Overcoming the Impossible: DataOps at Poclain Hydraulics with Rija Rakotoarisoa

David (00:01)
Welcome to a new episode of the ITOT Insider Podcast in which we will be talking to the group IT operations and industry 4.0 lead at Poclain Hydraulics. But first, two things. First of all, this is our very first sponsored podcast. So thank you Litmus for sponsoring this one and also to make sure that we can really keep on growing our channel, our reach, which means bringing more interesting content to you. And secondly, if you are still interested in joining our ITOT Academy,

We have five seats left in our January cohort, but we also already opened our May and September groups. Our January run is almost full. We tried to limit it at 40 participants, but for May and September, there are still plenty of seats available. So if you’re interested, please go to itot.academy. And that’s it from my side. So ⁓ sit back, relax, and enjoy a new episode of the ITOT Insider Podcast.

Speaker 2 (00:58)
Welcome, you’re listening to the ITOT Insider podcast. I’m your host David. Subscribe to our blog or go to our Academy to know more about our work. ⁓ And today we are going to talk about industrial data ops being ⁓ in the ITOT domain from the perspective of an end user. And I’m super happy that I have Rija Rakotoarisoa ⁓ if I pronounced that well, ⁓ with me. ⁓ Rija is the…

or used to be the IT manager of a large French manufacturer active in the automotive industry and today he is the Group IT Operations and Industry 4.0 Lead at Poclain Hydraulics, which is in Industrial Machinery. ⁓ Rija thank you for joining us. ⁓ First of all, I hope I pronounced all the names correctly and if so, can you introduce yourself?

Speaker 1 (01:48)
Okay, it’s very good. Hi David and thank you for having me. So I am Rija Rakotoarisoa So basically IT operations and the industry 4.0 manager for the Poclain group. So the scope of this is all of the manufacturing project applications, but also the supply chain side for all of the groups.

We are supporting all of the Poclain plants ⁓ on the all operations action manufacturing and supply chain.

Speaker 2 (02:22)
Could you start with positioning Poclain? What type of manufacturing, what type of products, just to get a of a rough idea?

Speaker 1 (02:30)
Okay, so Poclain is an digital company. we are producing motors, pumps and the valve. So basically we are one of the leaders on the hydro transmission

Speaker 2 (02:45)
Interesting. What does an industry 4.0 lead does on a daily basis in this company? What is your role?

Speaker 1 (02:55)
So my role in this company is to support the operational, let’s say, simplifying the usage of the data in the shop floor to address the real pain points like quality, like maintenance, yeah, even in data collection to, let’s say, to make, to take decision easily. And the other point as well is like,

Previously, we had a lot of manual actions and you don’t have the data on a daily basis. You have it the next day if it was written somewhere in a paper or Excel or something. And data consolidation is as well painful earlier. So our journey, part of our journey is like automate data collection from the equipment, from where we need it in term of performance, quality.

and on and so on and then ⁓ provides ⁓ automatic data flow to where the users need it. So the idea is to support them to concentrate to their real job. For example, quality should focus on quality but not on gathering data and cleaning data. So we try to as much as possible to work them, prepare data flow.

push it in the global database and from there we create automatic and live data and helping them as well, for example, for the ⁓ improving quality by providing some live indicators, but live Pocahontas as well. For example, producing one part on a business unit and the next step of the process when you’d like to use the same part.

by scanning this so we can see if the part was okay or not okay in the previous one so we can reduce the customer claim as well on this so that’s some real life example.

Speaker 2 (05:04)
I have a ton of questions about that, before we do that, I think originally you come from the IT sites. now you are in, call it industry 4.0 or ITOT or at least I would say in a manufacturing context. So I would love to understand like what was that change for you? Like when did you change from an IT role to an

ITOT or manufacturing or whatever role and how did that went to you? What were the differences? You know a bit about your personal journey from IT into ITOT.

Speaker 1 (05:43)
Yes, that’s a really good question. So basically I have a master of science in IT first. you’re right. So I started more than 20 years back in the industry as the IT developer first. working, starting working at the industry for a given plant as an IT developers. And then I moved to IT manager for this plant. So managing.

developments, local application, network, ERP and so on. So after five, six years, I felt like I already see everything. I thought, but it was not really the case after analysis, but at that time, 2010, so I would like to do something else in top of the IT.

