Zen and the Art of Manufacturing – Episode 3 ft. John Broadbent

Transcript

Bryan Sapot (00:06)

So hi, I’m Bryan Sapot the CEO of Mingo Smart Factory and also your host for Zen and the Art of Manufacturing podcast, where we talk about manufacturing technology, software, lead initiatives, and then also workforce development. And today I have with me John Broadbent from Australia. And John, if you don’t mind like giving us a little bit of an introduction and a bit of your background and then we can kind of get into it.

John Broadbent (00:29)

Hey Bryan, thank you for the opportunity and good to reconnect post COVID across the big divide. Yeah, my background is in manufacturing, predominantly in Australia. I have done work overseas, but next year marks my 50th anniversary in the Australian manufacturing sector. So like you, I’ve earned quite a few great years. And I find that my experience over that time, particularly in education, coaching and mentoring role, has helped quite a few organisations understand what the whole journey of Industry 4.0 and digital transformation is all about. So I’m looking forward to sharing some of my experience with you and your listeners today.

Bryan Sapot (01:08)

Yeah, thanks. It’s interesting that you mentioned, you know, across, you know, the world you being in Australia, but also the last time we did a podcast was either late 2020 or early 2021. And, you know, when we reconnected recently, we started talking about how there’s still a lot of resistance to digital transformation on the factory floor, IoT, whatever we want to call it. and we’re surprised still three years later, three and a half years later at why that’s still happening. Why do you think that’s happening? And really what can be done about it?

John Broadbent (01:47)

personally think that some of the larger end of town vendors have a bit to answer for because they’ve gone out there, particularly in Australia, which is a fairly immature market and tried to sell them the silver bullet. know, cloud-based, I’m thinking AWS, Google, Microsoft IoT Hub, sorts of rockwell, those sorts of plays in the marketplace. I ran a survey last year in Australia here. I got quite a few respondents and the question was how much of Australian manufacturing do you think is being made on already fully depreciated kit? And I reckon it’s over 80%. And in fact, the survey showed that I think 80 % of the people thought it was over 80%. So we have a lot of data and equipment. As a country, we’re a low population, 27, 28 million, mostly coastal. in a very large country. And as a result of that, we have to make lots of variety of things. So we don’t have clients that have 25 manufacturing lines. know, one of their biggest dairy companies has 38 lines, but it’s across the whole of Australia. So one plant will have five, one plant will have four, one plant will have two in smaller population areas. And the result of that is that we don’t have investments in high volume equipment here. We have investments in equipment that can make lots of variety. And as a result of that, have lots of changeovers. have low OEE generally across the board because we very rarely have a line that’s dedicated to making one thing like you might have in the States where you have loads of populations. And as a result of that, those manufacturers don’t have kit that is even industry 3.0, obviously we’re compliant, but it means they don’t have the ability with those pieces of equipment to close out industry 3.0, which is computerizing, so you’ve got something like a PLC or other controller on there that you can get data out of and that can network it, so connect it to a network which then sets the stage for that information to be collected and used in an industry for digital transport environment.

Bryan Sapot (03:55)

Yeah, that makes sense. never really thought about, was on a call with another group from Australia yesterday who was like interested in being a partner and they were sort of explaining their customer base and the small machine counts and even like our pricing models probably, you what you were talking about with Google and Microsoft and AWS, their pricing models probably don’t work for those smaller plants.

John Broadbent (04:21)

In fact, I was selling an E-engine here out of Canada and they were saying, you’ve got to get into the car industry and sell hundreds of these things. I went, we don’t have a car industry. The government nine years ago made a decision to no longer support it through subsidies and so Ford and General Motors and Toyota said, you later. So we lost 80,000 manufacturing jobs upstream and downstream. as a result of that decision. And so where manufacturing was I think 14 % of our GDP 20 years ago is now down to 6%.

Bryan Sapot (04:52)

Yeah. Well, and you’re surrounded by manufacturing too, right?

John Broadbent (04:55)

We have a lot of work. We are, but I wrote an article a couple of years ago that you may not have seen called Time to Pay the Piper. And I won’t go off track too much, but just to give you an example, it was a company in 2018 who had a roll forming machine making racking for warehousing, doing really well. I went and had a look at the machine and I reckon it was 1980s vintage. So it would take flat sheet, it would roll it into a square tube and weld it. And that machine was being exported to Malaysia. And I said to the guys, why are you exporting the machine? lower labour rates. So rather than invest in a modern machine that didn’t use labour, it was easier to export the machine to a lower cost labour market, make it there and then import it back. When COVID hit, supply chain disruption happened. They couldn’t get stock to sell. They’re now screwed because they now can’t bring an inefficient machine back to Australia. So what they would have to do, and that’s why I called it time to pay the piper, they would now have to find land, build a building, buy brand new up-to-date state-of-the-art equipment, find people to then run those machines and basically start again. And the capital investment to do that is probably way more than the money they saved by offshoring the inefficiency in the first place because the labour rates in Southeast Asia are now starting to climb. East Coast China I believe is matching Singapore. So they’ve basically painted themselves into a corner. So reshoring in Australia is difficult unless you already have some sovereign capability.

