Aug. 18, 2021

Digital Lean

Digital Lean

Augmented reveals the stories behind the new era of industrial operations, where technology will restore the agility of frontline workers. 

In episode 36 of the podcast, the topic is: Digital Lean. Our guest is Edward Atkins, Head of Customer Success, Tulip

In this conversation, we talk about achieving customer success with classic Lean Operations and beyond. We explore performance transparency, how operators can produce parts tracking procedures digitally, creating quality gates through inspection applications and being conscious about quality measurement. What am I making? At what quality? We also go into more complex things like work material flow, Kanban loops and taking inventory and how no-code is an exponential uplift since you no longer are burdened by logging and can collect as much information as you want.

After listening to this episode, check out Tulip's and Edward Atkins' profile on social media:

Augmented is a podcast for industry leaders and operators, hosted by futurist Trond Arne Undheim, presented by, the frontline operations platform, and associated with, the industrial upskilling community launched at the World Economic Forum. Each episode dives deep into a contemporary topic of concern across the industry and airs at 9 am US Eastern Time every Wednesday. Augmented--the industry 4.0 podcast--industrial conversations that matter.

My takeaway is that implementing digital lean on a client site especially enabling improvements to be done near 100 percent by the clients themselves is now possible with the leading frontline operations software and is game changing for industrial performance. 

Thanks for listening. If you liked the show, subscribe at or in your preferred podcast player, and rate us with five stars. If you liked this episode, you might also like episode **22, Freedman's Factory: What is nocode?, episode 24, Emerging Interfaces for Human Augmentation, and episode 14, Bottom up and Deep Digitization of Operations

Augmented--upskilling the workforce for industry 4.0 frontline operations.


#36 Digital Lean_Eddy

[00:00:00] Trond Arne Undheim, host: [00:00:00] Augmented reveals to stories behind a new era of industrial operations, where technology will restore the agility of frontline workers. In episode 36 of the podcast, topic is digital lead. Our guest is Eddy  Atkins, head of customer success at Tulip. In this conversation, we talk about cheating customer success with classically in operations and beyond, and we explore performance transparency.

[00:00:28] How operators can produce parts, tracking procedures, digitally creating quality gates through inspection applications and being conscious about quality measurement. What am I making? And at what quality we also go into more complex things like material flow, Kanban loops and taking inventory and how no code is an exponential uplift.

[00:00:51] Since you no longer are burdened by logging and can collect as much information as. Augmented is a podcast for [00:01:00] industry leaders and operators hosted by futurist Trond Arne Undheim presented by, the frontline operations platform, and associated with MFG networks. The industrial upskilling community launched at the World Economic.

[00:01:16] Each episode dives, deep into a contemporary topic of concern across the industry and airs at 9:00 AM us Eastern time, every Wednesday, Augmented the industry 4.0 podcast. Industrial conversations that, 

[00:01:34] Eddy Atkins: [00:01:34] These  are the moments I live. I live for him. And then the wine, one of the reasons I'm two lipid.

[00:01:41] So I was on, so one of our customers that large. Chemicals customer doing? We've deployed to 13 sites. They're just like they're incrementally adding sites. I really like the way the team there and just the taking it step by step. So they're really like starting simple and going step by step. They've had success about three sites, 13 sites [00:02:00] globally adding the next three sites.

[00:02:02] We're on a I'm on an onboarding call. It's the second onboarding call that we've had with this site. So they've the first onboarding call they've done some basics into Tulip post, had a look at Sheila university. Now they're really starting to think about what use cases do they want to take forward and do.

[00:02:18] And one of the, one of the ladies. So she really wants to build this tracker for near misses. So safety near misses in the shop floor if they've had a, like an incident, they want to capture that so they can solve it before it becomes a real safety problem. And it's like a pretty standard, like a manufacturing.

[00:02:36] And the problem is the right word, but something that people really care about and we want to make sure people are safe in the park. And then someone else on the call had mentioned actually I'd seen another site within the same company that had the same use case. So we weren't really sure exactly what they built.

[00:02:52] So we were able to call the team there was able to call up the instance of the customer, the other site. And they're able to look at the app and [00:03:00] go, oh, look, this app does exactly the same as what you're looking for. And they were, we were able to download the app. We were able to get it on the lady's instance for her in her side.

[00:03:09] And she was able to start like playing with it and starting to put it into use for the next week. And I think like for them, they'd probably never knew what their site on the other side of the world was doing that definitely never been able to share like a best practice or a way of working that quickly.