So I follow up a master in finance at the party. Yeah, master in finance. then because this was the only way, let’s say to ⁓ remove the tag as a geek because it was complicated to move to another position and the same company because you are tagged as that. This is the IT guy, it is the geek. if you’d like to.

Speaker 2 (07:09)
For the record there is nothing wrong with geeks.

Speaker 1 (07:12)
Okay, no problem. And yeah, and so I asked my manager at that time that, yeah, I’d like to do something else. And I would like to do more business in terms of the IT having more impact in the business because in fact, during my life in the plants, I always in contact with people in production, in finance, in separation and so on and so on.

I was aware about all of the requirements from the plant side. So I would like to provide more values. And I do this master in finance. And after this, I stay at the same group, but move to another plant as a finance and IT manager. So I have both positions. And ⁓ it was very challenging because…

was really let’s say the level zero in finance to be honest you know but you already have the position to manage this for for all of the plants and we had as well a plan that ⁓ we lost money for this so one of the big challenges let’s say to say that we should put this plant in positive not only me but with the whole team but at least I use my background in IT to put in Plastum Automation to support the business team to have

all of the information that they need. we called it at that time, each people, each manager, each individual, each single people should know the cost of their doing. we wanted that we would like to make them responsible of their doing. So providing an automatic reports, automatic analysis.

I already had the knowledge of the ERP for the group at that time. it’s more or less. So my background in IT was a big help for me, even if I was not, let’s say, superstar for a company.

Speaker 2 (09:22)
just creating a very big Excel spreadsheet.

Speaker 1 (09:25)
Exactly. Yeah. So it’s collecting information from the RP pushing to the right people at each level, at each level of our business unit. And I did it from 2012 to 2016. And from this, I worked closely with the operational people, how to improve, to reduce cost, to improve the operational excellence on the…

It give me, let’s say, really, it’s really a big game changer for me because I have another eyes, let’s say, another eyes. It’s not just a focus on IT, but focus on business and focus on the, okay, we can reduce cost, but we should produce with the right quality. We should ⁓ take care as well about how the operator are working, the working condition in the shop floor and facilitating and so on and so on.

And we started to see that, yeah, we have a lot of things that are not automated. We have to enter everything on paper because I was also the manager of the ERP for the plants and the team that entered manually the information. You ERP is part of my team. said like, yeah, we need maybe to automate something here because a lot of time, two, three dedicated people to enter information only. There is no value added for this.

And I started to work with the, let’s say it’s like a shadow. It’s not a group initiative, but in our local plants, started working with the maintenance team and collecting information on this. yeah, so after two, three years as the finance and the IT manager, so our plants, at the first year, our plant was positive. So that is the real good things.

The years after, ⁓ a lot of work was done by all of the teams, the local steering committee team. yeah, so the plant were going well and there is less challenges. see. it’s like, okay, you need to prepare the business plan. You had to check, you to generate the P &L, and so on and so on. ⁓

it’s less exciting as a job. So I am like working, let’s say, having new challenges. So I asked, so it’s in 2016, so I asked like, okay, finance is good, but finally,

I would like to do something that’s to digitize the manufacturing. This is one of the things I would like to have. So I asked there as well to see if ⁓ there is a position or a project for me. And at that time, the company, okay, so they create a new position for me to work this time on the global group ⁓ to work on these topics. And at the same time,

We were bought as well by a big company, by Alain Ammar, and I was part as well of the global group as well for the 3.4.0 project. And it is where the transition. So work with them during three years, years. Yeah, learning a lot. at that time we had 56 plants.

across the world and like 25 in Europe. yeah, so seeing various ⁓ examples in various plants, in various processes and so on and so on. and yeah, experiencing as well some good and bad practices. Yeah, but at least I learned a lot on this. And ⁓ after three years of this position, I moved to to Ponglien.

It’s like for the again for the industry for the zero, it’s a green field. no point is a creation position as well for me because the company at that time decided to yes, to put someone on this position to support the team and the operations. So with my team at that time, so we were two guys.

to start from nothing. ⁓

Speaker 2 (14:06)
I will come back to that but you said one thing you mentioned just a minute ago you saw good examples and bad examples. I’m sure our listeners would be really interested like what would be a bad example what would be a good example?