Bryan Sapot (06:36)

Yeah. That makes sense. Yeah. Sorry for the detour there. So to kind of go, go back to your original point on, you know, the aging equipment in Australia, I would bet if you did that same survey in the U S the numbers would be the same. mean, just across our customer base, I would say probably 40 % of it we’re connecting to PLCs and controllers and the other 60%, like we have an overlay like add-on solution to be able to pick up the data. And I don’t really see that. changing very much. mean, in the food industry, they build new lines and new plants more often than, you know, people that work with metal or plastics or things like that. But yes, I don’t really see that changing very fast. It’s just so expensive, right, to be able to swap that equipment out and really, really upgrade it. I think maybe that’s the mistake. A lot of these other vendors, again, pointing at the big cloud hyperscalers are assuming that, you’re going to have modern equipment that we can use. know, MQTT and all these crazy new protocols to be able to talk to it versus maybe we could retrofit something or maybe there’s something, you know, on the line that we could take a look at that’s going to give us the information that you need. Maybe there’s one like new piece of equipment where you could get that info, even though the rest of the line might be from the seventies or eighties, right?

John Broadbent (07:56)

Well, yes and no. I did a job for Coca-Cola Amatil in those days now, Coca-Cola Euro Pacific Partners, they were soft soft owners. It’s an injection moulding facility that makes the PET preforms and caps for the bottoms. It is fully integrated. It’s an SAP site. The cost to make that smart, so a digital factory, was 1.5 % of the overall capital budget. It is so automated that orders from SAP are released to the factory floor. The orders are executed on the injection molding machines through the Husky, the vendor, Husky’s own manufacturing execution system. But we did all the integration between SAP and the factory floor. And the hub that we put in place also talks to everything else in the plant. So the whole factory is basically run by five people. So they avoided the labor issue because the labor component of the cost of goods is very, very small. And so they make for the whole of Australia, they export to New Zealand and they export to South into Southeast Asia. And that plant is running 24 seven, very high quality product and one of the showpieces in the whole global Coca Cola crown. We did a ready mill plant, they talk about food and beverage, did a ready mill plant in 2015. They started off at 100,000 mils a week, smart factory. They got up to 600,000 mils a week. They’ve just spent 90 million on expanding to the land next door that had space. They’ve installed another eight lines and I was there on Monday and they are now nudging a million meals a week. And the budget to make that factory smart was around about one and a half percent of overall capital cost. So when organisations think that this is a big ticket item, it doesn’t have to be. If you do it well and do it smart and integrate the factory to allow data to move around at real time or as close to real time as we can and you make good decisions based on that. you, mean, the guy there, the head of operations or the head of manufacturing said to me that when they started back in 2015, their variance on 60 million cost of goods, a million dollar P &L, waste, sorry, variance to plan in terms of usage of materials was about $120,000 a month negative variance. They now have doubled the plant. They now have gone from a hundred thousand mills to nearly a million mills and they’re now losing $20,000 a month. And he said, that’s insignificant now in the cost of everything that they do. Because we record everything, we weigh everything, we measure everything, we report on everything and they’re able to use the data they’re collecting from their process to do continuous improvement. And that’s really where it boils down to. If you can’t measure. what you’re doing and you don’t understand the cause and effect relationships, what are the lead measures in your business that will affect your leg measures like profit and loss and you don’t know what they are in real time. You’re effectively driving your car with the windows blacked out, no dashboard, the view in the rear view mirror where you were a month ago and the managing director in the passenger seat asking, we there yet? Yeah, sadly that’s the state.

Bryan Sapot (11:15)

Yeah. But what about, so I mean, like these examples for the most part are like Greenfield or Greenfield ish, you know, like the ready meal place that they’re expanding next door, right? Those kinds of things. Like what, let’s say you’re a, you know, 50 year old manufacturing company, you’re growing, you know, maybe 10 % a year, but you don’t really have the business to build a new building, that kind of stuff. Like what can you do to get there to achieve these benefits without, if you don’t have the cash to. build out something new.

