Rumors of the death of American manufacturing have been greatly exaggerated. Factories across the country still churn on—and are growing. In fact, there aren’t enough machists to meet the demand, and compounding the issue, the more experienced machinists have been moving on, replaced by a younger workforce.
To complicate matters further, managers lack the insight tools to gauge productivity and respond to downtime. After extensive research into these issues, John Joseph co-founded Datanomix, a company that offers automated production intelligence for the manufacturing industry to help manufacturers understand what’s really happening on the factory floor.
More information: https://datanomix.io/
John Joseph, a graduate of Worcester Polytechnic Institute and Clark University, is the CEO and co-founder of Datanomix, a software company empowering precision manufacturers to transform their business through data. John has an established track record of commercializing technology, having held leadership roles at several startups and their acquirers, including the successful acquisition of Equallogic by Dell.show more
At Datanomix, John is helping build an agile team that is using machine learning and advanced analytics to automate the delivery of deep insights directly from CNC machine data, all without any operator input. Dubbed automated production intelligence, the Datanomix platform offers a window into a factory’s operations in real time, and delivers deep insights into productivity and profitability trends over time. For more information about Datanomix and its data-focused manufacturing solutions, visit www.datanomix.io.show less
DISCLAIMER: Below is an AI generated transcript. There could be a few typos but it should be at least 90% accurate. Watch video or listen to the podcast for the full experience!
John Joseph 0:00
In order to meet the increased demand from their customers and increase demand from the marketplace, they need to amplify and scale their businesses rapidly and do it in an efficient and profitable way.
Alexander Ferguson 0:18
Welcome to UpTech Report. This is their apply tech series UpTech Report is sponsored by TeraLeap, learn how to leverage the power of video at teraleap.io. Today, I’m excited to be joined by my guest, John Joseph who’s based in Nashua, New Hampshire. He’s the CEO and co founder at Datanomix. Welcome, john, good to have you on.
John Joseph 0:37
Thanks for having me. appreciate it very much, Alexander.
Alexander Ferguson 0:40
Now Datanomix, I’m saying correctly, is a production monitoring and performance analytics platform specifically for factories? I understand. What was the problem that you initially saw and set out to solve?
John Joseph 0:54
Yeah, I think when we first started the company, we met with lots of customers about the problems they were having, managing their production environments. What they were seeing was that their cut their operators were changing in the demographics and capabilities, they were losing the experienced operators that were deeply experienced machinists. They were hiring younger people without the deep machine experience. And they needed a way to work with this new demographic, in a way that allowed them to see productivity, respond to productivity, respond to downtime situations in the factories. And we compared that to products and technologies that were on the market at the time that we’re addressing or trying to address these issues. And we felt like there was a significant gap. In the pain point we were hearing from customers and the solutions that were available on the market. So we went to work canvassing several companies within our region, to talk to them about a number of different facets of their operation, to try to understand how they ran their businesses, what they needed for metrics around their business. And the rest started to unfold from there.
Alexander Ferguson 2:13
So you saw a need for really, this, this infusion of technology in the process that would help these new folks coming in that didn’t have that, that same know, how am I getting that? correct?
John Joseph 2:25
That’s correct. The, the problem that they were having was first, they’re looking there they are looking for it’s not words, it’s a current situation, they cannot find enough people to operate these this machinery. It which is one massive problem they’re all trying to solve. The other massive problem is, the amount of reshoring of manufacturing operations back into the North American continent is happening at a epic, epic proportions. And so in order to meet the increased demand from their customers, and the increased demand from the marketplace, they needed to amplify and scale their businesses rapidly, and do it in an efficient and profitable way. If they’re not profitable, they don’t last very long. So they needed a technology to assist them to do that amplification of their capabilities. Then you
Alexander Ferguson 3:23
pay you paint an interesting picture here of it’s kind of a happy problem one could say of, okay, lots of businesses coming back, US soil factories are have lots more requests, but the problem that comes with that is okay, now we need to actually have the people to meet that demand. And we there’s already a skill shortage or a skills gap there of having those people. And ideally, that’s where technology can can help is stop the gap or help with that, that close the gap. Close the Gap. So you saw those knees and start talking to the factories what what year was this that you started to do this help me understand
John Joseph 4:02
