Doug Glenn with Heat Treat Radio interviewed Jeff Halonene, CEO of Steelhead Technologies to learn more about process improvement with Job Shop Software.
To watch the full episode, visit Heat Treat Today.
The following is a transcript of the conversation:
Doug Glenn: We’re looking forward to talking to you today about what is a growing interest, I think, to a lot of people in this industry, primarily because of the growing of technology and data and things of that sort. I know your company is heavily invested in job shop, plant-wide, organization-wide operating systems. That’s basically what we’re going to talk about today.
The first question I wanted to ask you is, what do you think are, for job shop-type businesses, the two or three main reasons that make them less profitable than they might otherwise be?
Jeff Halonen: Great question. A job shop is a unique business. They don’t have the luxury of centralized planning, or planning in advance and architecting a plant, like they might do if they are making Chevy Tahoes where an immense amount of engineering goes into build a highly efficient operation so it’s highly predictable.
A job shop, on the other hand, is the opposite of predictability. Down to what parts, what hour, what time of day, how many customers, by week, by month. . . everything changes constantly. That buys a lot of inefficiency, not because what they do is overly complicated; it’s the complexity that grows both with a high quantity of customers and a high quantity of specifications they need to build to. Then, obviously, being at the end of the end-customer’s production cycle, the lead-times are often very high pressure because you have to make up time for all of previous mistakes and missed deadlines. You’re the last step before it gets on a plane or a train or goes straight to the customer.
We’ve seen room to improve these operations from a profitability and a data standpoint. Data is not something you’re going to be able to sell out in the market. In your plant itself, it is not very valuable, but it’s the decisions you can make with that data. Specifically, this is the case when it comes to profitability. For example, if you have 5 - 20 different processes or lines, which one is more profitable than the other? That should be something that you understand. This is what your business does: different types of business, even different part numbers, different customers, different lead times that you offer for customers, what’s your cost versus return on that.
The other opportunity, aside from analyzing the commercial side of all of your work, is your production. If you’re a job shop, generally you are parts in and parts out, go, go, go. It’s more of a service business than a slow moving, very technical manufacture.
So, basically, every time a job comes through your shop, you should be studying — what does it take to make that job go? Anything from the quote, to closing the sale, to receiving the order, to the physical paperwork, to notifying the floor, to building a schedule, to scheduling capacity, to executing production, providing instruction information, how you collect that data, even something as simple as how do you notify customers. Are you notifying customers on the phone or an email? You do twenty to thirty orders a day, 100 orders a day — it adds up quickly.
Job Shop Invoicing — how do you generate your reports?
Aggregate Data — that’s the other thing that we see in job shops very frequently. The actual cost of executing an order on a marginal basis can be extremely high if you have many systems; or it’s paper-based; or it’s not streamlined.
Doug Glenn: Not just the processing of the part, in the sense of thermally processing the part, but you’re talking all the way from preparing the quote all the way through to shipping and anything at the end.
Jeff Halonen: Absolutely, Doug.
In our experience, the job shop manufacturers are pros at what they do — the actual process. They’ve been doing it for years or sometimes decades. They have it absolutely dialed.
In the physical processing of the part, sometimes there is room for improvement; but a lot of times that is a fine-tuned machine. Generally, it’s the business systems that surround it.
One thing that we like to point to is the very heavy investments that the industry makes in physical, tangible goods: natural gas, land, equipment, people, automation systems (physical automation systems). Where we see a huge opportunity is all the things you can’t actually touch and feel. It’s the digital operating system that your plant runs on, all the systems and processes that makes your business tick.
Doug Glenn: You’re talking about data in, from the very beginning of the process, all the way through. I don’t know too many companies, especially job shops, that are actually gathering that type of data. Is it possible? What are the costs involved, let’s say, of starting to gather that data from that point all the way through? It seems like that would be expensive and difficult to do.
Jeff Halonen: Absolutely, Doug. It feels like an asymptotic target. You can try really hard, and you can make progress; but that rate at which you make progress drops off really quickly. It’s what we see as a balance of cost of data and value of the data.