John Broadbent (11:47)

So one example mentioned in that ready mill plant was that their giveaway on their lines now sits less than half a percent. So let’s use the example of check weighing, which if you’ve got food and beverage customers listening to your podcast, I’m sure there are people who understand that a check weighing is a piece of equipment that sits at the end of the line as product comes across it. The check weighing knows what it’s weighing, it’s a dynamic weighing. So the check weighing has to be calibrated for each product. So let’s say we’re making a 500 gram jar of coffee. The check-way was calibrated to know that the jar weighs a particular amount when it’s empty. And so it’s only weighing the weight of the ingredient that’s in the coffee jar. Most organisations that have these check-wayers have them at the end of the line prior to packaging to make sure that the weights are compliant. And as the product comes across those, the check-way will display locally what the weight of that coffee jar is. No one seems to look at that because no one seems to care. The check-way effectively does reject. So if there’s a jar that’s underweight or a jar that’s overweight, which is unusual, it’s mostly underweights, it’ll sweep it off the line and they’ll count those at the of the shift and go, you know, we had a one, two, three percent reject rate. But those weights are often held inside the check way and no one really goes and does any analysis on them. So how do you know unless someone goes and takes a look and often takes down a USB stick, sticks it on the check way, if it has that capability, can extract the data into a flat file. that can then take it away, do a spreadsheet and work out what their yield was, what perhaps their giveaway was for some average weights. That’s industry 3.0 or even 2.5 I’d call it. If we could connect that check-way to a network and we can abstract that information from the check-way to somewhere else, to a control system, a SCADA system, where we can collect that information and store it in real time, We have now taken the first step of the Industry 4.0, which has four maturity phases. And that first phase is being able to see information away from the source of that information. So imagine we created a little screen somewhere and on that screen, all we were showing was the pack weight of every copy jar that went over that check way. We could then take that weight and we could store that into a database. So every single jar, which is stored in the check wire anyway, usually, but every single jar has now been collected and stored into a database. Step one, step two then says, well, okay, I’ve now weighing these jars and I’ve got 495 grams and 505 grams and 510 grams, but what am I supposed to be? So step two is to then. find somewhere in your business where the standard weight for that product exists and that’s usually in an ERP system like an SAP or an Oracle or whatever it is. So you get the order from that system for example and you understand that the standard weight is supposed to be 500 grams. So now you have a frame of reference, you have context to see if the jar is overweight or underweight away from the check-weigher. That’s step two of industry 4.0, that’s the understanding piece to giving the data context so we turn it from data into information. If we then get really smart, we could now in our little SCADA system build an SPC chart, a statistical process control chart, and we can start trending the weights of those coffee jars. And we can now see where they are in terms of their standard deviation and mean average to where they need to be. And we can watch the natural variation of the production based on the filling of those coffee jars. And all production in that style has a natural variation. We have a bell curve. We have weights with underweights, we have standard deviations either side of that. But if we plot that on this PC chart, we can actually start to watch the trend. Then should, for example, we start to see weights in our sampling start to get heavy or light, because we’re starting to fill underweights, we know we’re going to hit an out of specification limit, say a low limit, where we’re now going to be making reject product. We can tell someone who cares, and they can go and make an adjustment to the filling machine. And now we’ve achieved level three of the industry for maturity, which is the predicting phase where we can take the information, we trust the information, we compare it to our standard and we plot it to show what it’s actually doing. The holy grail of industry four is the fourth step, because in the fourth step, if we now know that we’re about to make out a specification product, we could feed that back in a closed loop way to the filling system and automatically adjust the filler to adjust the weights. wait some time and test again and continue the SPC process. That step four is where we are then using the information that we’ve collected to now adapt and optimize our filling process. And that’s where the goal is. That’s where the significant amounts of money can be saved. And that’s where you can get down to example, half a percent giveaway. Because you’re now monitoring those weights in real time and the control system is making automatic decisions to ensure that you’re not over filling or under filling. And with the e-weights these days, what they call the T1, T2 e-weights, where you’re allowed to have a certain amount of product underweight in a thousand units, you can actually play close to that edge and you can even lower your giveaway even more. And then in that fourth step, you now have data that you can start feeding the machine learning and AI tools in order to then analyze that significant amount of data for you and perhaps do analysis of, giveaways or lines or shifts or crews against particular skews to see if there’s any patterns of, this product made on this line with this shift or this team makes a better product or has less giveaway, whatever it is. And then you can start to do some root cause and understand how to optimize it and more. So it doesn’t require significant amounts of investment to show that by doing a project such as that, you can add significant value to the bottom line. of the business because every dollar saved in manufacturing is a dollar added to the bottom line profit.