2017? Okay, I coined a phrase that I call digital leverage. And I use digital leverage to mean that owners of these companies needed a way to get mechanical advantage on managing the productivity of that factory. And so in 2017, we set up several focus groups with large companies in the northeastern part of the United States. those focus groups were with the leadership team, some some call them executive teams, some just call them leadership teams of about a half a dozen companies. We invited them into these meetings and ask them some very simple questions because what we were doing was we’re trying to explore their knowledge of this industry, industrial, internet of things for Dotto capability that people talk about this manufacturing for Dotto phenomenon that’s being advertised and marketed by all kinds of people out there. We learned first and foremost that it’s a rather ambiguously defined term, it’s not well understood by people in manufacturing, to, I think their interpretation of manufacturing for Dotto means that you’re putting sensors all over your factory. And you’re using those sensors and the Internet of Things to collect data from your factory. And three, when you get that data, what are you actually going to do with it? Do you have the skills to manipulate that data to give you the kinds of outcomes that you are expecting, when you initially scoped out where to put these sensors? So our focus groups started with three questions to the to the to the organization, it was about a half a dozen, two to 10 people in each of the focus groups we ran, we asked them the first question, which was, if you could put sensors to build on this sensor conversation, if you could put sensors throughout your factory, where would you place them, and we got 26 different locations, all over the factory from air intake, water intake, all the way to coolants. screens and filters, motors, motor temperatures, tolerances, all kinds of things that you might imagine, factory owners would be interested in keeping an eye on. We captured that. And then we asked them the second question. Now that you have these sensors placed all over the factory, what would you like to deal with them? For example, if you could run correlations between sensors? What types of correlations would you would you run would you execute? And they said, you know, we’d like to correlate incoming coolant temperature with machine tolerances. And understanding how tolerances drift over time and temperature. Those are all great things. And it occurred to us after we started doing this, that it wasn’t crystal clear what they were trying to solve their what they really wanted, was a way to manage their business in predictable ways. using technologies that were simple for everyone to understand from the top floor all the way to the shop floor. And it occurred to us that what we were doing by asking these questions was we were getting into the minds of the management to understand what these things were that they’re looking for, for from their business. The third and final question that we asked again, to drive a conversation, and these were very healthy conversations. The third and final question we asked was, now that you correlated all this data, what would you use this data for, relative to managing your business? What business drivers? What business levers are you trying to pull by using this data? And more conversation ensued? And I should also say that these were conversations where no holds barred, say, What’s your feeling? You know, a lot of people aired a lot of frustration about eirp systems that were antiquated the minute the data was entered into the RP system, meaning they were not getting real time information about their business when they needed that information. And so what they would then do is they would go out into the factory space, and they would start asking questions, and collecting data off the factory floor, bringing the data back to their offices, consolidating it into spreadsheets, building spreadsheets, and trying to look for patterns and trends in data and predictors of cycle time predictors of productivity output predictors of the amount of scrap that they were producing over time predictors of whether a job was running on plan and add speed or Off Plan and off speed, and what would they do about it. So we had lots and lots of these conversations, we took all that information, we came back to our company and we we whiteboard it out, we just we started to deeply think about each of these things. In parallel with that we were also examining this, this protocol called mt Connect, which is a data standard that is used on the control systems of CNC machines to collect parametric data off of these machines, and broadcast that data to people who are looking to capture sensor information from that machine as that machine was going through its production cycle from one part to the next. So two things were happening simultaneously. One, we were drilling deeply into the mind of the buyer to understand their pain point. And two, we were looking at latent data that existed from their machine tools, that they didn’t have the skills and where with all to derive the kinds of objectives around production management that they were looking for. Long story short or shorter. We converged on Something that my co founder Greg McHale and I were really focused on at the time, which was to develop a productivity index.
And change the game from a game of machine utilization to a game of, of production, scoring and scoring production and giving people a real time indicator of their of their production, as compared to benchmarks that we saw as the machines cycled through production from park to park department. And so there’s a much deeper disrupt discussion and description of the problem, the software itself, but I think for where we are in the conversation, that’s, that’s sufficient.
Alexander Ferguson 10:44
There’s so much there to unpack for sure. One thing that immediately comes to me though, is the fact you did a focus group. I mean, I feel like a lot of other tech companies, I’ve interviewed them, they’ll just jump into building a product. And then they’ll try to say, Okay, now who needs this? Now, I’m understanding on you. Your background is not in factories or this stuff. How did you get to this? Where did you come from that that you jumped into into this industry?