[00:03:26]I think it just, yeah, it was like, for me, it was. This is a, these are the sort of moments that really gonna make this make the sharing possible that we've like it's like brings to reality that they're sharing is really possible service. 

[00:03:39] Trond Arne Undheim, host: [00:03:39] Eddy I'm curious, is this the first time that you, did you get a sense that this was something that this client experiences on a regular basis?

[00:03:48] Or is there something different about that? 

[00:03:52] Eddy Atkins: [00:03:52] I think it's like something they've said two things to me that they've been wanting to happen. And I think really there [00:04:00] about, so they're what now nearly a year into their journey. They started with one site then slowly expanded. I think this is the first time where they've really been able to like copy and paste something.

[00:04:16] Or something that's relatively complex, shall we say? So they've taken like pretty simple, here's a, here's an application for doing a very simple checklist or something else, but this was now really a, an app that had been like developed over here without anybody knowing about it really and have now been able to be taken and used over, hit over in another site on the other side of the world.

[00:04:40] So yeah the bit of re let's bottom up to bottom up sharing was the first time that had really been seen or achieved. I think we've done a lot, was like, let's build an app that everyone can use and deploy from the top. But really this bottom up sharing is, was pretty.

[00:04:54] It was definitely the arm  and the team was like, wow, this is cool. Let's do more of this. [00:05:00] 

[00:05:00] Trond Arne Undheim, host: [00:05:00] Give  me a sense of how this client got onboarded and you know this, or are the apps, how are they built with this client? What's the approach? Did they build it completely on their own or was it a guided development?

[00:05:11] Eddy Atkins: [00:05:11] Yeah, I  really like the approach this client's taking over. I think they're in a position where they can, that it's important to them, what they're doing with and it's giving them a lot, like a lot of value with relatively simple apps. So they're in a position where they can take things like step-by-step and they're really focusing on building simple things, making sure plants build simple things themselves first.

[00:05:35] So they learn the basics and then,  incrementing on the the complexity. So you have like example of how they're like they're taking this the first site that was onboarded built a what they call a cycle count app, which is like an inventory counting application. There's probably three data entry points and that they needed almost everybody on site to be using it.

[00:05:59] So they [00:06:00] took this simple app and rolled it out very quickly and they learned a huge amount from doing that. They're like,  they learned the basics of how to build an app in, but they also learn a lot about how to effectively get an app into production. And I think a lot of our customers maybe I think a lot of we all underestimate like that change management piece of how you actually get this new process that you've built, whether it's on digital or in paper, and actually go and put it into production and get three shifts using it seven days a week in like, All the 24 7 and just like really brick-making becoming part of the process.

[00:06:34] And I think more like what I like about the way they took it forward was really just a built, a simple thing that, that they're able to learn then how to get it in production. And then, what we've maybe how they've onboarded to so that's their approach. Then we've been working with the sites guiding them through this onboarding process.

[00:06:53] And typically the sites go through like three kind of phases. The push is to say, first of [00:07:00] all, get them trained, get them to build something really simple for themselves. And that's using all our online they utilize our university. So they have a policy where they basically say, you must go through these certain university courses before you get access to deploy stuff into production.

[00:07:16] So really like making the learning very central, the second part they have in their deployment processes. You take a really simple, what we call a template application, but they'd like our library applications as they have. They then have to prove they can get that into production and get that up and running.

[00:07:33] And then the third part is they then start working, I suppose it becomes guided from us to build more complex applications. And I, at that point they then building like a, this end, like someone somewhere is off building this end to end incident management app or whatever, their missus management app that we just talked about or somebody else's meant by building an app that monitors the levels of all of their tanks and brings that data into tulip and has some [00:08:00] alerting to the different, like different email groups based on the levels of those tanks and things like that. So it's cool to see taking them through,

[00:08:08]Trond Arne Undheim, host: [00:08:08] How long does it take to implement an app. Obviously it depends on, you should just completely fresh off the boat on the software, but in this case, like that app that we're talking about here, how can you come about? How quickly and how was it built?

[00:08:25]And what was the process? I'm just reminded because you gave me homework the other day. So I was just reading this and it's the journey to me is the book, which I understand to have been able to make it. Circuit when you were there, but it's interesting because the headline here is trying to lean  and making operational change stick.

[00:08:45] And, your story here just made me think of that. I was reading these descriptions of some factories in the UK, and it really is a human process. In addition, obviously, to having effective systems like no code [00:09:00] app software platform like Tulip .  What are some of the content on your context that you can give us to get started?