Bryan Sapot (18:05):

Yeah, it makes sense. You know what’s wild is I’ve seen plants where they do step forward, but they don’t look at the data and record it anywhere else. Like the machines automatically adjust, like if they’re underweight or overweight, like the checkweights communicate back, but they don’t do anything with their data. They don’t monitor the lines. So they have other issues, just not that one. It’s kind of interesting how segments is everything is like I think you did a really good job of explaining like these different effectively five the AI piece at the end different levels of maturity and people are in different lines within the same planner like all different places on that scale right it’s not like this linear thing

John Broadbent (18:52):

especially in Brownfield. So if you’re to buy check wire next week, why would you buy a cheap, perhaps something made in China that doesn’t give you open comms, connectivity, data extraction capability, and it’s a pretty rudimentary island of kit, why wouldn’t you spend a little bit more money and buy something like a Loma, which I’ve had a lot of experience with, that gives you everything that opens and shuts. And you can. And then start mining that and use that as a project to do as we talked before, Bryan, a proof of value rather than a proof of concept. I see organisations where they still get stuck in the proof of concept thinking, we have to work out that if we connect to this, we can take data and we can build an SPC chart. Really? We’ve been doing that for 30 years. How about we accept we can connect and extract and display and do something with it? But the question for the business is if we do that, what impact will it make financially? Because when you go to the CFO then to roll this out across multiple lines and multiple plants, the CFO wants to know what’s in it for him or her in terms of bottom line savings. So organisations that often try the proof of concept piece because it’s a technical question will hit a ceiling where the senior leadership team, often not technical, may not approve those projects. And I see many organisations get stuck in what they call pile up herd. But if you go back to the business at the start and say, hey, we believe there’s a significant amount of money to be saved here and you want to run a proof of value pilot, you’re far more likely to get the funding to do so and you’re far more likely to get the support and executive support that you need. Because then when you come back and show that you’ve, know, okay, you’ve reduced the giveaway from 5 % to half a percent, what’s that worth per year across that asset and across those products?

Bryan Sapot (20:47)

Yeah, it’s a good point. A lot of the time, though, folks don’t know what they don’t know, if that makes sense. I mean, I know it does, but you just don’t know how big the impact is until you start to look at something. We had a customer talking about the check weirs where they have 17 different lines. They all kind of do the same thing. They put stuff in a bag, right? And a couple of the lines had really terrible performance, terrible performance, great quality, terrible performance. And it turned out that the line lead always hit the exact target to the gram of what they were allowed to put in the back, right? So never allowed under, she never allowed over, it was perfect. So like they were pulling bags off that actually were fine, like within the tolerances, you know, that you’re allowed to do, like you talked about. And it’s like these crazy things where, well, our weights are good, right? Yeah, they may be, but like there’s this whole other thing going on that you might not even know about. You can’t even measure until you look at it, right? Which is always an interesting challenge because you’re selling this sometimes. You’re trying to convince people of an unknown, you know, like it’s there. Trust me. No, no, no, we’re perfect. You’re not like, I know. Have you ever run across that? Like how do you, how do you. And even internally, folks need to sell it that way too, because I may believe as an engineer or plant manager that I know I have these problems instinctively, but I don’t have the info to really be able to back them up. Especially if you’re at the, you know, we found people who don’t have these systems, they’re usually in the 60, 65 % OEE range. And to somebody who’s not in the plant every day, it looks like that place is running pretty well, right? But You know, you got another 20 or so percent to go realistically easy in that plant and trying to sell that up can be really hard when you just don’t have the info.

John Broadbent (22:41):

You’re very common and you need champions at the factory floor level who can tell the story well and sell upwards. And therein lies the challenge because as an engineer myself and many years in this role, in my early days I was a terrible salesperson. I was terrible at marketing my ideas. Thankfully, in my early days of engineering, I did have a young management team above me.

Bryan Sapot (22:42):

How do you

John Broadbent (23:09):

I think the ops manager was three or four years older than me and the general manager was only another 10 years above that. So I managed to get their ear and we did a lot of work in the early days in emergency appeal systems and distributed control systems. And we were spending a million dollars a year on electricity in one plant and we would spend four and a half thousand dollars on a chart recorder and a meter and a relay that we could set to turn off some of that process heaters it was called. and we spent $4,500 and saved 100 grand in the first year. It was those sorts of spectacular returns on investment because there was massively low hanging fruit. So don’t go after the hard stuff. Look at your P &L, look at your cost of goods, look at your power, water, gas, labour and raw material. I mean, that’s it. You take any process and I defy anyone to make it more simple than this. You take any process where you convert something to something else. And the inputs are power, gas, people and raw material. And the outputs are finished goods and waste. Now you can apply that at the front gate to the back gate, or you can apply that to an individual process. So look at your inputs and see where the big ticket items are. I’ve seen an organisation, for example, who was struggling because they wanted to take three or 400 people on their team. They wanted to take like two or three people out to save some labour cost.

