Master Data-Driven Decision-Making with Greg McHale of Datanomix, Ep #1

[0:00] Speaker 3: So Mike, all of our training, everything we’ve done has built up to this moment where you now have your own podcast. You’re about to release your first inaugural episode. Do you feel like you’re prepared?

[0:14] Speaker 2: I feel like you drugged me into this.

[0:17] Speaker 3: I drugged you into it, there was no drugs involved.

[0:21] Speaker 2: No, not drugs, drug as in drag.

[0:25] Speaker 3: I dragged you, all right, good. No, but really excited about it.

[0:29] Speaker 2: No, it’s exciting.

[0:30] Speaker 3: Buy the Numbers, the new podcast under the Making Chips umbrella here, and I think it’s only suitable to explain why we have a podcast named Buy the Numbers. What’s the point of it? What’s the reason for it?

So I don’t think we can explain that without telling the Mike Payne story a little bit. We don’t have to go back to the day you were born. Nobody needs the graphic details.

But Buy the Numbers, what does it mean? Why are we starting the show from your perspective?

[0:56] Speaker 2: So we spelled buy, B-U-Y, because so much of business is numbers and dollars, right?

[1:03] Speaker 3: Yeah, making purchasing decisions.

[1:05] Speaker 2: Making purchasing decisions, for sure, yeah. So whether that’s companies, capital equipment, people, all those types of things, numbers really just drive your business. And yeah, so maybe people that listen to Making Chips know my background’s really more on the business side, finance side of things, not so much in manufacturing.

If you look at this industry, most people grew up in the shop and then maybe have transitioned to shop ownership either through acquisition or buying their own or- Well, that’s my background. Yeah, or family business.

[1:38] Speaker 3: I remember my dad had me going and fabricating parts at Enid because I owed him money. Right. Only later did I realize, I should probably learn about this.

[1:45] Speaker 2: That’s the same story my son will have.

[1:47] Speaker 3: Yeah, exactly.

[1:48] Speaker 2: I came from the opposite side. I mean, I spent 15, 20 years doing private equity across multiple industries, including manufacturing. Started out my career in shop floor data collection.

So we had a software company that collected data off the shop floor and tried to drive that data to data-driven decisions. I really value data, I value numbers. I think you can use them to drive your business forward.

As I came into the industry and of course, getting to know you and recording with various people from around the country in the industry, there just seems to be a thirst for a little bit behind the numbers. What can I learn from my numbers?

[2:27] Speaker 3: I think it’s probably safe to say that the average person at, like we’re at Fabtech right now. You know, you might hear some noise in the background. It’s because we’re at a trade show amongst the community of manufacturing leaders.

And I think it’s pretty safe to say that most of the people here are probably better at making parts than they are like making high-level business decisions based on real data, understanding the value of their business and how to increase that value or the value of another business. And if they were to acquire it, what would be like the right price or maybe boosting their value so that they could sell their business and all the kind of stuff that you bring to the table. I can tell you personally, I’m really excited to learn about that.

It’s an area where I have a lot of room to grow. As my family business grows, we have opportunities to buy companies and stuff like that. So it’s like, I’m gonna listen to every episode of this.

And I don’t listen to every episode of Making Chips because I’m usually like one of the guys- You’re on it. But yeah, I’m excited about it, for sure.

[3:24] Speaker 2: There’s a ton of information we can dig into, right? So you’ve got, of course, finance and accounting type decisions. You’ve got lots of operational data.

You can even boil it down to, Matthew, Nick and I talked not too long ago, like Nick’s companies is the best places to work, right? That is all KPI driven, right? So what is your investment in training per person and all those types of metrics?

Well, all of those things can drive you to be a better company, achieve certain goals. And I think that’s a lot of what we’ll talk about is, I mean, we will talk about mergers, acquisitions. We will talk about capital investments and taxes and those types of things.

But we’ll also spend a lot of time talking about just operational data. Like how do you align your goals and what data should you be tracking to know if you’re on track to achieve those goals?

[4:12] Speaker 3: Sure. I’m excited about like the micro numbers, you know, like the spindle utilization on a particular machine tool, all the way up to the macro numbers. Like, you know, what even a multiple can you command based on the business you’re trying to buy or sell.

So really excited about it. But enough interviewing you. I feel like you’ve been on Making Chips 58 times.

I think what we need to do now is have you interview somebody else and kick off this new pod. So let’s buy the numbers.

[4:40] Speaker 2: My first guest is Greg McHale. He’s the CEO and founder of Datanomics. When you look at operational data and the things that Datanomics and Greg are doing in that space, they really can drive your company.

And we talk about aligning your goals and priorities with KPIs to measure that. Greg’s got a grasp for that unlike most people. He’s got a background in big data, enterprise data, and launched into this industry, I think in 2018, something like that.

And has really built a mission to provide shop owners in the manufacturing world with real data that can make an impact. So anyway, with that, here we go.

[5:19] Speaker 3: Let’s bring him in. I’m looking forward to hearing it.

[5:21] Speaker 2: Let’s do it.