[00:09:07]Eddy Atkins: [00:09:07] I don't know this I'm like, I could talk about similar apps like this incident tracking one other than about cause it was done, but I can a similar kind of complexity apps where your the there's some it's.  So the. I think a simple app way or trying to like this title that you're trying to collect information from someone about an event.

[00:09:29] And then the event needs to follow like a process. And that process could be like reviewed and signed off it could be like reviewed, signed off implemented. So I think there's a, you have to have a bit of a vision for where you're getting to. But what I like this type of application I've seen others be successful with is they will literally build maybe that first of all, they build the data collection.

[00:09:53] So what is, what do they want to start understanding? So like very quickly you can build a simple form that says, what was the event? [00:10:00] What was it an almost in a week or two weeks, build that simple app and deploy it to different places across your facility and just start collecting these events.

[00:10:09] I think I also see it go very wrong the other way, where people try to design the perfect end to end  management system for these events. And they never, they can never agree on how they're going to do anything can never get past the first hurdle. So I think this is when I see customers being successful with us, they do really simple like from the simple phase and then iterate and add. And I think that's how, from what I heard, this is how that one was developed resource how they're doing a lot of the things that at the customer we're talking about. 

[00:10:37]Trond Arne Undheim, host: [00:10:37] I was just curious is given your experience, are you almost able to predict whether a customer culture is gonna lend itself to this kind of app development?

[00:10:48] Or if it's gonna take a little longer because their culture isn't ready for it, how did you describe a situation that makes these kinds of changes easier rather than [00:11:00] harder. 

[00:11:01] Eddy Atkins: [00:11:01] So actually I used to be good at this. I'd say after doing just purely and I got very good at okay, what are we going to be to capability gaps that we need to close?

[00:11:11] And I honestly, I don't think I'm that I'm not, I don't have that same, the thermometer that I used to have that kind of still figuring out exactly like how much, how fast can we say we can go and how what are the capabilities that really need to be built? And how do I assess whether someone has, or doesn't have those capabilities?

[00:11:30] And I think maybe that's a mix of being remote for the last year and not getting to sit next to these people and coach them. But also I think it is, there are still some like technical barriers, I suppose that need to, or that need to be overcome. There are still I don't, I think it's like the PE I think there's a simple kind of kind of something that makes things go quickly is always, if someone is familiar and can pick up it, basically, it doesn't mean they have to be a programmer or no [00:12:00] no, I know, the it background, but I definitely see there's still a resistance or no resistance has friction saying in the system is like the operators aren't used to using it. Or there's there's not someone that there's not a group two or three people that get it and get themselves into the system and learn it.

[00:12:19]So I think that's probably my first, like checkpoint is a bit like, how is the organization set up in terms of like being able to deploy, get something to deploy. I think from, and then the others, then the second side is okay, once we've been, go on, like what's I think the same, how are they going to be successful or not?

[00:12:38] And I think you really it's about prank, I think, what do I see being successful as like people that can be like pragmatic and sit like explicitly say, this is the problem we're trying to solve. And then understand that or another technology is a system to solve that problem versus, oh, wouldn't it be great if we had this wonderful solution and here's a really cool [00:13:00] technology, that's going to come and solve it for us.

[00:13:01] And I like, that's the kind of other end of the capability.  Mindset that we get and then we end up just not having a clear problem, this actually going to be solved or any like drive to solve that problem. And it just we just ended up on it. 

[00:13:14] Trond Arne Undheim, host: [00:13:14] I mentioned that actually last time when we spoke about this particular client, you said we're not doing much for them they run it themselves. I thought that was a pretty profound statement because you're also saying, they're very successful. So it's a bit of a paradox to me that a client can perception, at least I'll talk to them is that they're pretty happy with this solution, but on the other hand, they do a lot of it themselves.

[00:13:39]So how does that really work? Is it basically that they have ownership of the problem in some senses is a very crucial component. 

[00:13:49]Eddy Atkins: [00:13:49] Yeah. I would agree. So they don't, they own their problem. They're motivated to solve it and they like are asking I taking our guidance.

[00:13:58] Like here's a playbook [00:14:00] of how we've done it with other customers that works. And then they take that and go, okay, we'll do it like this, we'll change it a bit, but they could, if they take it on and they go, they put their hand up and they go, we've got this kind of problem. And we need to, we don't know how to solve this part of it.

[00:14:13]Can you help us solve that part of it, for example, So it's quite a collaborative with those, like the leadership team that it's quite collaborative way of working is it's really good perspective. 