Bryan Sapot (24:15):

Yeah.

John Broadbent (24:36):

I was like hang on guys, your PNL is 100ml, it’s 60ml cost of goods, you’ve got 15 % waste which is 9ml. How about we take a third of that and take it from 15 % to 10 % and save yourself five million bucks a year off the bottom line, which is way more than you’re ever going to get trying to cut heads. Why do you think that’s what it is? Well, they’re the numbers off your own P &L. And the challenge for most organizations, Bryan, is that whoever created the standards all those years ago can be unbelievably soft.

Bryan Sapot (25:10)

Sure.

John Broadbent (25:11)

And rather than an organization saying, well, let’s get tighter on our standards and let’s stretch ourselves to reduce that standard cost. And let’s look at the components of that standard cost. And okay, if we do that, we’re going to have negative variance for a while. So we get approval from the CFO and the CEO that we’re actually going to deliberately create negative variance by raising the standard. In other words, reducing the cost of the standard, which will hit us with negative variance to start with. But that’s what we need to claw back. And then when we get back to zero, we do it again. When we’re back to zero, we do it again. And it’s the plan, do, check, act cycle from good old Deming that most organizations that I see plan and do, but they don’t check and act.

Bryan Sapot (25:51)

Yeah, well, yeah. And the ERP world has made that process kind of hard, right? So one of the things that we have found, like if you have another system that can help you, like Mingo or something else, it’s Sketa system, doesn’t matter, but can help you actually figure out your actual cycle times, how much downtime you’re having, labor absorption, like reality.

John Broadbent (26:01)

Yes.

Bryan Sapot (26:17)

Right. Because ERP standards, they don’t change very often because they mess with the financials. Right. Just like you were saying, we could change the standards then CEO CFO gets involved. can have problems with bank covenants. Like there’s all kinds of issues. But if you can track this for real in another system and then look at it automatically once a year, once a quarter, whatever, and lower those targets, you know, so you’re doing things faster and less waste, all of that kind of stuff. And then you can really make a good case for, we really need to update this down in ERP system. It makes a lot of sense, but it’s damn near impossible to do, especially if you’re a company that makes hundreds or thousands of different products, you’re not going to have somebody out there with a stopwatch doing time studies on these things all the time. And you may not even know from the financial statements, like where are these variances and, and, know, waste and all that kind of stuff is coming from specifically, right. Depending on how you do your account.

John Broadbent (27:10)

Funny case in point, many years ago I was importing a little black box OEE engine. So it was literally a Windows CE embedded system software database HMI built into it. All you had to do was plug in a monitor and a couple of sensors and would do your availability performance and then with two sensors one further down the line you’d also get your quality component. Very simple, very rudimentary but really smart. We put this on trial in a dairy company for three weeks and collected data on a particular line they were having trouble with. So I go back and since it’s state, I’m in Sydney, they’re in Melbourne. I go back to Melbourne, I kept the data out of this unit and I then go and do a presentation and I sit with them and I show them that the OEE is 52 % and they want to get to 54%, no big stretch. And the guy called Gary, I his name. looks at the results and he goes, well, we know the OE is 52 and your systems come up with 52. So we don’t need your system because we already know where we’re at. I went, hang on, let’s have a look at the components of OE. And because we’re doing in-counts and out-counts, the quality component was the same. But there was a difference between his performance and our performance and his availability and our availability. And the numbers were actually swapped. And he’s scratching his head and he’s going, well, that’s a bit weird. And I said, yeah, I said, look at this. The trend shows that when the filler is filling, it’s doing 180 bottles a minute. said, and that’s the rated nameplate speed of the filler. And he goes, well, that’s interesting because we didn’t think it was running at rate. I said, so just to ask a question, said, how did you work out your actual production rate then? He said, well, what we did was we know that they have eight hours worth of runtime. We had an hour’s worth of downtime. So that gives us seven hours worth of runtime by the number of bottles that we made. That tells us our production rate. I said, but what happens if your downtime number’s wrong? I know that can’t be. We record it on paper. So I said, how about we have a look at the chart that we’ve collected from your machine? We called up this production chart and it was an absolute picket fence. Red green, red green, red green, red green. And he goes, oh my God, what’s that? I said, they’re called microstops. He said, well, I’ve never seen that before. I said, well, you won’t. I said, if your downtime recording is out, then your production rate that calculated is wrong. That’s why you haven’t been able to get your production rate. It’s not because the machine’s not running at rate. I can prove that it is. It’s because the downtime you’re recording is incorrect. So we go down to the machine, we go down and take a look. We go down to the machine and we suddenly see that these milk crates are coming through. being fed down a spiral gravity spiral chute. And when the milk crate wasn’t in situ, the gate allowing the bottles into the filling machine would stop the bottles going in. The filling machine would empty out, but it would continue to rotate. because it does, and it would exhaust the bottles that have now been filled and pack those and queue them in an accumulation space waiting for the milk crates to be there. And the crates were sticking in the spiral chute. And I kid you not, the operator was a female. She had all her white gear and safety gear and everything else on, and she reaches over behind her and she grabs a piece of four by two with a piece of carpet wrapped around the top of the thing. No, and that is the crap out of this spiral gravity. shoot and then the crate sort of rattles down and then gets back into place the gate opens up and the starts to fill and I just turned around and looked at this continuous improvement engineer and I went I think you found your problem and he was devastated that they’d spent a year on a continuous improvement project to find out why the filler wasn’t filling at rate because as you say they didn’t have the data