It’s not that manufacturers don’t value the data; they often do make an investments to get the data. They realize quickly that, with help with the proper tools, cost ramps up really quick. They essentially hit that floor or that asymptote, if you will, really quickly. They say, “OK, we are clearly extending more effort than it’s worth; so we’re going to dial it back. We’re going to go right here. That’s where we’re setting the knob, and we’re going to run our plant.”
Unfortunately, with the tools available, that setting or their status quo tends to be paper, Excel, systems like that. If you have no system at all, not even paper, it’s very obvious that we need something. We write on paper, we highlight, we put it here, we do this, we do that, we have a whiteboard for scheduling, and you see a lot of yield from essentially those data systems.
They are rudimentary data systems, and the cost is very low. You spend a week or two, and you deploy it. You make it, run it, and get the value. Then you say, “Now, I want to track what part was in what oven at what exact time and exactly what went through. Now I want to see all the parts that went through this or that type of process. We’ve run that part number forty times — I want to see its performance every single time from a commercial standpoint. All of a sudden, the cost of obtaining that data — now you just have a wish list. You look out and say, “That’s going to be impossible. That’s way too much time to get that data.”
That’s really where deploying a centralized platform that takes in everything from the business — anything from inventory, to quoting, to operator instructions, to timing each process, to part numbers, to shippers, to invoicing — it all really is intertwined. When you can track every single movement of every single part going through the entire process, the cost of that data drops precipitously. Especially if you have reporting integrated. Now you can generate reports that will passively collect data as your operations go on. Whatever you look at on a daily, weekly, monthly basis to essentially run your plant.
We like to say, “Digitize, automate, optimize.” To us, these are three core steps to reducing the cost of data.
First, you need to digitize.
If you have no system, or you have paper, it’s very difficult to obtain actual data and actual insights from your plant. Even if you do digitize successfully, but you just have a mass of huge databases, that’s very low usability because now you need to be essentially a data scientist to go in and make it useful.
Next is automation.
We need to automatically tailor the data to present profit margin. We need to automatically tailor the data to track reworks. That’s the next step because that gets us to actionable data information.
Then, the last is optimization.
The optimization itself often actually happens at the plant level where the management team is now in, what I would say is, a luxurious position of having a lot of their systems digitized then automated as well.
They have the data. They have the time to address it. Now they’re in the position of optimizing their facility. That could mean focusing on marketing; that could mean pursuing a new industry — developing a new process, training your operators. It could mean anything in the world — getting a new permit from the city, for example. These are things that optimize your plant to provide as much value as possible to the world. You have the ability to do that because you’ve done your homework to get your plant into a position to action that data and have the time available as well.
Doug Glenn: Right. You’re collecting the data that you need to make those types of decisions. I find it interesting some of the scope of the types of decisions you just mentioned, that job shops could make based on the data. That’s pretty interesting.
Let me ask you this, Jeff: do you have any examples in the heat treat world? Or where, for example, in the heat treat world do you envision there being some real business efficiencies here? Let me preface this one other way, too: a lot of the people that listen to this podcast are not necessarily job shops; they are what we call “captive heat treaters” who run high volume/low variability. But there are some who actually run high variability — a lot of different parts, maybe not exceptionally high volumes, or at least smaller volumes than our captive heat treaters. I guess, addressing those guys, or even the guys that run low variability – what are the business efficiencies we could see?
Jeff Halonen:
It’s going to be different from plant to plant as far as the value you’re going to have because everyone has different proficiencies and profiles like you just mentioned.
If you’re in a situation where you have a very high degree of mix, one of the huge inefficiencies that we’ve seen is they want to keep a really tight lead time; but they don’t have the ability to quickly and easily schedule efficiently. They keep all of their equipment at maximum capacity, so they keep as much throughput in the plant as possible. The inefficiency essentially shows itself in the quantity of ovens – they just simply invested in more equipment and just keep more ovens hot. They essentially overcapacitize which is both acap efficiency and an opex efficiency because you have to heat, maintain and run that much more equipment. That’s an environment where you have a lot of complexity, and the scheduling required or the inability to schedule efficiently is leading to capital inefficiency.