[00:14:26] Trond Arne Undheim, host: [00:14:26] So it's early days for Tulip, but on the other hand, we're starting to scale some of the apps to larger numbers of users.

[00:14:34] Have you  started reflecting a little bit about how different it is, because I'm assuming you, when you were doing lean consulting, working hustle with very large deployments. 

[00:14:45] Eddy Atkins: [00:14:45] Yeah. I think like when we, so there's a bit of like where, there is definitely a, where am [00:15:00] I? So yeah, large deployments we're getting there with Tulip. I think people come from maybe a mindset like I can download an app on my iPhone. It needs to be that simple. So there's a duplex the bar is pretty high for customers when it comes to that level. And there's the way Tulip and I suppose other IOT platforms would be set up as there's a little bit more.

[00:15:22] There's a bit more red tape versus just being able to really share applications easily and that way, because you want to like do you want to be able to connect the data up and you don't want everything to just go in completely different directions. So I think there's a bit of it's not as fresh.

[00:15:41] It's not as easy as it's always illustrated or people think it might be, but it's I think also from the other side, we found that there was a lot, like if from a lean perspective, we really took things like very step-by-step. And proof like took a lot of time doing [00:16:00] building capabilities and working with people side-by-side like, it would accelerate. So we probably do one wave where it was one plant, which would take three months, we'd do another wave, which would be maybe three more plants and it would be another three months. And then you maybe do another three plants and then you could start to scale from there.

[00:16:17]I think we get like where there's some where we are now is somewhere in between those two. So we have a lot more of like online content, a lot more shareable content, the ability to work, go and see what work someone's doing by like logging into like sharing screen sharing as the cloud or seat and eye tracking what they're doing.

[00:16:37] So you have this ability to build capabilities much quicker for those more complex use cases. I think like more and more, I could see, like at least work, like whereTulip focused is like to be able to really get to that level where frictionless app sharing, where you really can share applications and and that will really help for, I think [00:17:00] from there.

[00:17:04] Trond Arne Undheim, host: [00:17:04] The question is I've been scratching my head quite a bit about that since I joined the company, just to try to understand how this actually does work and woodwork at a larger scale, because honestly, I do not understand how appealing it is, but when you have this feeling that you can really download it on your app.

[00:17:21] That your consumer  life, actually, it could almost happen at work. That is instantly very appealing but of course, like you said, there has to be some checks and balances. And I was more thinking about the issue that something that you in a team of five or 10, on the shop floor, you might think this is really efficient and we feel very good about.

[00:17:42] But similar to other need experiments where we have made small improvements, incremental improvements on one instance, there could be effects of this new process all across the factory. How how are we equipped to do that [00:18:00] correctly? And because I'm assuming this still is a little bit of a human in the loop.

[00:18:04] So you build one app without building another whether they're connected or not. That's a decision if they are convicted, maybe then, you start to serve, optimize or change your behavior, behave based on what the app is telling you, have you, what are some experiences we've had in that regard?

[00:18:21] Eddy Atkins: [00:18:21] So I think the, one of the biggest areas where I've seen as like light, like production line improvement or Owyhee, and that's a classic for. We have a machine monitoring solution, but actually it doesn't bring any value until you start building other solution. If you still so solving the problems, it tells you, you exist basically. You can act like that is, I think you connect to machine to Tulip  you start to see that changeovers are a big problem. You can then build an app for. A standard app for your change over process. And then you can use the data from that app to [00:19:00] identify improvements, to reduce parts of the change over, or do things in parallel or prepare better for the change over.

[00:19:06] So I think that's quite a standard the use case, we see it now, like proof of value concepts. Another really interesting one. The other day that I, another customer that I'm working with, she put Tulip into their, what they call their reverse DC, which is a small team that works in a, like a customer's distribution center.

[00:19:27] And they're managing all the warranty returns from that customer to them basically. And they're checking, using Tulip to go through a process to check whether it's like a valid return or not. And that's that gatekeeping is saving them like warranty, significant amount warranty cost. But on the other side, they're now collecting a lot of debt and I could nice debit of data, which is saying, what is the what's the value.

[00:19:51] What is being returned and why is it being returned? And they get like a photo of it and all this like nice data. The net of that, a step we're working on now is pushing [00:20:00] that, okay, where do they quick can they use that data? They have Salesforce and they have in Salesforce, they have all the information on there, like customers who signed up for for it.

[00:20:08] So they can make a link between this warranty return and the customer and why the customer's returned it. And that can potentially, they can spot customers that are doing more returns or they can stop a customer claiming for something they've returned before. But also they could start pushing it back into they're going to push it back into their product database and stuff like seeing what is being returned.