Speaker 1 (14:23)
Yeah, an example of bad example is like, for example, yeah, for a company and it’s normal that the company has a let’s say a global strategy. So we have to roll out globally on the group. So you decide it, sometimes you take information from the single plant or from the end user themselves and sometimes not. ⁓

But sometimes you say, this is the strategy, you should implement this. And you don’t ask the real problem at the plant and you push it. So it’s a real top-down. And at the end, you take time, you put money on this, and at the end, it doesn’t meet the real needs. And then you lose money, you lose time, and nobody use it. This is one of the…

One of the battles.

Speaker 2 (15:22)
this really top down. I want these five dashboards with these seven KPIs and I don’t know. That’s it.

Speaker 1 (15:30)
Exactly. Exactly. So in terms of a big company, it makes sense. It makes sense. But in order to sustain some changes, think it’s very important to take into account as well the real business needs. And I am told, I am telling this, the bottom up. So it is for me, one of the good practice.

to say that, okay, maybe we cannot say yes to all because it’s not possible. It’s not the same one, but at least we can highlight, you can highlight that this is the real issue. This can bring value for the end user. And then when you are working or addressing, let’s say the real problem in the manufacturing side or in the operation side, this will change the real life.

the daily basis life of the operational team and then they will support you and they will use it and they will apply it on a daily basis. It is what I am saying, ⁓ good practice, but it’s not simple. It asks a lot of change management, a lot of animation and so on and so on, a lot of network as well.

because you have to have your ambassador at each plant or at each business unit and so on and so

Speaker 2 (17:05)
Yeah, And so and then you say, okay, you made the move to Poclet and then you said from an industry 4.0 perspective, it was a greenfield. So what should I imagine with a digital greenfield?

Speaker 1 (17:14)
Exactly

So basically,

the companies, think it’s among or for most of the companies, the companies were strict. But ⁓ from the IT side, we have a strong footprint on the IT for IT, let’s So on the other things, ⁓ for example, to managing the network and so on, that is very good at the company. The global system and the company is very good as well.

But for the manufacturing, there was no dedicated people to supporting plants. there were a lot of ITOT projects that was managed only by operational people, by OT people. And for example, put in place some CMMS, for example, directly managed by the operational people for maintenance, for example. And then

Speaker 2 (18:06)
I have.

Speaker 1 (18:20)
there are potentially some gap that is not filled, that is to the IT constraints. So, and that is one. And after this, there as well, for example, a good practice, let’s say for you should make something more scalable, taking technology, have to take into account the updates. Or for example, you’d like to take into account how can we have some,

some global configuration. if you let the end user to say like, okay, I’d like to implement this. don’t know. I don’t want to tell they don’t care because they don’t know. They care the, let’s say the final use, but in between.

Speaker 2 (19:03)
No, no, definitely care.

Speaker 1 (19:13)
They don’t know, they don’t have no idea about what is the impact in terms of technique, in terms of maintenance, in terms of everything. Cybersecurity. Cybersecurity, standardization, and so on and so on. And if you implement something that is not standard for all, that means that if you have 10 plants, you have to manage 10 various versions of the same application.

Speaker 2 (19:20)
Security.

Speaker 1 (19:37)
at the end, you have a big nightmare everywhere. But of course, the motivation was there because they have to have something to work and to perform the task. So the good idea had by Poclain at that time is like, OK.

it is no more possible to continue like this. So we have to structure this putting some dedicated people working on this to standardize this and it is where myself and my colleague at that time, welcome to Poclain.

Speaker 2 (20:15)
I can imagine that it’s like the is this saying like how do you eat an elephant piece by piece or bite by bite. I can imagine, Rija if you like if you start it’s in the beginning it’s this it’s an elephant right and like where did you start?

Speaker 1 (20:36)
Yeah, so ⁓ it was really complicated because even if so I have some memories on that when we started starting discussing with people in the shop floor here in France. So they said that from when the IT come to the shop floor, that was the first question of the guys and they are not really happy because I said they have a

They had, let’s say a bad idea of IT.

Speaker 2 (21:11)
I’ve

never heard of it. Never, never, never heard about that.

