S2:E17 Data-Driven Approaches in Manufacturing with Bryan Sapot

Bryan Sapot is the CEO of SensrTrx, a manufacturing productivity and analytics platform designed to collect and display actionable data from the shop floor.


In this week's episode of Masterminds in Maintenance, we are excited to have Bryan Sapot, CEO of SensrTrx, on the show! Traditional maintenance and manufacturing are undergoing a huge transition to be more data-driven. Bryan and Ryan discuss this shift towards data and how to get teams on board for this revolution! Listen today!

Episode Show Notes

  • What myths hold teams back from embracing IoT?
  • What pitfalls are common after adopting IoT?
  • What trends are you most excited about?

Podcast Platforms


00:03 Ryan Chan: Welcome to Masterminds in Maintenance, podcast for those with new ideas in maintenance. I'm your host, Ryan, I'm the CEO and founder of UpKeep. Each week I'll be meeting with the guest who's had an idea for how to shake things up in the maintenance and reliability industry. Sometimes the idea failed, sometimes it made their business more successful, and other times their idea revolutionized an entire industry. Today, I'm super excited, we've got Bryan Sapot here on the show, Bryan is the CEO of SensrTrx, a manufacturing analytics platform designed to collect and display actionable data from the manufacturing shop floor. Welcome to the show, Bryan. I'm really excited to have you as a guest, you and I have been chatting for quite some time, we finally got you here on the podcast.

00:41 Bryan Sapot: Thanks for having me, Ryan.

00:42 RC: Bryan, could you share a little bit about how you got started in this field of maintenance and reliability in your background?

00:49 BS: I started my career as a software developer, a network engineer, but I always wanted to own and start my own companies, so I created a custom software company in the late 90s, and my very first project was building a custom ERP system for a chocolate company. So that eventually grew from order taking to everything happening out in the factory, like pick, pack, and ship integrating with machines. My journey through manufacturing just kinda continued for the last 24-ish years from there.

01:22 RC: I love it, and obviously I kinda know what you're up to now, which is SensrTrx, which is this awesome, amazing data-driven platform for the industrial IoT... What are some of the most common metrics, most common things that you're seeing in terms of the newest adoption of IoT in the manufacturing shop floor?

01:43 BS: What we see mostly because we're a productivity application is people wanting to really understand throughput, capacity, and downtime and quality issues on the factory floor, and downtime is really where, and a little bit of the throughput, but really where the productivity side and the maintenance side come together. And they have to play nicely. So once you understand what your downtime reasons are and why you're having waste and why you have losses, a lot of times maintenance has to get involved to figure out what's happening, and in a lot of our customers, they're using data to really understand that. So if you have a quality issue that starts popping up, maintenance can dive into the machine data, for example, and figure out, "Oh, somebody changed the settings, the pressures aren't right over here, we have a leak in a hose," whatever the root cause of that is.

02:38 RC: As you look at the next five years, the next 10 years of maintenance, reliability and manufacturing, what are some trends that you're most excited about?

02:48 BS: I hate talking about this 'cause they think there's a lot of false promises with it, but it really is artificial intelligence and AI, or building models around maintenance and reliability of your assets. The biggest trend, I think that we're gonna see is, or the shift is, there's a lot of big companies today who are trying to build these models on their own, like makers of things, they're not the ones that make the pumps and the motors and the machines, but maybe they're like Ford... And they have really big reliability issues on some of their equipment, so they're trying to build predictive models on failure, and I think one of the biggest shifts that we're gonna see is the actual makers of the motors or the things that you're... In the machines that you're trying to predict failure on are gonna start building those models and connecting to those machines because they're the experts in those. Somebody like Ford is definitely gonna have internal people that are experts in whatever kind of machine or robot that you have, but a mid-size manufacturer is not, and they could really benefit from that technology and all that aggregated data across all of those different robots, machines, and motors out there in the field across the world.

03:57 RC: I totally agree with it, because I think what's happening is that pump A from company and brand A even as the same rotations per minute, same horsepower as pump B from a different manufacturer from a different company, operate in slightly different ways, and we can't necessarily say everything that happens to this pump is gonna be directly relatable to hub B, but if you are actually the manufacturer, I'm sure you can set a standard for what's in range, what's out of range, and the manufacturer is actually gonna be in a better position to tell us as the operators is this in line or out of line.