[5:25] Speaker 4: Welcome to Buy the Numbers. Each week, we’ll explore the numbers that drive your business, from accounting and finance to operations and the exciting realm of mergers and acquisitions. Whether you’re a seasoned veteran or just starting out, our discussions will equip you with the insights and strategies you need to rein in the power of data to elevate your manufacturing company.

Let’s crunch some numbers.

[5:54] Speaker 2: So Greg, welcome. Thanks for having me, Mike. Yeah, I’ve talked to you before about launching this Buy the Numbers, and it’s going to be a lot of things about a lot of different types of numbers.

But one of the things I really think is important is data-driven decisions. And when I thought of who’s the first person I want to talk to about manufacturing data, it was my good friend, Greg. You’ve helped us a ton, and I just thought we should talk data.

[6:18] Speaker 1: I appreciate that. First of all, Mike, I love the name of the podcast. I love the Buy the Numbers, the B-U-Y, because I think that’s something that every manufacturer out there is trying to figure out how to do, is how do I increase the value of my business?

How do I increase the value of my shop? Which ultimately has to be done through the numbers. And look, we’re all in this because we’re capitalists, so.

[6:42] Speaker 2: Right, yeah, it’s not a non-profit work, is it? The B-U-Y just works. Yep, yep, yep.

Well, it’s been fun so far, used the opportunity while we’re here at IMTS to get in front of some bright minds such as yourself. And so based on that, first question, how’s IMTS going for you? Unbelievable.

[6:57] Speaker 1: We’ve been blown away so far. So my pattern for IMTS historically is, Monday, of course, is always great, but a lot of people fly in on Monday, right? That aren’t exhibitors.

Then Tuesday’s usually your peak, and Tuesday’s like just nuts. And then Wednesday’s a little less than Tuesday. Thursday’s a little less than Wednesday.

That’s not true this year. Really? So Monday was exactly what I expected Monday to be.

And then I told the team, and we had a lot of folks first time here, I’m like, Tuesday’s tsunami day, guys. Just like, if you gotta go to the bathroom, it’s in and out, right? We’re not reading our phones.

Right. Back to the booth, we got work to do, right? So Tuesday was as hot as expected.

Wednesday was hotter. Wow. Wednesday, we were like, what is going on?

Here we are Thursday, and it’s been hot up until I walked over here. That’s amazing. To do the show.

I would just have to guess, anecdotally, I mean, I know we’re doing by the numbers, but anecdotes matter too.

[7:54] Speaker 2: Yeah, sure.

[7:55] Speaker 1: This IMTS year has just had to be incredible, really, for everybody, based on what I’ve seen.

[7:59] Speaker 2: It’s interesting you say that, because about a month ago, we recorded a Making Chips episode with Peter Eelman, that’s the show boss. We were asking him about attendance and expectations and so forth, and he was talking about, they’ve made it so easy to register. They don’t know, really, until the show’s over, what they’re gonna expect to be there.

Who actually showed up. Because you can show up today, and you don’t have to preregister. He said, 15 years ago, you had to fax it in.

They knew who was showing and who wasn’t showing. But they had an expectation that this might be one of the busier years they’ve had of the last maybe 10 years or so, I think he had said.

[8:33] Speaker 1: I totally believe it.

[8:34] Speaker 2: And I think they were fully expecting to be well back up over 100,000 visitors. And I would just tell you, based on my experience, walking around the show, visiting booths, that has absolutely been the case. It’s swamped.

[8:44] Speaker 1: Oh yeah, I think I walked past 100,000 people trying to get to the talk I did the other day.

[8:50] Speaker 2: It’s been a great show for me, and I’ve enjoyed being able to reconnect with everybody and talk and meet people that you only get to talk to online. It’s been awesome. One-of-a-kind event.

Yeah. So tell us, we’ve talked a lot, but go ahead and tell the listeners a little bit about your background and Datanomics. Give them a little overview so we know why I got you in here.

[9:10] Speaker 1: Sure, yeah. So founder and CEO of Datanomics, and I started the company in 2018. We do production monitoring software, help manufacturers use real-time data off of machines, business systems data, like ERP, quoting software, financial data, et cetera, and provide all the automated insights and valuable feedback that help people improve efficiency, improve profitability through the use of operational data to help folks get better.

In terms of how I got into it, I actually don’t come from manufacturing at all. I come at this from the outside in, which I actually think has been a significant advantage just for us as a company to be unconstrained by sort of some of the assumptions of the past and to be able to push boundaries on different areas of the problem. So I come from enterprise data collection, enterprise infrastructure, storage appliances, analytics appliances, things of that nature, and really all that got commoditized by the cloud, and I wanted to start my own company, and I was like, there’s nothing to do in that space, and I am not going to work for Amazon.

So I really, outside in, top down, just started looking for new and emerging data types. What’s going on with machine data? What’s going on with sensor data?

What’s going on with this whole industrial internet of things topic, and do people have data? Are they using data? Is it really valuable?

And so I just started pulling on that thread. I started spending time with manufacturers, just through networks, started getting introduced to folks that own precision machining companies, go in, walk the floor, and I mean, I barely knew what a CNC machine was. I had never been in a machine shop, for sure.