Bryan Sapot (31:05)

You know, that’s a common, I we sit here and laugh about the stick with a piece of carpet on it, but it’s really common because that woman, you know, was probably taught that’s what you do, right? And it’s just part of the job and nobody even brings it up. And they don’t like, like you said, it’s a microstop, so they don’t write it down. And when she sees the filler running low, she probably proactively walks over there and it wax the thing. And it’s just the way it is. We have a customer that makes battery canisters for very well-known battery companies, make billions of them. And they had a little situation where the canisters would get turned upside down and it wouldn’t get a coating sprayed inside. And for every one of those, the customer would charge them back. Like it’s fractions of a penny, but it doesn’t matter. When the percentages of those are high on billions of parts, the numbers get big. And so the machines would stop when they detected this backwards can. And it was the same thing. All the operators, you know, they have decent turnover at the operator level every couple of years. They get new people, they’re promoted, whatever. Nobody thinks about it. And they didn’t even realize like how many times it was happening. It happens hundreds of times a shift, nine hours a week. Like it’s crazy the amount of downtime that it added up to. Like their eyes, and these are really efficient. I mean, you can imagine making like four product numbers and you make billions of them. I mean, you’re really high OE numbers. And this was really the biggest inefficiency there and nobody knew. just because you’re not gonna track it, right? You’re just not. Nobody’s gonna write that down. 15 seconds, five seconds, just not. And it’s the, don’t know what you don’t know until you put the systems on.

John Broadbent (32:40):

And this product also had the capability to measure what was called break creep. So the shift pattern was put in that at 9.30, you 15 minutes of morning tea, and you’d see that at 9.20, the line would stop. 9.30, 9.45 to morning tea and at 9.55 the line would start again. So you’d have 10 minutes on each side of the break, for example, that was there. And when they started, you could also measure the under rate. So it wouldn’t start up at rate. It might take another three or four or five minutes to get up to break. When you analyze that over number of shifts per week, over every year, it’s a staggering amount of production loss. Or because they didn’t know because it’s not being measured.

Bryan Sapot (33:22):

Yeah. Yeah. Yeah.

John Broadbent (33:28):

So I see the whole IIOT, what I call edge enablement piece, getting equipment on the network. It may be that initially you don’t know what to do with it, but just to collect the data to understand what your process is doing. The ready mill plan, when we first built it, we put a process historian in that plan to record everything that opened and shut. A year later, I a phone call from one of the continuous improvement guys, he goes, I’ve got a real dilemma here, I can’t work out. We have some sterilizers, big horizontal tubes where they put ready meals in trays and they pressurize it and cook it at 90 degrees for a period of time to kill all the bugs and then off it goes out to market. He said, we’re capacity constrained. He said, but I can’t work out, we should be getting 13 to 15 cycles per day out of each retort, each sterilizer. And I went, okay, well, you know, you can go and mine the process historian and check the data. He goes, well, how do I do that? So I went to plant. sat with him, showed him how to use the client tool and the MIME this and we could see the pressure temperature cycle per retort. So I showed him how to export that data. Anyway, he comes back to me an hour later really excited. He said, I think I’ve found it. I said, what do you mean? He said, well, I can see that on a good run, back to back cycles have about a 15 to 17 minute gap, but on a bad run, can be up to an hour and a half. He said, I can’t work out why it’s going to take an hour and a half to load and unload a steriliser. I said, why don’t we go down and take a look? So we go down to the factory floor, there’s the operators, there’s some meals on trays, in baskets, on trucks, ready to roll into the steriliser. And Adam, guy, says, yeah, by the way, to the operator, look, we noticed here, we got this gap and that gap, and he goes, oh, do know the skew that you’re making at the time? goes, oh, yeah, we’re making this skew for this and this skew for this. guy goes, oh, yeah, yeah, we don’t have enough baskets. And Adam goes, what do you mean you don’t have basket? He said, well, with that skew, we’ve got enough baskets we can preload. So when we finish the last cycle, we empty and we just roll new trucks in. But with that skew, we don’t have enough trays and baskets. So we have to unpack and stack. And he said, that takes us a good hour and a half to do that. And this has run for 12 months and capacity constraint. And Adam’s scratching his head going, why did no one say anything? That would never have

Bryan Sapot (35:47)

job. Yeah.