[00:20:29] Why is it being returned and use that and tie that then into their product development process. We felt this very simple for app for them that they're using or built with them, I should say. And we're now building those connections like the app is running, but we're now building out those connections for them. So that data starts getting pushed back to those other systems.

[00:20:49] Trond Arne Undheim, host: [00:20:49] The only thing I wanted you to talk about, which I found fascinating is when you, if you are a leader in an organization that's considering making obviously [00:21:00] investing in lean processes and you're sitting there trying to figure out what are the steps in the process.

[00:21:06] I know you have a pretty clear idea about what their priorities should be. Would you just walk that through with me in terms of what the decision processes for committing to a lean changeover? 

[00:21:23]Eddy Atkins: [00:21:23] I would sorry,Trond it  came to my mind. I'll come back to that. Ask that question again, but I'd like to ask answer a bit about lean, talk about leading checklists and digitizing those that have an interesting story from this week.

[00:21:34]Trond Arne Undheim, host: [00:21:34] So Eddie, tell me a little bit about lean checklists, 

[00:21:44]Eddy Atkins: [00:21:44] Or you should also talk to Mark Fredman about this as well, cause I'm sure he has a good opinion, but a classic thing to do in lean is I think we talked about this last week. I don't know, last time, but is to put a checklist in place to standardizing  it's just found it becomes a [00:22:00] standard There's QC a very generic 5s checklists, 25 questions.

[00:22:07]Is I really standard thing, like big organizations tried to push everywhere and it, it creates a nice measurable, like how well is an area doing within the plan. Someone goes and does it every month. It creates a nice mentality of people going on to the shop floor and going to see what's going on and coming up with ideas to how to improve it.

[00:22:25] So it creates some really good lean, like behaviors that you want to look at. I think something that we've been like thinking about is like that type of audit is very paper-based and yeah, you can put that audit into  an app and you do the same process, but all the data or you no longer have to the admin of the paper-based system that you have this kind of same system in, in digital.

[00:22:47]I think what I'm trying to do with small customers is actually take one step further with that and say, what is what is a digital version of this look like? What does a digital kind of are there checklists look like [00:23:00] it's actually like the root of what you're trying to do with the checklist is one make this maturity level so that you can create a bit of competition.

[00:23:08] You're trying to create this behavior of someone going on the shop floor and talking to it's like talking to an operator, like generating problems or issues or identifying improvements. And you're trying to You're trying to spot deviations from the standard and reinforce the standard. So they, one of the ideas that that mark and I have discussed a couple of times is like this this like almost like a Tinder for five S basically.

[00:23:34] So you like you do, you go on the game is like someone goes around and takes photos of what stuff should look like. You go on the shop floor and you get these photos and you just say it  it like this or not. And if it's not like this, you have to go and come up with an idea how to improve it and like this, instead of like people going through a checklist, this is now the kind of, you've got this randomized list of photos and you have to go and find them on the shop floor, for example, or you have to go and ask [00:24:00] someone like, where is this photo and they can't tell you. And you can say, is it in that condition or is it not in that condition? And then with that, you're starting to, you could. It's like more fun. You're using the digital technology and you're starting to, and you can do the same scoring cause you get like the photos get randomized, like how many deviations per area that have been.

[00:24:19] So in the background, the system can do all of the, are these checks happening or are they. Are they happening are they not happening or how what's, how many are getting deviate? What areas are getting percentages of deviation? So you can do all the KPIs you want to do. But you start to use the technology in a different way.

[00:24:36] And I think we're, I definitely see where my customer, my customers are not, we're not there at that level of like creativity yet. And still wanting to put their standard lean checklists on paper on glasses we talk about. And I think that might, that mindset is something I think we all can, we all need to do a better job of coaching and challenging, like our customers about trying to do [00:25:00] something different.

[00:25:00] It's very easy to say yes to the paper on glass checklist and she does it. 

[00:25:04] Trond Arne Undheim, host: [00:25:04] It's a great point because I think, if there's going to be transformational value out of these digital investments, you just do  stuff, 5% better or 10% better because you're not doing the admin around the paper.

[00:25:17]We're trying to do something completely new ideally or vertically change around the process, obviously never responsible way. But yeah, this brings me actually to another question I had for you.What are the main  use cases in like a one-on-one deployment of digital lean tactics. So you talked about checklists that would seem to me to be one. But can you just list a few of the, like the top five or top 10 that maybe they all do come from the paperworld it doesn't matter what me, but just give me a sensce what the typical first 10 apps look like for it, for a customer. 