I have no mechanical skills whatsoever, but I started meeting with these owners and operational leaders, and I listened to what they do, and how authentic they are, and how hard they work, and how important what they make is to literally the viability of our country and our economy, and I was like, what data do you have, right? And people show me stuff in the ERP, and I’m like, so what can you do with that, right? And it’s, well, it’s, you know, it’s daily.

[11:24] Speaker 2: It’s about an hour. And probably in a lot of cases, not even from an ERP, right?

[11:26] Speaker 1: Or by hand.

[11:28] Speaker 2: A bunch of sheets that have been collected and maybe put in an Excel spreadsheet.

[11:31] Speaker 1: Exactly, right. I walk up to the controller, I write this down. We go over to the whiteboard, and we do this.

We look at our ERP, and I look at that, and I’m like, well, okay, but what about the machine data side? So I’m like, well, I got to find people that are doing stuff with machine data. So then I found folks that were trying to do that, and I’m like, this cannot be state of the art, right?

What I saw was just like, this is not possible, because there was really this wide spectrum of on the one hand, you had a lot, and there are a lot of monitoring companies out there. We can come back to that later, but there’s a bunch of solutions that collect data from the machine. Is it on or off?

And what percentage of the time? And I’d look at people and say, well, does that change anything about your behavior? No, it just tells me how much I suck.

I’m like, okay.

[12:13] Speaker 2: It just makes me cry after the job’s over.

[12:16] Speaker 1: That’s an amazing, right? That’s an amazing feeling. I’m like, so, okay, is it that we need more data?

What’s missing? So I just branded those things too simple to be useful, right, not helpful. Then on the other side were these other solutions that were collecting everything under the sun and throwing it in your face and saying, man, if you had the time, if you could go crunch the data, if you had your own analysts, if you had your own this, if you had your own that, man, is there a treasure trove of insight in there?

And then I looked at these manufacturing people I was spending time with, and I’m like, they don’t have time to even think about that, let alone figure out how they’re actually gonna do it, who they would hire.

[12:53] Speaker 2: Best case, they’re dumping that in and trying to create a pivot table that tells them something, right?

[12:57] Speaker 1: Exactly, right. It’s just too much work. And it’s like, wow, we missed here.

And what we missed was the middle road was this has to be super simple, out of the box, make all the data just work, contextualize it, what part number am I making, what should my performance expectations be, when am I doing a good job, when am I doing a bad job, what do I need to do about that, how does that affect my profitability, how does that affect my revenue? And you have to do that all with taking the human out of the equation. I mean, that was one of the other things that I saw was the technology, when we started the company, every system depended on operator input.

Collect data from the machine, but we didn’t really do a good enough job at making the data make sense, so the human’s going to fix all that. And it’s like, there’s just no way the lowest common denominator being everyone, get everyone on your floor all day, every day to make the critical operational data make sense. I mean, we all have a hell of a time doing that with ERP.

Now, hey, let’s add another system that actually bugs them 100 times more than the ERP does, and they got to make that make sense. And so it was just so obvious that what was really missing to make this category viable was the appropriate technological solution that understood the way manufacturers really work, and I couldn’t help but work on that problem.

[14:20] Speaker 2: I always love these conversations with you because you know my background, some of the listeners know my background, but if you talk about it, when I was in the software business in the late 90s, I was doing shop floor data collection, and I’m going into CNC machines and mixers and all these things and pulling weights and part counts and all types of data and dumping it into the ERP system. Now, at the time, because there was a couple of factors going on. One, that was harder to do.

Maybe it wasn’t harder to do. I bet it was a lot harder to do. It was a custom solution that, I mean, my company might charge you three quarter million dollars to come in and do that, right?

So fast forward 20 years almost, I mean, 18 years, 15 years, whatever, and like you, I’ve entered this industry in 2018, and I know there’s data in that machine that I can do something with. And actually, when I came to IMTS in 2018, one of my primary goals was to find some way to get that data. And there was, like you said, there’s people on the market, and I looked at some of those, and there’s good and there’s bad.

And I mean, there’s a lot of solutions out there, and I think probably most of them all fit different niches, and that’s great. But that data is valuable, right? If you’re doing something with it.

It’s been an interesting journey for me just to see what was attainable but unattainable. You know it’s there, but I can’t get to it. Right, why is it so damn hard?

Right, and to see just from even like 20, like you’re talking about IMTS 2018 versus now, the things that we’re doing, whether it’s directly like what you’re doing or in AI operations of, I mean, just a million different paths here, where we have so much more access to all this stuff. But yeah, I think that next phase, and you’ve already touched on it, and you guys are taking a cut at this, and everybody’s taking their own approaches. But now what do I do with it?

That’s the part that we want to focus on. There’s a lot of value there if you can do it, and you do something with it. So talk to me a little bit about that.

I’m a big, big, big believer in data. I can overanalyze where I’m going to stop and get gas, right, so. Hopefully before you run out.

Right. Some of my guys at the shop, they laugh about how I can get wrapped up in overanalyzing some relatively minor decisions, but I can make big decisions quickly. And I think that’s a lot of times it’s because the data, right?

I mean, the data can prove itself on a big decision sometimes a lot quicker than a small decision.

[16:45] Speaker 1: Or it might just require less data to know that it’s the right thing to do.