If you’re in a position where you’re on the other end of that spectrum, as you mentioned, Doug, where you have lower variability by part number higher volume; some of the inefficiencies there can definitely be more on the data systems — the data processes. Again, scheduling — if it’s a very simple schedule, you can schedule it very quickly and easily and very efficiently because you don’t have a lot of change. It is tracking that information — where are all the parts, for example.
A lot of times it’s the administrative time. People are walking out, still trying to find where the parts are. Simple things like order status or we ran this part with this lot number or purchase order or shipper number or whatever identifier you have — we ran this six months ago, what happened to that part? Often times, something as simple as that can turn into a 20, 30, 40-minute exercise instead of just quickly finding that information.
Doug Glenn : How about equipment inefficiencies? How can a system like yours that’s gathering all the data – how can it help us increase efficiency, just of the equipment that we have?
Jeff Halonen: The first step, as they say, to addressing an issue is knowing you have a problem. The first step is measuring what actually happens in the real world. You can walk out into any shop and just open your eyes and things are happening, right? But your knowledge of what’s happening is almost just instantaneous. As soon as you go to the ball game later, you’ve forgotten all about that pile of parts or that empty oven that you saw. If you have a situation where at the end of the day, week, month, quarter, you have the raw numbers – we know it can do X amount of parts or racks or pounds or whatever the metric is – and it did Y and that’s X percentage below that number, the numbers speak for themselves.
Jeff Halonen and Doug Glenn
Source: Heat Treat Today
As far as what the team can do to address the problem, that could be any number of things from a plant management standpoint. That’s where it really gets into the optimization side. Although one thing might be scheduling efficiency. If we find that turnovers or essentially thermal profiles or run speeds or whatever it may be, if we have an opportunity to batch production, increase density which might be a scheduling exercise, we might explore that.
That could also come back into business strategy — it’s all very related. For example, you might offer to expedite fees — fast turnaround at a higher price — but then lower price for long turnaround time because that now gives you the privilege of organizing your work in a more efficient way and providing the same value at a lower price. If that end customer has the time in their schedule, they can say — hey, I want to save money and have my job done in a week or so instead of 24 hours where I’m ready to pay a huge premium. That cost multiple can be dramatic and for good reason.
Doug Glenn: I have two more questions for you. One is a little bit hypothetical. I want you to kind of take a guess about the rate at which software and digitization, of the sort that you’re talking about, is coming into job shops. That’s the first thing. The second and last thing I want to ask is for you to just kind of take us on a walk through, to the best that you know, a heat treat business. How might a system like yours look?
A lot of people say, “Listen, I’ve got my system in hand, I’m okay the way I am.” First, is it coming? Do you think people are going to have to get to this digitalization sooner or later?
Jeff Halonen: I believe so, yes. I understand you can definitely run a shop on paper. You can run it on Excel. As far as macrotrends out in the world, consider first the technology available. The team at Steelhead is obviously very competent with a lot of really great software engineers, but the tools available have progressed dramatically over the last couple of years. We’ve run into heat treating shops that are using a system they may have installed in the ‘90s. It does functionally work, but we do see a lot of symptoms, essentially adverse symptoms, to the business by choosing to use a less effective system. That usually comes in less actionable data. It does facilitate the physical paperwork needed to run your plant, but it’s not a competitive advantage. It just is kind of there, and it’s just kind of barely acceptable. The employee morale is not great – people don’t love using systems where they have to reenter data. They can’t work from home. There are so many different currents that are coming in there.
The other one is the end customers often like more and more visibility and more and more traceability into plants. Transparency, visibility, traceability. I don’t think it’s a situation where it’s going to become unviable to run a shop without really advanced technology. I do think, over time, it will be a situation where the competitive profile is one where there are clear advantages to the shops that embrace the wholistic business advantages that come with that.