John Broadbent (35:49)

just part of the job. So they spent 30 grand on additional trays and baskets and got 15 cycles a day out of two store. But as I said, 10, massive improvement in capacity. And I see this repeated. I mean, you’ve got war stories, I’ve got war stories. Everybody I talk to in the automation integration game has war stories. And I really feel for the customers and manufacturers who have no idea how much money I say has been left on the table every single day that with some judicial investment, some judicial measurement and a team that cares, a continuous improvement group that actually want to go and mine this data and look for that initially 10 % and then the 5 % and then the 2 % and then the 1%, they’ll find it. It’s absolutely guaranteed to be there.

Bryan Sapot (36:39)

Yeah, I agree. mean, the trick there is they’ve got to collect it and then you’ve got to ingrain it in the culture that we’re going to use this information on a daily basis in our regular meeting cadences to try to figure out how to improve these processes. And it’s there.

John Broadbent (36:52)

And if you start using that information to beat people over the head with as a big stick, you’re on the wrong side of the game. KPIs like this are not measured to, I believe, shouldn’t be measured to affect people’s performance or their bonus or whatever it is. We reward the wrong behavior because then we naturally assume that things want to cheat to get our bonus and we’ll do crazy things. It should be for the culture as a whole to excel and become a process, sorry, a culture of continuous improvement and use these tools to understand where the money’s been left on the table because organisations that do this well have amazing work cultures where people actually care about what they’re making and the way they’re making it.

Bryan Sapot (37:39)

Yeah, we have a big tier one automotive supplier for Subaru as a customer and they, I can’t take all credit for this, but it’s exactly what you’re talking about. Like they, they’re very strong general manager, like extremely strong team, but they had problems with turnover. They had problems with quality. They had problems with downtime performance, et cetera. And they started monitoring this stuff. And they ingrained it into that daily, weekly, monthly meeting cadence. They always looked at it, cross-functional teams in these meetings. They went from like mid sixties in OEE to 96%. And they got awards and stuff. I mean, it’s like crazy over the course of 18 months, two years, their employee turnover rate went down to like low single digits from, you know, like the mid thirteens or something. Like it was crazy. And it’s all because they changed the culture. Like, yes, the data is a little tiny part of it, but the culture change is the big part of it. And they foster this, what you’re talking about, that people care, right? Part of the team and we all, you know, trying to make it better. We’re not trying to beat each other up, yada, yada, yada. It works. It’s hard, but it works.

John Broadbent (38:48):

I guarantee the lady with the lump of 4×2 has probably raised that as a maintenance issue and no one’s cared enough. So the company overall suffers a whole year’s continuous improvement investment made on the wrong thing, on the wrong problem. So yeah, it’s wild. It is a staggers move and I see it over and over and over again. even the ability to do some rudimentary troubleshooting seems to be disappearing, critical thinking seems to be disappearing as well, unfortunately. these tools at least help you quantify and qualify what’s actually going on in the process. And without it, you’re just going be flying blind.

Bryan Sapot (39:31):

Yeah. And you you bring up a really good point throughout this whole discussion that it’s really not that much more expensive. Like when it really hit home for me was when you’re talking about the checkware and buying the cheap one that has no connectivity versus the nice one that does. I what’s the price difference there? Ten percent, maybe.

John Broadbent (39:51):

You know, even if even if it’s double, you’re going to end up with a piece of kit that’s an island that you’re not going to get information out of that’s going to break down that you can’t get access to the code, you know, blah, blah, blah. Even if it’s double the price, you want to sweat that asset for 10 years. Think about the cost per day incremental difference between the cheap one and the really good one.

Bryan Sapot (40:15):

Yeah.