John Joseph 11:12
Well, I think that goes back to my co founder and I, we have a really excellent balance of skills, I come to the table with a mechanical engineering background, having been educated there, having worked in factories for a lot of my early career, specifically in machining factories, as a way of earning extra money on weekends and things when I was younger and more available. And he comes to the table with years of enterprise data management experience having worked in storage companies from when he was in high school, through college, and for the last 15 or so years that he’s been out of college. So you bring these two skills together, one person looks at a problem mechanically, the other person looks at a problem from a data data science perspective, you bring these two things together to produce some wild results,
Alexander Ferguson 12:05
where you and Greg Ray, your co founder, spitballing. This idea for a while that it just suddenly culminated in in in was 2016 or something to start to go What was that? What was that process? Like what Tell me about that journey?
John Joseph 12:19
Greg first started examining data, opportunities, data, volumes, volumes of information that were out there that were nascent, nascent markets, nascent pools of data that were available, to be mined by someone, and not being mined. And through a number of introductions of people in our alumni, community, venture community, corporate community, through all these introductions that we were presented with, we zeroed in on manufacturing as a great place to focus our energy relative to this data stream that was being produced. That was one thing. Also, the two of us have a passion for helping apply digital leverage to American manufacturers. And so I’ve got a soft spot in my heart, having come from a manufacturing family, and having watched manufacturing in the in, in the United States in the 1970s and 80s, depart the country, and now a resurgence. And so how do we help that to be successful? How do we protect intellectual property within the borders of our country, and not outsource it to people who really could care less about intellectual property, so I have a huge passion for that. And so, you know, we’re here to make manufacturers in America successful, right? It’s, it’s amazing when you meet these people, how innovative they are, how authentic they are, how hard they’re trying to make product, high quality product and give it to sell it to customers at a profit to to continue to grow their business, and invest in their businesses. We’ve had some of the best experiences with these manufacturing leaders. In the last four years that I’ve had in the last 35 years, I’ve been in this in the computer industry. So it’s been it’s been fantastic.
Alexander Ferguson 14:19
I feel and I can definitely sense that the underlying passion you have from from your, from your own history, is that okay? There’s got to be a way we can help, then manufacturing here and technology is the vehicle or tool to empower the mortals to make that happen. Coming back to this journey of the shared of Okay, you and your co founder seeing this opportunity from different perspectives converging you do this focus group, what year did the focus group happened was curious
John Joseph 14:48
to happen in 17. As we were unpacking the data, we were unpacking the focus group feedback at the same time,
Alexander Ferguson 14:55
and then how long until you start to be able to build based off of that, that insight.
John Joseph 14:59
It took us about two years to to create a product architecture and then start to build out and develop the architecture with our software people.
Alexander Ferguson 15:09
And then being able to roll it out. I’m curious does for those folks in manufacturing, do they get it like to click with them right away? Or what’s that process of adoption look like?
John Joseph 15:19
Yeah, that’s a great question. And it allows me to pivot to a another topic that’s very important here. manufacturers that we sell to our experts in subtractive, and additive machining, right, so you understand 3d printing, that’s a form of additive manufacturing, you understand that a CNC machine is a subtractive process that takes a piece of material, a block of material, and removes portions of that material to create the finished parts, those are the manufacturers in America that we are targeting. And what they’re really, really good at and wake up every day to go to work to do is to perfect that process. Because the better you are at that process, the higher the quality part you produce, the higher the profits, the lower the scrap, the higher employee morale because they feel like they’re making a solid contribution every day. And so they’re 100% focused on manufacturing materials, we come to the table knowing that that’s their expertise, and that their expertise is not data science, we bring the data science to the table and marry it up with the material transformation. And we do this digital transformation for them. And we want to they asked us in these focus groups very specifically, they said, Please do not ask our operators to be data input operators, instead of machine operators, we want our operators to stay focused on machining, and not focused on giving you the information that you just tabulate and feed back to us that’s not helpful. The other thing they said was that machine utilization, while interesting is not at the core of what we’re looking for, we need something bigger, better, simpler, stronger to indicate visually, which way our production environment is headed. And so two important things came out of that the Prime Directive for data, nomics was Do not ask operators to input data, let’s get the data directly from the machine. And so we focused only on the data set that was coming from the machine and any auxiliary information that was being produced from sensors around that machine, on that machine, etc. And the other was, we needed to present them with a user interface that presented in the exact same logic flow, that they think about their business and how they manage their business every day. And so the culmination of all that was what we call data anomic fusion. And when we started to present that to customers in the field, as we came back out with the the launch of the product in early, alpha and beta phases of the software, they said to us, when they saw the the early instantiation of the product that they asked us, how many years have you worked in a machine shop? How many years have you worked in manufacturing? We said, none? Yeah, I wasn’t there. And they said, none. We’ve never worked in a manufacturing environment. Well, then how do you know how I think about my business? This is an exact mirror image of the way I think about my business. And you presented data to me in that flow and logic pattern. The point is that the product needed to deliver to the customer, a point and click environment, which said, I’m going to give you a visual cue on whether you’re doing well to benchmark or poorly to benchmark. If you’re doing well, to benchmark you’re well suited to ignore that production cell work. So if you’re not doing well, to benchmark, that’s where you need to focus your energy. And you need to focus the the scant few manufacturing engineers that you have working at the company to go solve problems. So
Alexander Ferguson 19:15
it’s taking information that can be complicated, and they don’t really care about but simplifying it to exactly the information they need. Yes, good. You’re doing good. You’re doing bad focus here. You’re good here. Don’t worry about here. Yeah.