04:48 BS: But when you get into a complex piece of manufacturing equipment like a CNC, for example, where you can do anything, and so there's so many different failure modes on that thing, depending on what you're doing... So the way it'll fail if you're cutting plastic or wood is different than titanium, and the tools make a difference too, so then having the ability to combine all that data together and contextualize it, and then know what the failure modes and what you're predicting based on what the thing is doing and what it's supposed to be doing, I think is really important too, 'cause a lot of what we focus on today is just... Do we have an anomaly? For example, is a vibration pattern anomalous, is the temperature to high, is the flow rate wrong, current draw wrong, things like that.

05:32 RC: Let's say the future moves more towards the predictive analytics, goes more back on to the manufacturer, what do you think happens to the internal maintenance reliability team that was previously running all the analytics in-house?

05:48 BS: I think that the manufacturers can build better models, but the people, the men and women on the ground working with the equipment every day is gonna know it better in that environment than the manufacturer is, so the manufacturer can predict a lot and understand a lot, but they can't understand everything. That team still has to do the work. Somebody actually has to do the PMs, do reactive maintenance when things break, so I'm not sure that it changes that much because I think those systems are still gonna be making suggestions that have to be interpreted by people.

06:21 RC: You know, a lot of people talk about AI machine learning taking over all of our jobs, all of our roles, but when you think about it, what it really does is it surfaces and highlights problems that ultimately we need to take action off of, and that's gonna be the internal team that's actually on the shop floor.

06:39 BS: Where we're headed, and I think where you're headed with AI, machine learning and really making smart systems is to have the systems do all the grunt work that the engineers have to do today. Collect all the data, put it into spreadsheets, type it into systems, look for patterns, and automate that. Be the eyes on the data for the engineers, especially when you talk about the combination of the kind of data that we collect and systems like yours, you can move towards prescriptive because once we start to understand the patterns and failure modes in the systems we're monitoring, you match that up with historical maintenance records, you could say, Hey, last time this happened, this is what we did, and this is likely the cause and this is the likely fix for it... That's huge, 'cause you're not reinventing the wheel and trying to really understand what happened maybe two years ago when that thing failed last time.

07:30 RC: Kind of on the same gear here, Bryan, but you and I were chatting probably... Was it two, three weeks ago about the common myths that holds maintenance teams back from fully adopting, embracing this, what we believe is gonna be the future of technology in the industrial space, so I'm kind of curious, Bryan, any highlights from that? What are some common myths that you think are really holding people back from fully adopting industrial IoT?

08:01 BS: I think the biggest one is that... And people will disagree with me, I think this is the LinkedIn conversation that you're talking about, [laughter]... people disagreed, but they think it's expensive and it's complicated, and it isn't... And it doesn't have to be. I know it's all relative, but this is a tens of thousands of dollars thing, not a millions or hundreds of thousands of dollars things, so that you can get started using data on your machines and monitoring them. I think that's one of the... Probably the number one in my opinion.

08:34 RC: You're right. Cost is an important problem, and I think what people called out... It wasn't just the cost of sensors, it wasn't just the cost of even purchasing the hardware and getting it set up, it was actually the cost to implement, train, and also have an ongoing person or people to be able to analyze this data, versus getting it first set up. So I'm curious, any thoughts around that, Bryan?

09:04 BS: I think that as engineers, we tend to over-complicate things sometimes, and want the perfect solution that does everything that we want, when in reality, maybe a good starting point is like a check engine light solution. I don't need to know down to the itty bitty detail of every little vibration, every millisecond. I just need to know that something's gone haywire to move me down the path, starting there and being able to prove value with something small, which is something we talk about a lot, start small, think big, move fast... If you start with something small, all of the costs that you just described are a lot less than if you're trying to do a much larger system, collect a whole lot of data sensor up something with a lot of very sophisticated sensors. That's one of the myths, is that you don't have to go to the perfect solution and that you can start with something small and still get a lot of value out of it, and that's where my, "It doesn't have to cost a lot of money," comment comes from.

10:06 RC: Absolutely, and I think you hit the nail on the head where a lot of people when they talk about IoT solutions, it's like, "Alright, we're gonna set up a digital twins. We're gonna run all these different scenarios of what happens," but then it's like, "Hold up, where are we right now," and you kinda look at the existing set up and you're like, "Alright, well, let's get there," but there's so much value in just taking the first step. I know we're gonna get to that predictive modeling, setting up digital twins, running 30 different scenarios, but let's just get that check engine light because I think we can get half of the total value of this for 1% of the total cost. When we do finally make that purchase and we say, "Alright, we're gonna make the investment into more predictive analytics and getting ahead of our failure modes and breakdowns, finally make this investment," what are some common things to actually avoid? Where are some common pitfalls that people make after they've made the jump?