John Broadbent (40:16):

And it’s dollars per day or cents per, pennies per day. know, when you actually extrapolate that over 10 year lifespan, there’s pennies per day. Like, why wouldn’t you? But you’ve got some, some person uneducated in procurement who now gets involved in the buying process and goes, no, you can’t buy that. I’ve done three quotes and this one from China’s, know, half the price and procurement then takes over. And then the poor bloody manufacturing people end up with kit that they just can’t maintain, can’t program, can’t get access to, can’t get data out of. and they’re excluded from the buying position, that’s completely and absolutely wrong. I actually have a 11 point equipment purchasing checklist for equipment like fillers and stuff for manufacturing. it’s, can you get access to the code? Is it in English? Is it a standard PLC like a Rockwell or Siemens or whatever, General Mise-a-Bizion, whatever your flavor is. and it just takes you through an 11 point checklist just to ask some basic questions before you buy a piece of kit.

Bryan Sapot (41:19):

It’s important. mean, those things are really important. Like, cause that there’s a lot of vendors that’ll hold you hostage, right?

John Broadbent (41:27):

It’s out of words.

Bryan Sapot (41:28):

they’re not gonna work on it. The guy that wrote it’s not there anymore and they won’t give you the program. like, you just made my million dollar machine worthless.

John Broadbent (41:39):

I battled with a global company who’d been on massive acquisition role. have offices in Scotland and the US and I battled them for 12 months to get the PLC code from a machine to get the password because it was our machine. It was our switchboard. It was our PLC and therefore the code in it was also ours. And they’re thinking, you know, we’ve got IP. You really think that I’m going to take the code out of your machine and go to market and try and sell your code to one of your competitors. I wouldn’t even know who they were. Like the fact that you’re that paranoid is a problem for me. And I will sign an NDA. I will sign a perpetual licensing. Like, well, you can find me a million bucks if this code ever gets out and gets to one of your competitors. But really, let’s just play adult here and have a mature conversation about why we need the code when you’re in Scotland.

Bryan Sapot (42:28):

Yeah, yeah, that makes sense.

John Broadbent (42:32):

I won the battle by the way, took me 12 points.

Bryan Sapot (42:35):

Good. Good. We battle that stuff all the time. We help customers get what they need. But yeah, I guess, know, to kind of summarize this thing, and maybe you can add on to this a little bit, is that you can get a lot of value for not spending a lot of money, right? And even, I like the fact that we talked about Greenfield versus Brownfield and how

John Broadbent (42:55):

You can.

Bryan Sapot (43:06)

You know, these things have been available for a long time, so they’re almost like sort of standard. You just have to ask for it, right?

John Broadbent (43:15)

Yes. And you need the education to know what to collect, why to collect, how to collect. I mean, my philosophy is we need to collect information and give it to the right people at the right time, in the right context, in the way that they need to consume it, by whatever device they want to consume it on. And when you do that, it’s game changing.

Bryan Sapot (43:42):

Yeah, it makes sense. Makes sense. Well, John, hopefully everyone will listen to us and do this.

John Broadbent (43:49):

I do the stuff. Catch up.

Bryan Sapot (43:51):

not to sell software, like forget the podcast and even the company, like I truly love manufacturing. it fires me up. I love walking in the plants. I love seeing how stuff is made. Frankly, I wish I owned a manufacturing company, but I just happen to be good at software. And I just want to make it better anyway, you know, any way that we can.

John Broadbent (44:11):

So. You look at countries that have manufacturing as a high proportion of the GDP. They are better societies. They’re more integrated. They have better standards of living. They seem to have a pride about the stuff that they can make stuff that then gets exported all over the world. I mean, in Australia, sadly, we dig shit up and we just send it overseas for the value add and we buy it back. You know, we’re notorious for having one of the best resource rich countries on the planet and yet we don’t do much with the resources that we mine. just sell them overseas. We have a very poor sovereign fund, not like Norway, and we buy that stuff back in valuated things like cars and other materials. just, yeah, and trying to get that changed here is a real battle.

Bryan Sapot (45:03):

Yeah, yeah, makes sense. All right, John, thank you. Yeah, it’s fun. Sounds like you’d battling 12 months for PLC program from a guy in Scotland. So, yeah. Well, thank you. Yeah, thanks for the time. And we’ll post John’s info as part of the show notes here in Australia. Want to reach out any help on digital transformation, right?

John Broadbent (45:26):

If anybody listening reaches out on LinkedIn, can find me, John S. Stephen, John S. Broadbent on LinkedIn and just reach out and say that you’re on the Mingo podcast and I’ll ship them the PDF with a pre-purchasing checklist, for example. That way they can go and buy good gear.

Bryan Sapot (45:43):

That’s great.

Bryan Sapot (45:50):

Yeah.

John Broadbent (45:50):

Thank you, appreciate that, take care.

Bryan Sapot (45:53):

Yeah, you too. Thanks.

John Broadbent (45:54):

Thanks, Bryan.