[00:25:59] Eddy Atkins: [00:25:59] So the first, like the [00:26:00] first app for me is always I, we call it, I can call it performance transparency, basically.

[00:26:05] And that's so if we take a cell, it's like what is important for, what is that cell doing on a daily basis? And that app might simply be like, just a count of production versus target for that cell. So we start with something really simple. There could be like an AOE monitoring application. It could be a, like just a, an operator, like some tracking of an operator producing the parts.

[00:26:31] So I think that's, for me, it's like first thing I want to understand it so what performance is. We can then add, we had a few additional apps and maybe that's within that cell, if it's one operator or maybe it's a as an adjacent cell is some kind of quality gate, sorry, a second inspection application or quality measurement.

[00:26:52] I start to understand what am I making and what quality am I making? And that's the first couple of apps that always deploy [00:27:00] from there we can make that data can start to be put as a hourly tracking board on the production floor, like a date, a shiftly review board that can be used in shift meeting.

[00:27:10] And then a like can be cascaded into a performance board for the sign, for example. So you start to getting this core performance data. You're starting to have ways of looking at this performance data. So the next part is things are going to come up there's going to be a quality issue. There's going to be something that the operator sees can be improved. And there's going to be things that come out of these performance meetings, like actions that need to be followed up on and tracked. So next thing I put in would be like a action tracking or events manager is what we call it which sort of connects all these events that are coming like deviations or ideas that coming out of all this discussions about the data and you can start tracking them into allocating them to people, tracking them to completion. And that starts like the improvement that starts to close out, like the improvement loop that you want to do.

[00:27:56]I think then some other app, other applications. [00:28:00] I would start looking at, then there's like the SAP's that underneath what's happening in the cell. So you're actually starting to give the operator some information about what they need to be doing to produce putting a standard in place for how they produce things.

[00:28:14] You can start then based on what you're seeing in the quality data, you can start connecting error proofing into that SAP as well, like connecting to IOT devices and things like this. So you might want to start collecting the measurement. I don't know, you might want to start tracking the operator has done the right talking for for a particular process, for example. And then I started looking at more complex things like work material flow. Maybe we have a, I work use a Kanban application for creating Kanban loops and doing material and an inventory flow in a plant, for example. From there, we start building it into something that's more of a, sort of an end to end material management for the site.

[00:28:56] Trond Arne Undheim, host: [00:28:56] So this way you have, when we [00:29:00] built it up from scratch, but of course they are somewhat mirroring existing processes. What about this process compared to a more traditional enterprise process or even just the way that things were working before, like if you go into a brownfields factory setting, you don't have to Tulip..

[00:29:20]How, what does this look like then? Let's say a new manager comes on board and starts to look at these things. Let's just say that they don't have a,Tulip type system. Let's say that they're in their head they're thinking either just manual improvements or they're thinking that's putting a massive enterprise software legacy system in there.

[00:29:42] What are those two situations and how are they different from what you were just describing with this sort of like bottom-up.

[00:29:49] Eddy Atkins: [00:29:49] I suppose like the manual system for me is very very similar to the no-code process, but I have I'm burdened by what I can do, but so to throw a few [00:30:00] things, I'm like if I putting a, I might want to measure a process I'm burdened, I could put a piece of paper or a whiteboard that someone will measure that like we'll use to measure the process or look their activity.

[00:30:12] And I then I then have to make sure someone is filling it in every hour. I have to make sure if we filled it in, we put it into an Excel and then we use that Excel run some analysis. So there's a change management piece that's needed that I would say is still there with the digital app, but you have to get people to use it, but it's more can be lower.

[00:30:31] And and there's a piece of adding this additional administration burden. So you usually as if a process engineer you usually limit yourself to, or a manager usually limit yourself to how much paper you put down or whiteboards you put up or the complexity of what you ask people to record, because you know that like the changes you've got to get people to do it, and you've got to like actually analyze it.

[00:30:52] So now we have within the no code environment, you can F you can really like, you want to do stuff quickly and get stuff. But you can [00:31:00] collect as much information as you want really and you can do once you've done an analysis, once you can then like, it's an infinite, every week it's refreshing or every hour it's refreshing, you can see like updating.

[00:31:10] So you get a, like a huge exponential uplift in your capacity, basically. Based on the numbers, on the date, having everything digital. So I think that's that's the difference between this like very manual based process. And they could the coming from the other side, if I said, okay, now I want I want to take this traditional and I want to put an MES in pro in place.