Think about your customers — they’re buyers just like anyone else. Everyone loves shopping on Amazon or their favorite website. They get the notifications; they can track everything. You order a pizza at Dominos, and you can see them rolling out the dough and putting on the sauce. The predictability and the visibility is something that is really important.
I think, over time, it is certainly something that the end manufacturer is really going to continue to drive (even automotive, traceability and everything like that). As the requirements of data go up, and the competitive landscape starts to adopt more and more of this technology, it’s going to be more and more of a disadvantage in the competitive landscape to be not at the cutting edge of that technology.
Doug Glenn: Pretty soon it’s not going to be cutting edge, is my thinking. As more and more people are adopting it, it’s going to become easier to adopt. I’m assuming it’s going to be easier to install, easier to launch into any business.
Take us through that, as a job shop. Any of either our captive heat treaters that have a high variability of products or a lot of our commercial heat treat shops which have high variability of products — how would it look? Start us as early as you can in the process and walk us through it.
Jeff Halonen: Starting at the beginning of the process, if you say, “I’m interested. I need a system. I can just feel it. I can see it. I know there is opportunity to improve our business.” Or, “We feel like we really optimized everything about our business, but we still feel like we have restrictions on growth.” This could be — it takes a lot of energy from the management team, the ownership team that are involved. One question I like to ask is: if the business going through your business or business unit doubles, do your headaches double? Do the amount of hours you work double? To me, that’s symptomatic of room for system improvement.
Early on in the process it starts with some level of interest, some level of intrigue. I think this is something we can improve on; our plant manager is complaining that they can’t work from home. Whatever all these issues are that manifest themselves; or you’re losing jobs in quoting, and you’re not sure if you’re a high price or a low price or margins might be unclear; you’re not satisfied with the level of data. Basically, that’s generally where we start our conversations with plants.
We really take an approach of both informing about what’s possible on the tool and what’s impossible on a modern platform. Then we also do a value assessment. How do you do this? How do you do that? That kind of third party provides a great, almost, sounding board to say, “Okay, here’s an itemized list of room for improvement.”
Honestly, that’s what really makes it compelling. When we can sit down with a potential partner and identify how we can actually improve the business in very real and definitive terms, not just insights that you can use to drive your business. Also here’s how much time you spent here, how much time you spent there. It’s a combination of calculating hard costs and hard opportunities and then the upside with the actionable insights from the system.
From there are the deployments at Steelhead.
We do direct-to-direct service, so no third parties. A Steelhead team member will essentially go through a lot of meetings and plant walks, in some cases, and essentially reverse engineer your plant.
Where do parts come in the building, and then what? Then they go here? Do they always go there? Sometimes. Can they be run together? We will ask a million questions to essentially reverse engineer the flow of your plant, and then we design the system to reflect your current operations. Not – here’s a software, now bend your plant over backwards to match the software. Whether it’s quoting, workflow automation, or reporting or any of the other parts of the system, we’re going to want to ask as many questions as possible to configure the system to match your plant.
Job shop technology will continue to improve
Doug, you mentioned the cost is going to go down, over time, of technology. Absolutely. Because one thing we can do very quickly since our platform is essentially kind of like a build-your-own; it’s almost like Excel except it’s Excel that is highly, highly engineered to work for this specific industry. One thing we can do is we can prove it out really quickly. So, in just a few hours or a couple of days, we can essentially build your plant, or a part of your plant, and then run your work through your plant and say, “Okay, pick up a quote and let’s write this quote. Here’s a purchase order, let’s do the purchase order. Let’s take this tablet and go out to your plant and step it through your exact operator interfaces. Let’s collect that data. Let’s track it; let’s sit down in the conference room. Let’s pull up the reports, the inventory.
That’s one thing that we really like to do early on — whether it’s in the late sale cycle or evaluation or sale cycle or early on in deployments — is prove it works. Because one of the things we’ve seen time and time again are the horror stories with plant operating software: we’ve tried for 18 months and we put in tens or hundreds of thousands of dollars, and we pulled the plug. We went back to whatever we were using before. So, proof of concept really quick.