[16:48] Speaker 2: Right. So let’s say I’m collecting information from my machines or I’m collecting information from my accounting package. I’ve got all these sources of data.

I mean, my entire shop now can give me data. Yes. What the hell do I do with it?

[17:03] Speaker 1: That’s a common question. This industry figuring out how to just live and breathe alongside data all day every day is really something that companies like Datanomics have to help master for the industry to achieve at the levels that we all want it to. One of the most important ways to frame it is we all have a business plan, right?

For your shop, you’ve got a revenue target. You’ve got a gross margin profitability target. You’ve got a set of assets and there’s some inputs and outputs that have to take place there for your business to hit its goals.

So we’ve got to start there, right? It’s not just, you know, hey, let’s go collect data and see if it fits like a nice new pair of sneakers and maybe it’ll work or maybe it won’t or something like that. So first we’ve got to be anchored in the goals of the business and knowing that, all right, I’m trying to generate this many dollars per hour revenue on my machines and my labor’s this and if I want to hit gross margin, well, I got to generate this much revenue per labor hour.

So we got to be thinking like that first and if you’re coming at it top down like that, then it’s getting the data that aligns with very simply your business plan to make sure you’re delivering on the right amount of uptime, the right level of machine performance, the right utilization of your capacity and your assets to deliver to that business plan because that now gets you thinking about really the right handful of things for how to run the business. Like number one, do I have capacity to sell the type of work that I need to be selling to hit my profitability metrics?

If I have more capacity than I realized, what types of things should I be out selling more of to try to get that work in to boost the uptime and utilization of my assets so I can be generating the incremental revenue that I need to hit my targets?

[19:02] Speaker 2: If you think about, and this could be many, many industries, you could say the same thing, but as we’re walking around an IMTS right now and we all live in this space, there’s two things people usually say they can do to increase capacity. They need more people and they need more machines. And those things help.

They do.

[19:18] Speaker 4: Absolutely.

[19:19] Speaker 2: I’ve got to have good people and I got to have good machines. For sure. I know you’ve heard me talk about before too though, like having the data to see maybe how much machine capacity am I really using and then being able to affect even just small decisions.

So in my case with roughly 40 machines, if I can save an hour a day, my maybe being more efficient or opening up some capacity here or there, I just hired a person. I just gained 40 hours. If I can gain, I mean, I have to gain it on every machine, but if I gain an hour a day per machine, I just hired a person without hiring a person.

I’m adding to my top line and adding to my bottom line. Right. So no incremental costs for a higher level of throughput.

There’s a ton of ways I can gain that hour. I can gain that through better people, better processes, better data to make decisions about those things. I think we often default to, and you hear it all over this show and you hear it from everybody.

In fact, I need more people. I need more people. I need more people.

And I think everybody recognizes there’s not more people, right? I mean- The pool is not there. Right.

I mean, there are, and we can go find them, but it’s a little bit of a zero-sum game. If I hire somebody, I just hired them from somebody else, which helps me. But if you branch into like the bigger aspect of where we’re talking about this gap in the market, we’re not growing the market.

We’re growing maybe our company by adding 10 people and that much capacity. But all we did is take it away from somewhere else and we didn’t fill this gap that I’ve talked about it a few times now. You’ve got buildsubmarines.com here at the show. They have a shortage of 6,000 people, right? So we didn’t add 6,000 people to the market. We just moved them from one place to another.

I think collecting data, analyzing data in a correct way, we can add 3,000 people to the market just because we’re more efficient.

[21:10] Speaker 1: For sure. And I think what’s important about making sure it’s connected to your business plan is if it’s not, what happens is the data just overwhelms and paralyzes you, right? So now there’s all these red lights.

I’m chasing every red light because I don’t know what the hell it means. Wow, that’s running at 20% utilization? Man, I thought that was a 80% utilization job, right?

Everything seems worthy of your attention because it’s not anchored in, well, what really is the business plan? So you have to believe, what am I trying to hit for revenue? What do I gotta produce for uptime hours to get there?

And start just really, really simple. And then any action you take in your mind is always connected to more revenue or more profit as opposed to just what can feel like random ping pong activity on the shop floor where you’re like, does anything work around here?

[22:01] Speaker 2: I know what that, I mean, yeah, I’ve got that red light, but I don’t know what it means.

[22:04] Speaker 1: Right, do I care or do I not care?

[22:06] Speaker 2: Yeah, it may be fine.

[22:07] Speaker 1: There’s a lot of stuff we can get pissed off about, but if it’s not anchored in, is this really affecting the performance of the business? There’s too much that it can overwhelm you.

[22:16] Speaker 2: So we’ve both been in the industry for a short time. I mean, relatively. What I’d like to hear your take, you’ve done such an amazing job.

I mean, watching you for, I don’t know, maybe three or four years now, watching what you’ve done. I mean, you have developed a great product that I think has a lot of value. I use it.

I’m very happy with it. But I also know the other thing you’ve built is a great community around you in this space, right? And that puts you across the table from a lot of manufacturing leaders.

So what I’d like to hear from you, like from your viewpoint is maybe two examples. One of where you see the average person listening to this, like where could they start using data? So let’s talk briefly about like, what’s a good example of data I can get out of a machine that’s actionable?