[00:31:33]It's a very there's a lot of kind of mapping takes you map everything. Of of the process that you're doing, you measure it to go and measure the times you go and upload those to the system you  write down all that material, like the, what, the flows of the work orders, and you upload those to the system.

[00:31:51] This might be like yield calculations that you need to develop and  upload those rules to the system. So there's a very, like you're spending a long time [00:32:00] building the rules of a very big system for this manufacturing process that maybe doesn't always follow all those rules and wants to change quickly as well.

[00:32:08]And then at some point of time, that information gets comes once it's been implemented, it comes back to the shop floor, which It is I think back to your point earlier, Trond where you said you were making a decision, like you're deploying an app, but then you're making a decision whether you connect it to someone or whether it needs to be connected to something at all.

[00:32:25]And yeah, that, so like part that, that way of being able to just deploy or use like a use case or an app and it allow it to be standalone or it doesn't matter. But also then being able to make the decision to connect it to something. And like often, this is a deployment approach that we often take is like a customer wants us to connect to their ERP system, to push and pull data.

[00:32:46] And usually what we do is actually have. I put that have a wreck and I use a mock of that data or a replica of that data in a, to that table to start off with, to actually be able to quickly [00:33:00] get something working that fits for the operator before we then actually build the connection, the real time connection back to the ERP system with the engineering team.

[00:33:09] So it's versus how you would previously do it you've mapped out all the connections you want and then you would build them. And then the last thing that gets deployed would be the app to the operator.  

[00:33:18] Trond Arne Undheim, host: [00:33:18] I'm quite curious about this difference between Tulip  and an MES system. Also given that we were just classified as an EMEA system, so there are certainly overlapping aspects but how traditional MES systems then work with other type of software? Are they're not all all encompassing they also work with ERP systems unless they are in the ERP system. I'm just trying to get a much better understanding of how an NDS system works and how they have evolved over time.

[00:33:49]You give me a sense of, you worked with them sometimes.  Give me a better sense of what it feels like to work with an MES. What is that process like? Because I want to really [00:34:00] contrast it with what you are now using. 

[00:34:02]Eddy Atkins: [00:34:02] Yeah, so I think that a few fold, the, how do they communicate with other systems and so typically pretty like historic MES is, have been an on-premise software, so they will sit on premise and they'll talk so maybe an on-premise ERP and norm, like they would replicate data. So , there would be a once a day that would be a replication of like important data from the ERP into MES and then the MES, which use that through the day.

[00:34:35] And then that would be updated and where, and some have moved on to  be   more than that, but that's just a sort of example so like how to Tulip works is. There's we'll create real time connections with these real time updates to these the cloud-based versions of these ERP systems.

[00:34:55] And when data is needed, it will ping the ERP and bring you the latest [00:35:00] data back or when in the in the application, when the operator enters some data, it will push that data over a connector directly to the ERP system. We've got this kind of differentiation between system of record. Whether they live somewhere, the data stores and that could be like in the ERP. Some of that can be in Tulip  it could be in another database could be in like multiple places to the system of engagement, which is like the apps that you've built for the operator. And like in the past, they were always the same system. So the MES you'd always go onto the MES screen if you then needed to do something in SAP or the ERP, you then go into the ERP screen. So it was as an operator or even a supervisor or a manager you're working, there's like multiple systems to find all your information. 

[00:35:48] Trond Arne Undheim, host: [00:35:48] Wasn't a I, my guess is, what, if you are a vendor of an MES, the traditional assumption was, if you're having some success, your goal is to get grow and [00:36:00] eventually be more and more, a system of record like you pointed out, but is that everyone be the reality on the shop floor, even for a Tulip  system, that there is one system of record that sort of handles all information, or I would rather just moving into a day and age with interoperability and connectors and stuff like that, and somewhat seamless interface.

[00:36:25]Eddy Atkins: [00:36:25] Definitely. So I think the system of record is definitely is not going to be the same system. Like I think we can, we're always going to have a a historian on the shop floor that like can collect real time series, data really efficiently. And then the ERP that has the that has the, like the financial data that lives in it and then,a lab database that has all the lab tests in it, for example. So I think the system of record definitely is going to be different. It's like courses for the courses, for the data basically. I think the system of engagement side, [00:37:00] like from, I think that there is an aspiration and I think it's feasible in many cases to actually say that could be one system or look and see what, like one system to the operator.

[00:37:13] So I have a custom like many, multiple customers that are really selecting tulip because. It will be their one system of engagement. Definitely not there one system of record it may be in some cases like for some pieces of data and then connecting to let, to SAP and all these other systems and making that a really making an effort to bring all the data to the operator in an easy to understand way.