The other thing is the actual deployment going fast. We like to go hot and heavy, kind of rip the Band-aid off, not drag it out over months. We like to go 2-4 weeks, training the team, iterate, iterate, iterate, test, test, test, test, test, because everyone says how it should work. We build it, and we test it and then there are always a couple of tweaks.
From there, we have a customer success team, as well. Whether it’s a phone call, email, zoom call, or training, we like to have someone available to answer the phone. Our philosophy is that your digital infrastructure isn’t something that should be optimized for low cost because we feel that the service that comes along with it — including a team of experts — is critical. Like a wiring harness for your automobile, it runs your entire plant. When your plant changes and you have changing requirements, to have “pros on your bench” to make adaptations real time in your plant is highly valuable.
Our approach is full service. We kind of own the functionality of the system in your plant — anything from the deployment to ongoing support. We’re always one degree away from engineering, as well. Occasionally when there is a bog or something is not working properly, our engineers are right there on calls with customers to make sure everything is working properly.
Doug Glenn: They can jump into the system and reprogram, as necessary.
Jeff Halonen: Absolutely.
Doug Glenn: Run us through this one then, if you would: Let’s say there’s a company — and you’re going to have to take a very generic-type job shop — that has your system already. It’s fully integrated and everything is working well. What does it look like? The reason I ask this is question, is I think a lot of people have trouble envisioning what it would look like and how it would be different than what I’m doing now.
So, a company that has your system, they go to do a quote for a job, and they receive an inquiry. Is the process they follow any different than what they would have done before your system was there? Let’s just take that first step, process and RFQ: how is that going to change from if they didn’t have your system?
Jeff Halonen: Something like quoting can be a fairly dramatic change, depending on what the current system is. Current systems can be Microsoft Word and Microsoft Outlook and maybe an Excel sheet to track the quotes. A lot of time to administer them and then there is pretty limited visibility. The transaction that we’d see there with Steelhead is that you can actually build a fully automated quote so you can define the algorithm: all of your flow charts, your Excel sheets, square root functions, looking at the mass of the part, specific gravity, material, etc. We can even reference the current price of natural gas. So, you can build models to reflect how your business runs and reference the past performance of that job.
What that looks like for quoting is that you’re able to build the quote all in one system. You might have multiple people involved in the quote: one person is working with the customer to verify requirements, verify inputs, ask questions, gather more data, and then it lands in someone else’s inbox. Now, a streamlined workflow allows the owner or the plant manager or sales manager can now approve that quote and then sending it to the customer instead of sending it through email and downloading and moving things around, we can send it right out of the platform, attach a file. We can track and see when the end customer is opening the quote. Also, your sales team has a lot of tools. They can say, “Hey, what quotes have never been opened? Hey, we should probably make a phone call. We put a lot of effort into those quotes, why don’t we call them and see what’s going on?” There are a lot more analytics but also just speed.
Another thing we’ve heard a lot is that a customer calls, they send a purchase order that references a quote to go, and find that quote can be a challenge because it’s in someone’s inbox, unread. Instead, they can have one platform to grab the quote, convert it to an order, and now you’re running production! It’s all the same data, it’s your plant. You’re not going to start quoting using different laws of physics. It’s about streamlining what you do today and replicating what you do today, but improving the customer experience, improving your team’s experience, and quoting faster, which leads to more jobs.
Everyone wants the thing quickly. If you can go from two days to two hours or from five hours to five minutes, that can have a dramatic impact on winning deals. For example, I’m in Chicago on vacation, and a customer calls, asking, “Where are my parts?” or “What happened to these parts?” Option A is you’re making phone calls back to homebase, someone is rummaging around looking for the information, and it’s not very enjoyable for you, your team, or your customer. But, it could be you put that customer even on speakerphone or mute, look for 30 seconds, and just text them the file right off your phone in real time! It’s the speed of accessing data, customer service, and actual insights, but a lot of it is workplace satisfaction.
We’ve literally been in shops where people are threatening to quit saying, “Our system is so bad” right in front of the president of the company. . . “I’ve been saying for years, we need a new system. I’m at my wit’s end because what I have to do is so painful and repetitive every day just to make the plant go.” It can manifest in a lot of different ways.