John Joseph 19:29
So what percentage of people in manufacturing Do you think went to elementary school? or high school? 100%, right. And when they went to these schools, weren’t they given a letter grade on their performance at school? So So do you think that they understand a letter grade system? Absolutely. So why wouldn’t you use something that is familiar to you in nature, and use that as your visual cue of goodness and badness. And so what we present them with is a large TV screen on the production floor and in the executive suites. The production screen shows the product, each of the workstations that are producing parts. And we are giving them a letter grade from a plus to C minus on the performance to benchmark imperative that we do by watching without asking an operator to tell me what your cycle time should be, how many parts you need to produce per hour, etc, we learn that the software learns that automatically from the control system of the machine, we set the benchmarks and then we watch performance against benchmark over time.
Alexander Ferguson 20:33
I want to take a moment to actually talk about that for a second because we as humans don’t like entering data. I mean, we don’t live to say, Oh, I can’t wait to enter date i’d same on like salespeople don’t like to enter data into CRM, and manufacturing floor don’t want to have to spend their time on that. So ideally, getting the data that already exists pulling that in you focused on that, was it difficult, like with that data already exists? You plug into it? Did you have to do additional things to be able to get more data out of those machines?
John Joseph 21:01
Yeah, good question. The data needed to be connected to very simply, it’s there. Nobody really does anything with it. And so by using a simple Ethernet plug connection into the back of that control system, that data is now being transmitted to data nomics. We use a very lightweight, cost effective computer, to crank through that data to separate wheat from chaff, pull the parametric data that we need from that system, watching how the parts being manufactured. Understanding when we hit m 30s, m zero zeros, which are signals that the machine cycle has either started or is about to stop, we start to watch, we start to count. And we capture this into our algorithms that we developed as part of this process. Those algorithms then, are transformed into what the user then logs in through a web browser to experience the product firsthand. It starts at a homepage that describes all of their machines and how well their machines are running. And then as they click in, they go into a particular machine, they look at a particular job of interest. We give them setup times we give them setup costs, machine times and machine costs, we give them cycle times parts produce all of the information that they need to understand whether they’re going to be able to deliver a quality product to their customer at the end of the fiscal week or not. And if not, what are you going to do about it? So now that we’ve given you these visual cues, keep clicking in Why is Joe only producing 69 parts at a cycle time of four minutes, and Suzy is producing 100 parts at a cycle time of three minutes and 39 seconds? What’s the difference? Why is there a difference? Is there a machine difference between the two work cells etc. keep clicking in and understanding more deeply what what’s happening and who’s doing it.
Alexander Ferguson 22:57
You describe all this, it’s coming as a layman to this new manufacturing, I almost assume this is happening already, that they would know all this information, they’d have this data. But you’ve described at the beginning of our discussion here the pain point is they didn’t they weren’t they often they would be walking the floor you’re saying and the next day they finally tabulate it, put it in a spreadsheet and then maybe a day or the next day, you’d find out how everything is doing.
John Joseph 23:20
Yeah, I think there there were a couple of people that stick out my mind as I think back on trying to understand how they ran their businesses. First and foremost, they if they needed an answer to something, they needed to traverse the factory floor to find someone with who potentially had an answer for them. And a lot of times when operators feel like somebody is on a data finding mission, a fact finding mission, they kind of get a little bit quiet, right? So chasing that information becomes difficult. Other people said we’re in our offices trying to generate new business for the company, trying to order new raw materials, trying to order coolants trying to order new machinery. And we’re not always down on the floor able to watch things happening. But what we did was we put your product up on our wall on in TV mode, and we were able to see a job that was going awry, and go right to the production floor, essentially cutting our corrective action time in half. Because an operator’s first inkling is not to reach out to management to solve a problem for them. It’s to kind of wait, maybe the problem is going to go away, right? Instead of waiting for it to go away. Let’s proactively in predictively go to that person and help them out. In another case, a shop had seen an alarm code continued to trigger. And that alarm code told him that the the the guide bushing on his machine was about to fail. And so he kept seeing this pattern of performance to alarm code and he said, Gosh, the last time I saw this, the guide bushing was going south. He proactively replaced the guide bushing And saved all kinds of scrap that would have been made had it failed. So giving people the guidance system to talk is an amazing amount of digital leverage as I defined it a few minutes ago, amazing.