And then maybe what’s a good example I can get not from a machine, I mean, just other shop data that is actionable that I can use to either grow my top line or my bottom line.

[23:07] Speaker 1: I think the most important thing is getting baselines on what you really do. One of the first questions we always ask and Dave Kapkowicz from EBITDA and I did a session here at IMTS just the other day on quoting this way is killing your profits.

[23:20] Speaker 2: Sure, sure.

[23:21] Speaker 1: And a lot of good topics in there, but one of the ones we focused on is what’s everyone’s efficiency assumption when they quote? And I love this question because I’ve never been wrong about this. The minimum answer is always 80%.

Okay. 80%, 85%.

[23:37] Speaker 2: So when you say that, the typical person you’re talking to feels like they’re running their machines at an 80% efficiency rate. So it’s for every hour they’re getting whatever that would be, they’re losing 11 minutes or something.

[23:51] Speaker 1: Exactly, almost 50 minutes of uptime out of every 60 minutes. Sure. And you say, what do you do?

Oh, I got a five axis mill, 45 minutes cycle times, door opens five times during the parts. You think you’re getting 85%? Right, right, right.

And it’s not because anyone’s doing anything wrong. It’s because the data hasn’t been there to tell us what reality is, right? So everyone always says 80, 85%.

And then people say, well, what if I actually don’t know? If you don’t know, that means your answer is 100%, right? You say, oh, that’s a 10 minute part.

Then you think there’s no other time. Every 10 minutes you’re getting a part.

[24:24] Speaker 4: Every 10 minutes you’re getting a part, right?

[24:25] Speaker 1: And getting the baseline on how you really run puts you in the position to think very differently about the business and how long do things really take, both in terms of what you’re selling your machine time for and the labor hours required to produce parts of that particular value, just to drive it home. If I’m really only getting 30 to 50% out of the asset, which for non-palatized, non-automated equipment, it’s not bar fed, it’s not palatized, it’s not robotically fed. 30 to 50 is just as a good swag about where everybody is.

I’ve looked at data across thousands of machines. And so everyone always asks me like, oh, what’s a good expectation? Well, it depends on the equipment type, but let’s just say 30 to 50, right?

So if your business plan is 100%, I’m going to get a part every 10 minutes. And then on the floor, you get a part every 20 minutes. Either you’re pushing on extra shifts to hit your delivery date, which means you’re paying twice as much as you thought you were to hit the date, or you’re delivering in twice the time you thought you would, which greatly reduces the revenue the machine actually generates, because that’s in service of one job.

[25:38] Speaker 2: Your effective shop rate is actually 50%. It’s half, right, it’s half.

[25:42] Speaker 1: And then if your target gross margin is 40, 50%, well, you’ve basically given away all the wiggle room for things that happen upstream of that and downstream of that. And a lot of folks, Dave says this all the time, a lot of folks, they get to the end of the year, they’re like, man, I was targeting 40% gross margins. I actually did nine, 12, 15.

Well, this is where that is. And it’s not like you can’t have a great lifestyle business at 15% gross margin, but if you’re thinking in this by the numbers type mentality, like you’re trying to have us do here, well, changing those kinds of assumptions, getting the real baselines, and then saying, okay, now how do I get the business producing at the required level of throughput, required level of uptime, required level of cycle time to make sure that I do hit my gross margin targets is where a lot of the insight comes from. And there’s different ways for how to do that.

But that’s the big picture that really we need the first recalibration of.

[26:43] Speaker 2: Right, that’s such a wonderful example. And as I talked about on some other episodes, and I think will probably be talked about a lot, is when we talk about maybe you’re potentially positioning your company to sell, and you got to take that multiple into effect, that $100,000 decision is maybe really a half a million dollar decision, or that not decision, but that change in number. It’s $100,000 to your bottom line, but it’s $500,000 to the value of your company.

They’re even more important if you view it that way.

[27:11] Speaker 1: Sure, yeah. I mean, if you’re making decisions that are fundamentally going to affect your cost of goods sold, those have to return 5X, right? Whatever those decisions are.

[27:19] Speaker 2: I’m actually going to answer the question I asked you about maybe what’s some other data that’s important, and you kind of triggered it when you talked about it. These things that happen upstream and downstream from really what’s going on in the machine, and often that’s all we pay attention to. We’re like, okay, cycle times, like when we’re quoting a job, it’s like, it’s going to take me two hours to set this up, and it’s a 12 minute cycle time, and there’s two more ops, and whatever those numbers are, and we run off of that.

That’s our price. And that’s a really good place to start. It’s critical.

But one of the first things I saw early on in our quoting process was there’s going to be scrap. The best performing shops in the world are still going to run like a 1% scrap rate if you’re running perfect. So you’re going to lose one out of a hundred parts on average.

If you don’t account for that, it could wipe out the entire profitability of a job. Now, that created a different problem. Well, how do I know what my scrap rate is, right?

Because prior to me buying the company and implementing an ERP system, that data was collected on, if you make a bad part, you’re supposed to fill out this form and turn it in. Well, how often do you think that happened? It didn’t happen very often.