Doug Glenn: Give me a quick history on Steelhead.
Jeff Halonen: Steelhead Technologies was founded early in 2021 by a team of software engineers, mechanical engineers and manufacturing engineers. We started with the need of a single customer who talked about their pain. This particular customer was running on paper; whether it was contacting customers, instructing operators, tracking information, finding information, even something as simple as a customer calling with a phone call asking for information on a job, all were huge disruptions in their day. It took half an hour and they had to physically go find people and paper. Everything was in a huge storage room at the end. What the Steelhead team heard was a high degree of pain and essentially a plea for a solution. So, we founded Steelhead Technologies to address that need.
The team has grown really quickly, and we’ve raised 2.5 million dollars in investment capital here in 2022. The team went from an original team of about 6 to up to 16 now. As I mentioned, everything is kind of the direct service model. Whether it’s sales, deployments or customer success, it’s all team Steelhead across the board. Everyone is an expert in the platform, and everyone is one degree backed up by our team of engineers. We have six engineers that work full-time on Steelhead making it the best that we can.
Our customers are amazing. Everything that we work on is driven by a customer request, someone out in the field, using the tool and saying, “Hey, what if. . . ?” or “Hey, this could be better” or “This extra effort. . .” or “We would really like to know this. We are trying to achieve this business objective, but we need that.” That’s really where the Steelhead team excels is taking essentially what we view as requirements and building a platform to meet those requirements.
It’s really exciting. What we found is that [we can really help] the job shop industry, specifically, and also the high throughput — or even the high-mixed captive shops as you mentioned — where there are no build materials, no engineering team, no design release, and no supply chain. It’s go, go, go. The assumptions change by the day, hour, or minute, and everything is fast-paced. We’ve seen a huge lack of automation and data in the space. With high quality, purpose-built technology for this fast-paced, almost, service type of manufacturing, with specifically built tools, we’re able to achieve minimal effort to achieve each job. To minimize all of the administrative efforts, but also aggregate extremely detailed data down to operators, part numbers, and equipment, without driving up costs or sacrificing ease of use. That’s our core strategy.
Our customers are incredible, and a lot of our recent success has been just keeping up with our customers since they’re always pitching for a new tool or feature. We roll out a couple thousand lines of code to address this and they’re like, “This is good, but it needs that.” That’s the other thing — software is a service, so we’ve long accepted that we are never going to build the software and just ship it: it’s a continual living beast, whether it’s cybersecurity, speed, everything. So much of it is service because this type of software [requires] pros to extract the maximum value from it.
Doug Glenn: All right, last question. This is your challenge question for the session here. You, personally: What do you find most interesting and exciting about it? When you look into the future, what are you excited about?
Jeff Halonen: What excites me the most is the transition that shops can see by adopting technology and systems whether it’s impact to the bottom line — which we’ve seen in some shops, where it’s really stunning — or something else. A marginal labor cost where one customer’s job takes three minutes and another takes nine minutes. . . The manufacturer can now address that 3X marginal labor cost with the data where, before, they were just eating it. “This customer is not profitable; we’re not doing that; we are no longer accepting batch sizes below X, because we’re armed with the data.”
What really excites me personally is the journey that each of these plants are able to go on. I also get really passionate about when I’m able to see a difference between the way things are and the way they could be. That’s what every single interaction with our partners is: we see the way that they are and we work really hard to build the best one possible. So, it’s like, hey, look at this difference? Look what we can do? It’s that sense of accomplishment on a customer basis to really get as much as is possible out of their time, resources, people and their entire enterprise.
Doug Glenn: It’s kind of nice seeing the smile on your customer’s face, I think, right? When things go well, to be able to say, “Man, that was good!”
Well, Jeff, thanks. I really appreciate it. It’s very exciting. It sounds like you guys have got a lot of interesting things going on and, hopefully, some of our readers/listeners/viewers will be able to give you a call but thanks for joining us.
Jeff Halonen: Thank you, Doug. Thank you very much.