[00:37:39]I've got other customers that are taking, I think, which is really interesting taking the view of where we can build. This functionality and Tulip really quickly, but we still need to have this data coming from somewhere else or these, it can mean machine controls being done in a, like directly through the controller PLC.

[00:37:58] But what we're going to do [00:38:00] is like either tulip is going to take some of the real estate on the HMI for the for like most of the apps. And then we've got this other, like machine control part of the HMI and they're bringing it all on to, so the operator's perspective, it looks like the same interface basically. But they are seeing that they're needing to have kind of those additional levels of control. But I think, I think that's with Tulip that functionality to day, and I'm sure the The direction will be more like that. I don't know it doesn't matter whether it's one system or not, but I think from an operator perspective, you want to have, you want to have like fit easy to use interface that gives you the information when you need it.

[00:38:39] Not the kind of super complex Boeing 7 47 flight cockpit, that a new if you're a new operator you take years to get trained 

[00:38:48] on.

[00:38:51]Trond Arne Undheim, host: [00:38:51] That's interesting, right? Training times, because it's more standard sort of old school system, you have [00:39:00] massive amounts of courses and then those courses take hours and hours.  What is your let's go into some of the training aspects kodak, any industrial milk. What does it take and how do we typically approach it when a client comes and says,  we want to deploy on the training side? Is it a question of first like training, getting onboarded to the leadership or do you rather go straight into a use case and then try to train everyone and how physically are they trained on our system? 

[00:39:31]Eddy Atkins: [00:39:31] If it's a  user using the application, so an operator or inspector, the aim is really for them to like zero hours of training. Like we try, I try to push my customers to build stuff that's really intuitive. And with the operator can pick it up and it just makes sense. That's not always the case. They do need sometimes a bit of training and a bit of bringing on board. But I think it's you don't get trained in how to use Google maps on your phone.

[00:39:58] It just works. So that's the kind of [00:40:00] objective for that. And from a process engineer perspective, that's going to be building the applications. I think there's  is definitely some training and you need some training in how to like the training is not just like how to code and how to build a functioning app.

[00:40:16] It's also coaching them how to build something that is going to be easy to pick up from an operator and works well. And I think there's there's often a tendency with us, like the engineers who are usually very intelligent and want to stick by the nature often very detailed oriented. They don't build the most, like user-friendly all, that's always build the most user-friendly interfaces for the operators. So there's definitely some coaching that we need the coaching we often do at that side of it. How does it work as like a, an overactive doing something for real is when people really  give open access to the tulip university where people can go on and just play around and build applications. And usually more often than not customers give them a, create themselves, a [00:41:00] sandbox environment where anyone can just go in and build everything. But with a site, with an India, a group of individuals that want to deploy a site, I usually take them through we usually take them through a kind of an actual use case over a number of weeks.

[00:41:13] It could be over a week if they want to do something really intense or it could be over a day or half a day, a week over, over a month or something where we've worked together on a use case and really build it together and get it out on the shop floor. And I think that's the way people really learn from that very well.

[00:41:31] Trond Arne Undheim, host: [00:41:31] You have just listened to episode 36 of the Augmented podcast with hosts Trond Arne Undheim. The  topic was Digital Lean our guest was Eddy Atkins, head of customer success at Tulip. In this conversation, we talked about achieving customer success with classic lean operations and beyond, and be explore performance transparency.

[00:41:51] I know operators can produce parts tracking procedures digitally. My takeaway is that [00:42:00] implementing digital  lean on a client site, especially enabling improvements to be done near 100% by the clients themselfs. Is not possible with the leading frontline operations software and is game changing for industrial performance.

[00:42:13] Thanks for listening. If you liked the show, subscribe at Augmented or in your preferred podcast player and rate us with five stars. If you liked this episode, you might also like episode 22 Friedman's factory. What is no-code episode 20 for emerging interfaces for human Augmented. And episode 14, bottom up and deep digitization of operations.

[00:42:38] Augmented upskilling the workforce for industry 4.0 frontline operations.


Eddy Atkins

Customer Success Manager, Tulip

Eddy focus on delivering sustainable manufacturing improvement programs using the latest Industry 4.0 and digital tools as well as the tried and tested lean techniques. He worked for a wide range of national and international clients transforming manufacturing operations.

Client engagements have ranged from one-day diagnostic visits identifying improvement potential through to planning and executing sequenced waves of focused operations transformations.