Send me 10,000. I need them by Tuesday.”
We’d have a lot of beer getting sent over.
I think we all have high hopes for the future, and we had high hopes for this conversation. We’re going to reel it in here, I think, because we’ve got a lot to put on the virtual artificially intelligence enabled so-called cutting room. Probably some of it’ll make it through, but we’ll have to see what it is. This has been a good experiment so far. Thanks for doing it with me.
Thanks for having us. It’s been great.
Happy to be here, Eric. Thanks.
We’ll come up with a title and all that, or maybe we’ll just let the machine do it. All right. Sam Aldinger, Anders Green. Thanks for reading.
Important Links
![[GC2610] Solution_FCH Banner_[220x100] copy](https://news.fastenersclearinghouse.com/wp-content/uploads/2025/09/banner_solutionind.jpg)