Speaker 1 (21:17)
Yeah, and yeah, so we started, so we have some small project, let’s say, that we need to start with. For example, we had some quality application that normally this should be one application, but each plant have his own version. So in term of code management, in term of configuration, it’s something that is really not… ⁓

standardized. it is one of the things that we start with. we know that there are a lot to do. So we started with what is the most pain point for the group. And this one is typically critical for us because it’s the end of line quality control. So all of the product produced by the company

should be checked with this application. And we started with this one, standardized this, so moved from exact application to web-based application and centralized one. So first is everyone to align with this. ⁓ Same thing for the, for example, for the work constriction. So we have as well a very old application on this. We start as well with this one.

And this has a real impact on the operational team because they are using on a daily basis each time, for example, because we have a lot of diversity. And for example, in a given station, we can have 50, 100 various part number that should be processed at the same station. So you can do it by your head. you should have your constriction in front of you to say that, OK, this is the parameter. ⁓

parameter tolerance is not the same for this part for this part and then so on and so on. we should have something that is more dynamic. Those as part of the initial project that we put in place. And we had as well a big project to put in place an MES across the group. But this was canceled after the COVID-19 because no more budget.

COVID-19. at that time I said that, normally it’s part of my main mission when I come to the company. But since there is no more projects, said that, OK, what we can do? So go elsewhere ⁓ to start a new project or continue here. But ⁓ how can we do more with less? That is the real point. And how can we, let’s say, use the existing system

with less expenses. So we changed from the strategy to have a big system having everything by using the best of breed. So it was the strategy we chose and we did not have the choice as well because we did not have the bid rate to put place in a…

a big thing and I think I will say this that yeah, I know that it’s not a happy event for everyone, but I thank the COVID-19.

Speaker 2 (24:52)
You made a very good point here and of course there is a lot of very very bad things about COVID but the thing here is that because you were under budget constraints you had to rethink your approach.

Speaker 1 (25:05)
Exactly.

Exactly. So this was really a good thing for us. So it’s like a fictation. do what you have. And this is what we did. So we try, OK. We already have some siloed applications, siloed system that generate data that people take a USB key connected to the system, ⁓ put it to the computer, creating a

a cross analysis with Excel and so on and so on. And after one hour or two hours later, yes, we have a dashboard, but during this one, two hours, we did not do your real job of quality or of quality. You’re just generating data. And this is when I did some travels across the group plants. So this is one of the things that we need to work on. ⁓

I, in 2021st, so I started the Data Hub project. ⁓ and it is where we have various system. So I don’t want to name all of the partners we evaluated, but at the end, so we try to have some ⁓ educated, to take some educated decision about what we need in term of connectivity, what you need in term of compliance or.

with the database in terms of how to use it. It should be easy. So one of the key points for me is like something should be usable for non-IT people. Because I don’t want my team to be the bottleneck of the system. So those are some things that we did and Litmus.io was part of the candidates on the system on this.

And at the end, they check almost all what you need. And since 2021, we implement this system as our main data hub. And thanks to this, ⁓ we break the silence on the system. Step by step, we build on it.

Speaker 2 (27:18)
So you made a decision to have Litmus.io on different locations. One thing which I find really interesting is that from a technological perspective, I can imagine it ticks the boxes, but from an organizational slash data governance perspective, you can’t just say like, I’m gonna roll out.

Speaker 1 (27:21)
Exactly.

Speaker 2 (27:42)
edge devices or whatever you’re doing. I’m gonna roll out some historians and just start. assume you also needed to think a bit about how am I going around data governance? Like what data do I need? How I’m going to ⁓ structure it? How I’m going to store it, et cetera, et cetera. Any stories about that?

Speaker 1 (28:05)
Yeah, so ⁓ it’s a very good point because I am ⁓ involved as well in the ecosystem in France, internal industry for the manufacturing. So last month in October, I gave some conferences in Paris and exchanges with some experts from various companies. And ⁓ yeah, it’s really the pain point for all. So where to start and how to start and how to… ⁓

how to take people with us and take the journey with us. So it’s very, very important point. So for us, ⁓ took a lot of time to explain to our top management as well that yes, when you put in place a data hub, it will just an enabler. It is not magical, it’s just an enabler. So yes, it will have some costs.

it will cost at least for the beginning, but the idea is to implement some use cases in top of this. So that is what you explain. And we were lucky that the top management at that time understood this and spread as well the information in all the group because yes, when you have to pay something in the immediate timing, you have more or less nothing.