11:15 BS: One of the biggest ones is not including everybody, not everybody, but if you're coming at this from a maintenance perspective, typically you're gonna keep it within the maintenance department, and again, I'm focusing on manufacturing. So you're gonna keep it within the maintenance department and maybe not involve production or manufacturing, maybe not involve quality. But the three groups, those three groups, have to play very nicely together to achieve the goal of the company, so you wanna make sure that you're including other folks, and the same thing happens... Our primary, the people who primarily buy SensrTrx is the production teams, like VP of ops, plant manager, and they don't always include maintenance in the discussions, and then we show up with sensors or start talking about connecting to machines and they're like, "Wait a minute, what are we doing?" So you have to include both sides 'cause otherwise you're gonna be dust into failure and you need to think about that in the beginning.

12:16 RC: How do you align... In any suggestions, how do you align production and maintenance reliability?

12:23 BS: This battle, this exact battle, is one of the core reasons why we founded SensrTrx, because production, when they don't hit their numbers, often says it's because maintenance can't maintain the machine, and then maintenance says you guys set it up wrong and you just kinda go back and forth, pointing your fingers... And so our idea of what the common ground is, is data. If I can get good reliable, even if it's simple, data to tell me where the problems are coming from, and what's causing it, then it's no longer subjective. This is the data, I don't really care whose fault it is, let's just figure out a way to fix it. I think I'm being optimistic when I think that data solves everything, but at least it gives a common ground, a common language, a common set of measurements for both groups to work off of and talk about.

13:17 RC: It's very objective, it's not about finger pointing, it's how do we take this number and make it better.

13:22 BS: Yeah, exactly.

13:22 RC: What's something that you wish more people knew about within the maintenance reliability space?

13:27 BS: There really are high quality, low-cost solutions to start monitoring these things, and you can get started collecting data for a relatively small amount of money that will make a pretty big impact on your business. It doesn't even have to be SensrTrx solutions, there's a lot of really good stuff out there where I can put a really high quality temperature and vibration sensor on a motor in a plant and get that check engine light level of data and confidence, and even anomaly detection for a very, very low cost. I wish more people knew that because I think more people would get into it, and I think it would kinda bridge that divide that you and I have talked about between the very data poor in the very data rich.

14:15 RC: For all the listeners, it's basically this idea that you've got so much data, you don't know what to do with it, or you got nothing at all, and I think what you're mentioning, Bryan, absolutely, it's how can we that middle ground of the right amount of data that you can actually take action off of. Where do you go and find yourself learning from, and where do you continue learning essentially?

14:37 BS: I'm a LinkedIn junkie, so I learn a lot from LinkedIn. I actually think that's how we got connected. And I've developed a lot of relationships like this where I've learned a lot from conversations I've had on LinkedIn. Also, I follow a lot of Lean thought leaders, consultants, that post a lot of book recommendations, and I've read a lot of the books that they've recommended, and it's really kind of opened my eyes to things that maybe I saw the different ways to think about it and potentially how technology could help.

15:11 RC: That's awesome. And do you have a favorite book that you'd recommend?

15:15 BS: From a manufacturing perspective, "The Goal".

15:18 RC: "The Goal", alright, I'll have to check that one out myself.

15:22 BS: It's a story about essentially continuous improvement in Theory of Constraints, but it is told more like a novel than a dry business book, it's really probably the best business book I've ever read. I'd highly recommend it.

15:38 RC: Awesome, thank you so much for sharing that, Bryan. Just to wrap up up here, can you share with our listeners the different ways that they can connect with you, follow you on your journey.

15:48 BS: Yeah, the best way to connect with me is really on LinkedIn, so Bryan B-R-Y-A-N, Sapot, S-A-P-O-T, I'm the only one in the whole world, which is kind of cool, feel free to connect with me and ask me anything on there.

16:02 RC: Alright, there you have it, you got Bryan Sapot, the only one in the world, here on masterminds in maintenance. Thank you so much to all of our listeners for tuning into today's Masterminds of Maintenance, my name is Ryan Chan, I'm the CEO and founder of UpKeep, you can also connect with me on LinkedIn or shoot me an email directly at [email protected] I hope to connect with all of you guys soon, until next time thanks again, Bryan.

16:24 BS: Thank you.

Join the Masterminds in Maintenance Podcast!

Are you an industry leader in the fields of maintenance and reliability? We want to hear from you!

Please sign up to be a featured guest on our podcast here!

Stay tuned for more inspiring guests to come in future episodes!