Alexander Ferguson 25:16
Your if you were to go back to yourself when you first started this year about four, four years ago 2017 and share something that you’ve learned now, it could be about the industry, it could be about the product, it could be just about running a new SAS platform. What would you go back and tell yourself,
John Joseph 25:39
I would remind myself something that you said a few minutes ago, and that is the power of observation of watching people work and move through space and time within a factory is so informative on how you develop products. back to what you said earlier, build it and they will come mentality is absolutely the wrong mentality. We are a we are a customer driven product development organization. And we use our customer inputs to create new features and functions for our product that just left a meeting a few minutes ago where the manufacturing engineer in Pennsylvania said, asked us a question about whether or not we’re going to do a new level of functionality relative to part cutting, cutting routines, cutting instructions, etc. And we said we were already working on it. And we shared with him some futuristic looking product features that we were in development on. And the guy said, I’m getting in my car, and I’m coming over to your company to hug you guys, because I’ve been worried about this for years. And you’re the first company I met who actually thinks about this the way we do. So that’s that’s why we have succeeded here.
Alexander Ferguson 26:52
When you when you started dating. Did you bootstrap this? Did you get funding from the beginning? How did you begin that double effort?
John Joseph 26:59
Yeah, bootstrapped. I invested in the company personally, a lot of family and friends who I have worked with and have known for many years and believed in me and believe in Greg invested in the company. So I took it very, very seriously that I was taking money from people that I cared about, and using that to build this company. And so when that occurs, you’re authentic attention. Laser focuses in on producing measurable results. And so that was a driver for me, and still is today.
Alexander Ferguson 27:34
You there was a way before the data gravity was there much like action that was curious. Was there anything? Or is it just a separate endeavor that you read as a
John Joseph 27:43
separate company altogether? That was in the data management, and analyzing data within a storage device? technology, very different from what we’re doing. Now, that was an on premise hardware device? This is a cloud based SaaS model.
Alexander Ferguson 27:59
And data definitely, still do.
John Joseph 28:02
That’s the common thread. Exactly, exactly. back to what I said about being experts and managing data.
Alexander Ferguson 28:08
Looking ahead, that if you kind of look at the future, five years from now, maybe even 10 years, if you want to go that far, what predictions would you make for technology prediction, if you make more for this manufacturing in factory industry,
John Joseph 28:24
I think that five years from today, you will see a very different set of technologies that are running successfully on the manufacturing floor. I think the days of people chasing parts and paper and you know, people are gone. That machine learning capability. Artificial Intelligence is replacing that by by putting A to D add technology on the floor that converts a motion analog to digital. that replaces people doing things and makes binary observations or digital observations of things, and turns those observations into inputs to machine learning algorithms and produces quality inspection reports for people that are automated, relative depart manufacturing produces firefighting guidance systems. And I don’t mean fire in a literal sense. I mean, in a figurative sense solving problem solving guidance systems that point people directly to the problems they need to solve. We already have predictive maintenance and service technologies that know when a bearing is going to fail on a particular machine, or that a fan is no longer cooling instead of electronics. Those are technologies that are in place today. And a bunch of optimization technologies that look at a process and learn From historic performance and an benchmarks on that process to improve it going forward, this notion of continuous improvement is absolutely as a long pole in the tent and here to here for quite some time.
Alexander Ferguson 30:14
For those that want to learn more about Datanomix you guys should go check it out Datanomix, as I stated Datanomix.io, and be able to look they can schedule a demo. Thank you so much for spending time helping us understand the journey that you’ve been on and the future of really manufacturing and the opportunity for tech involved there.
John Joseph 30:34
Appreciate the opportunity Alex, thank you very much.
Alexander Ferguson 30:37
And we’ll see you all on the next episode UpTech Report. Have you seen a company using AI machine learning or other technology to transform the way we live, work and do business? Go to UpTechReport.com and let us know
YouTube | LinkedIn | Twitter| Podcast