Only if other people knew you had a scrap, right? I actually remember when we first kind of went through and did a cleaning of the shop, we’d move a work bench, and there’s rusty parts underneath it. Well, those were scraps that nobody told anybody about.

[28:38] Speaker 1: I have a customer, Mike, where they mowed the field behind the building one time and found a pile of scrap parts. You gotta be kidding. It was like Shawshank Redemption when they put the little rocks, the pebbles in their pants and shake them out.

It’s like you throw a couple of scrap parts in your pant leg and shake them out in the yard.

[28:58] Speaker 2: So the first step I had to do to that, though, is collect good data on my scrap. Before I could figure out what my scrap rate is, I needed to get good data collection of my scrap rate. Now, we were fortunate.

Our ERP system we deployed has a great quality module, and we had to really change the culture of what a scrap meant. We’ve adopted an approach of basically, if it comes off with any defect, whether it can be reworked or not reworked or used as is, doesn’t really matter. It’s a scrap, right?

So I need to know that. I need to know it for a variety of reasons. One, I need to know so that maybe we can improve how we run that part in the future.

Maybe I learned that I need to change a drill out every 12 parts, that type of stuff. But also I need to know, so on these parts, maybe my scrap rate’s a half a percent, right? I can run 200 parts without having a bad one.

We quickly, as we improved that, I mean, we found jobs that our scrap rate was 30%, but we weren’t priced for that. We had to take that into account. We gotta buy 30% more material.

And it’s not because we were not smart in how to run parts, it’s just a difficult part that’s gonna happen. And then we had to go back and say, okay, we’re gonna have an increased runtime and material in this part. Now this is what that part needs to be priced at to cover that.

Sometimes that’s a difficult conversation with the customer. We honestly found once we had the data to show them, they were fine with it. If I could show them, like, look, this is the last three times we ran it.

Here’s these scraps we had. They’re unavoidable because of the type of material we’re using, it’s gonna egg. Again, it’s not that we were doing it wrong, it’s maybe we’re using a casting and there’s gonna be porosity in it.

And then sometimes we were doing something wrong, right? And those are the chances for us to improve. But the data collection of that scrap was required to figure that out.

So now we have shop-wide goals on scrap and we have part goals, right? So we’re okay losing one of these every hundred because we know it’s gonna happen and we’ve built it into the pricing. I think that’s another example.

I’m not getting that necessarily off a machine, I’m not getting that from me. I’m collecting that data in a lot of cases. I mean, yeah, I’m relying on human input there into my ERP system, but then I’m getting information that I can make better decisions of how to price my part, how to schedule my part, right?

If I don’t realize I’m gonna have a 30% scrap and I’m making a hundred of these, I’m gonna have to make 30 more. You got to over-provision, right? And that might be an hour apart.

[31:30] Speaker 1: That’s 30 hours, I’m gonna be a week late. That’s right, that’s right. I mean, yeah, what you’re hitting on is basically you got to baseline every element that goes into your cost, right?

So you hit setup time, you hit cycle time, you hit scrap rate, right? And then of course the tooling and the other stuff, but really if we don’t have the data on what’s the real utilization of machine, what’s the real setup time, what’s the real cycle time and what’s the real scrap rate, how do you quote a job and guarantee that the business is gonna hit the target profit margin? You don’t.

You can’t. It’s impossible. You can’t.

Right, so it’s exactly what we’re talking about. You need the baseline because with the baseline, you can see reality. With reality, you can plan the job and plan the business around what really happens out there.

And look, when you get the data, you’re also gonna learn there’s a whole bunch of stuff you’ve got to optimize. It’s not like you can just take your current inefficiencies and inefficiencies, again, through no fault of your own. If you don’t have visibility into these metrics, you’re as inefficient or as efficient.

[32:37] Speaker 2: You don’t even know how efficient you are.

[32:39] Speaker 1: Yeah, you’re as efficient as you can be with no information.

[32:41] Speaker 4: Right.

[32:42] Speaker 1: I mean, great. You might be superhuman. Right.

You might be able to whip around your 40 plus machines in 10 seconds and figure out everything that’s going on.

[32:50] Speaker 2: Right, right, right.

[32:50] Speaker 1: People can’t do that, right? So you’re as efficient as you can be given the constraint of not having visibility, but once you get the visibility, it’s not gonna mean, oh, let me redo my business plan on the assumption that everything takes twice as long as I thought it did, right? You can’t just double your price.

[33:05] Speaker 2: Now you just lowered your top line.

[33:07] Speaker 1: Right, you can’t just double your price. You’re not gonna get the work. But what it’s doing is positioning you for the hard reality that you both need to get more efficient and need to plan around the actuals, and you’ve gotta squeeze those two things together as closely as possible to get yourself to be a business that can run the best possible way while achieving the goals that you have for revenue and profit.

[33:31] Speaker 2: That’s awesome. I love that summary of that. As we start to kind of wrap up here, we’ve talked a little bit about shop floor data collection, even as long as 20 years ago.

We’ve both talked about how much has changed in the last six years, in our, you know, six, seven years in our experience in the industry. You’re much more in tune with it. I’d love to hear what your take is on what we’re gonna see at the next IMTS and the next, you know, what’s that next five to 10 years look like of data collection as we deploy AI and all this automation and all this new equipment maybe coming into your shop, right, like where do you see us going over the next five to 10 years?

[34:04] Speaker 1: For sure, it’s funny. We’ve actually been talking about this in our booth because IMTS 2022, you couldn’t go 10 feet without seeing a UR in a booth, right? I mean, literally, it’s like- It was, and they weren’t practical uses in every booth.

No, I mean, they were cool. Like, hey, open my beer, you know, scratch my back, whatever was going on, right? I think I saw one doing massages.

You literally couldn’t go 10 feet without seeing a UR. And this year, I don’t know if I’ve seen five in the South building. The machining side, it’s a different story, but where we are on the East side, there were URs everywhere last IMTS, and I don’t think I’ve even walked past one over on our side where we are.

So obviously robotics was the theme of 22, right? And that was sort of in the peak hype cycle. It’s like automate, automate, automate, right?

We all need robots. And then when the dust settles, what you ultimately learn is there’s a subset of scenarios for which that applies. They’re very valuable.

They’re well worth investing in, but it’s not the, it’s going to cover 800% of all the things we do. And you know, you didn’t go out and buy 40 robots.

[35:10] Speaker 2: Yeah, I think the way I would reflect on the same thing, I actually had this conversation with one of my shop managers this morning. He was here last time. He didn’t come this time, but he asked about, is it still just robots everywhere?

And I actually have seen a lot of robots. I agree there’s not as many, but I told him the biggest difference I’ve noticed is, I mean, last time, yeah, they were opening my beer and they’re doing whatever, but this time there’s actually practical uses of the robots.

[35:34] Speaker 1: They’re reduced to the practical applications, exactly.

[35:36] Speaker 2: And they also aren’t drawing as much attention because I think a lot of people that, we’ve all accepted that we need automation. There’s value there. We’re doing it.

And it’s not as, oh my God, look, there’s one of those robots, right? Oh, that’s an interesting use of a robot. I hadn’t thought about that.

So it’s a much more practical, you’re right. Before it was just to draw attention to your booth. And now it’s, hey, here’s a practical application of automation.

[36:01] Speaker 1: That’s the perfect way to frame it, Mike, because in 2022, it was like, wow, I can do anything with this. And then the dust settles. And now this year it’s, no, this is what you actually can do with it, right?

And so if you stick with that, I think this year, the theme is, hey, there’s AI here, there’s AI there. And it’s literally parallel to what we’re talking about with the robots, right? So this year with AI, it’s, oh, what can I do with it?

And it’s sort of this open-ended question. And I think what you’re gonna see by IMTS 2026, the same idea, there’ll be less, everyone’s parroting AI, AI, AI, and it’s gonna be, these are very practical applications because we’re gonna have to reduce to the things you actually can do that as like a software vendor, I can actually sell a practical application of AI to solve a specific problem. And as a manufacturer, you’re willing to buy for that practical application, just like the robots.

And I think there’s fundamentally two things that AI needs to help the industry with. And this is certainly what we’re focused on in our company. And number one is a lot of great shops have historically relied on tribal knowledge.

Every shop has a person you can name, I’ll tell a story. You can name the person who he knows this, he knows how to do that, he walks by, he hears this sound, he does this thing, right? And what that really is, is accumulated tribal knowledge specific to your specialties, your capabilities, your culture, your operation as a shop.

So one of the things that AI needs to deliver for us in the next IMTS cycle, I like framing this in IMTS cycles. I didn’t do that intentionally because we’re at IMTS.

[37:43] Speaker 2: It actually works. It’s our development cycle in this industry, right?

[37:47] Speaker 1: It’s our reset as an industry to say, so where are we really at with all these different technologies? So one of the things AI needs to deliver in the next IMTS cycle is how is it automating the generation of the tribal knowledge that’s needed so that information gets passed along without needing that named person in every shop all over the country who just knows, oh, you’re on that lathe? Yeah, no, what you need to do is this.

Oh, you’re running that job? Let me tell you. In the setup, you’re gonna run into this.

Here’s what you need to do, right? So the automation of tribal knowledge is one major component of it. The automation of tribal knowledge also does something very important, which is reduce the amount of judgment required to make decisions, right?

That’s what that really gets to is when people go to Joe in the shop, what they’re really saying is, Joe, what decision should I make in this circumstance? Because I don’t have the experience, I don’t have the insight, and I don’t have the history to know when I run this part, when I do this thing, when I use this tool, what’s going to happen to me, and when that thing that I don’t know is gonna happen to me happens, what do I do about it, right? So tribal knowledge translates to reducing the amount of decision-making required on ultimately our operators, the folks on the floor.

That’s just one topic AI needs to deliver on.

[39:10] Speaker 2: So I need to chip in John? Yeah, that’s right. Just download the brain?

That’s right, that’s right.

[39:15] Speaker 1: That sounds easy. Yeah, just stick it in there, record everything that he knows, and get it programmed up. And then I think the second thing that AI needs to deliver is interpretation of cross-sections of data, right?

[39:31] Speaker 2: Massive amounts of data.

[39:32] Speaker 1: I gotta look at your finances, I gotta look at your ERP, I gotta look at your machine data, I gotta look at your quotes, I gotta look at your deliveries, and I’ve basically gotta lay out for you what are the things that are gonna hurt your revenue, hurt your profitability, hurt your business goals, what’s helping, and then very specifically where in that stack of information is the thing you need to go do now, or soon, or tomorrow, or whatever, to make sure that you’re staying on plan.

[40:02] Speaker 2: Yeah, where am I varying that’s urgent? Exactly. Where’s something I need to be thinking about this month, this quarter, or next year?

Right, right, there’s bigger things, there’s smaller things.

[40:12] Speaker 1: We’ve got a problem, but it’s not urgent, or we’ve got a problem, and you better go take care of it right now. Right, the phrase we use in our company is, people come to me with a problem, I say, is it a lion or a dandelion? Because I can ignore a dandelion almost forever, but if it’s a lion, I gotta go get a baseball bat and make some precautions here.

[40:28] Speaker 2: Yeah, good luck.

[40:29] Speaker 1: I wanna see you fight a lion with a baseball bat. I’m betting on the lion, I’m sorry. Me too, the baseball bat’s just something for him to chew before he gets to me.

It’s a toothpick when he’s done with you. It’s really those two categories. We gotta automate tribal knowledge, so that we reduce the amount of complex decision-making required for the folks in the industry, and then we have to interpret bigger sets of information to prioritize for folks like yourself, what decisions do I need to actually be thinking about and making, and how do I make those decisions and what are the direct implications to the business?

[41:04] Speaker 2: I think that’s great. I think I actually kind of parallel, kind of this week, I’m paralleling the AI to automation too, right? So automation, I think any form of automation is obviously evolving, but it’s a clear use in dirty, dumb, and dangerous applications.

Right, exactly right. I mean, it’s, that’s where- It’s spray and pray right now. Right, yeah, and if it’s dirty, dumb, and dangerous, let’s automate it, right, and make it less dangerous or take the dumb stuff off of your most skilled person and let them go focus on things that really use their skillset.

I kind of feel like that’s where we’re going with AI. I can break this data down and I can make decisions. If AI could make the dumb decisions for me, like if it could save me an hour and just present to me the nugget of that, it could absolutely take care of the dandelions.

Right, that’s right, absolutely. And let me focus on the lions. That’s right.

And actually, I’ve been impressed this week. I expected there to be a lot of marketing AI and not a lot of practical AI. And I actually have found a lot of practical AI this week.

Yeah, for sure. And a lot of it’s developing and actually the vendors are even saying that, right?

[42:17] Speaker 1: Yeah, that’s right.

[42:17] Speaker 2: They’re not saying it’s not perfect. Not over their skis. Right, yeah.

But I’ve seen some super practical applications of AI that have me pretty excited. Absolutely. We joked Monday on the stage, our talk was actually called Practical Uses of AI.

And I hadn’t even walked around the show yet, so I didn’t know what I was gonna see when I walked around, but we were joking. And this actually comes from somebody else recording a month ago, as we were previewing IMTS. But we were joking about if you walked around this show this year and played the AI drinking game, that you gotta take a drink every time somebody- I’m not gonna make it back to my booth.

You’re not gonna walk around very long.

[42:54] Speaker 1: And I think that’s- I’m gonna fall into a machine.

[42:55] Speaker 2: Yeah, I think that’s a true statement. I have been pleasantly surprised how much practical AI I’ve seen. And I agree with you.

I think over the next five to 10 years, I think the more and more practical uses of- We have massive amounts of data. I mean, you could overwhelm yourself with it. And just, if you can break it down from all these different systems, the merging of pulling it from a machine and pulling it from my accounting and pulling it from my quoting and really help me identify, here’s some real problem areas that you can affect.

Man, I mean, you’re gonna change my world and you’re gonna, again, back to that, if I can increase efficiency, whether it’s on the machine or in the front office, I’m closing that gap, right? I mean, if I can make somebody 20%, not even a machine 20% more productive, but my accountant 20% more productive, then I’m closing that gap we have in this workforce to really continue to grow this demand that we’re seeing come that is a problem for the United States right now.

[43:57] Speaker 1: Absolutely. I mean, and I love the way you said the dumb, dirty, and dangerous, because a lot of those things require a fair amount of brainpower sometimes or a fair amount of labor to address. And to the extent that those can be reduced and automated, that’s all incremental gross margin, right?

So if we’re businesses that are aiming for 50% today and we get this right, why shouldn’t manufacturing companies be aiming for 60, 65, 70% gross margins through the use of technology in the next five to 10 years? I think that would be amazing.

[44:29] Speaker 2: I hope that’s where we’re heading. Like I said, I’ve been pleasantly surprised at uses I’ve seen. I hope that’s the direction we’re heading.

And I think it is. We’ve got a lot of smart guys like you working on that. Man, Greg, I just appreciate you being on the show.

This is great. I’m so glad that you’re one of the first guys to join me and I appreciate it.

[44:46] Speaker 1: And thank you. This has been a blast and no better environment to do this in than live at IMTS. So thanks for having me, Mike.

Yeah, absolutely.