So at the beginning, it’s not easy. But what we did as well is like, we tried to find some quick win, something that has a meaning, meaningful. It’s meaningful for any people from each plant. So for example, we know that we have a test benches at each plant. We know that data from test benches are already available somewhere, but it is not collected automatically. There is no automatic dashboard on this.

It is one of the quick queues that we say, okay, just a plug and play the information on this, put in place data transformation to collect this one, create this globally, and we can duplicate easily because we have more less the thing with the MemSocial. This is one of the, let’s say, the first use case that helped a lot to say to say, there is something to do. People started to… ⁓

to give some interest on the project that yes, we can do something. And then we did as well some exchanges for each plant having a weekly workshop, a weekly meeting with them to understand, okay, what are your main point points? And we addressed the real ⁓ problem. So we did not try to digitize, let’s say something that they did not ⁓ master, but really.

Speaker 2 (30:54)
Yeah, yeah.

Speaker 1 (30:56)
address some real pain points. So this is how we start, not easy at the beginning. And we try to, let’s say, to push some use case first in the beginning. And then after this, ⁓ when each plant now that it can help us to do this and this, they give us the things. And we have a lot of backlog of the request from the…

Speaker 2 (31:23)
That’s really good. Do you, like from a data modeling point of view, do you centrally own some kind of a, yeah, do you centrally own the master data? Is it decentral or are the plants so different that it’s a plant by plant model anyways?

Speaker 1 (31:42)
So we keep mastering the data model from my team. We master data model, we create the model and we duplicate it. So for example, we have various application. We cannot name all of them, but for each application, we define some data model. And once it’s done, it can be duplicated for each data hub at each plant. And then you just, ⁓

they can manage it by themselves because we try to find some, let’s say, ⁓ scalable configuration, a template with parameters. So they just need to change the template, like which equipment, what information, blah, blah, the mapping and so on that is done by themselves. But the heart of the system is managed and maintained by my team.

Speaker 2 (32:31)
Okay, so that means that you have the template management is somehow done globally, but the actual translation to local use cases, that’s done locally.

Speaker 1 (32:45)
Exactly. So basically, for example, we have a motor line, for example, you have this bench for a motor line. Test bench motor lines more or less same thing for any countries, ⁓ for any continent. So we create once and then we duplicate it in their data hub so they can use it. And at the beginning, maybe it can take us a few days to develop the template, but for a new equipment, once, for example, the IP is done, the federal configuration is done and so on and so on.

it’s a matter of half an hour to connect it, to reuse it. And then they are using as well the same dashboard for the other plant. And so we try to have something that where we can have, let’s say, a dynamic selection from which plant for each business need, for each part, and so forth, et cetera.

Speaker 2 (33:34)
Yeah, Okay. Do you centrally store all that master data or is it in spreadsheets or is it in a system?

Speaker 1 (33:49)
So everything is inside the database, a global database. Yeah, and inside the global database. So we have the global database for all data, but we push as well some information to where we need it. Basically, we have some cloud application, for example, for reporting. So we collect information once, and we push it many where we need it. So that will avoid any manual inputs for a…

for any user. for example, for our software, we are using a software like Dynamics Fabric. So the software is in the SaaS application. So from our data hub, we make an API call to push information there. So more or less 90 % of the indicators in this is pushed automatically by our data hub.

Speaker 2 (34:39)
Nice. ⁓

Speaker 1 (34:40)
And

it’s a huge improvement for all of the plants, they don’t need to take care about formatting data, entering information. Of course, there are some actions, some manual action to be done by the operators or by some managers, but 90 % at least, it’s pushed automatically there.

Speaker 2 (35:01)
Well, you know, we’re getting close to the end of the podcast, but I want to do, I don’t do that too often, but I want to do a crystal ball moment here. Like, suppose I would say in your role or in your company or whatever, let’s say that we are, I don’t know, not too, not a decade, like five years in the future. Like what do you hope will be different or maybe what stays the same in five years?

Speaker 1 (35:25)
Yeah.

in five years. ⁓ I’d like to, our company will be a full digital twin. So everything will be there. ⁓ The predictive maintenance is just normal.

Speaker 2 (35:45)
It’s like duh.

Speaker 1 (35:48)
It’s just normal things and the manage directly by the users. In less than five years, I hope that we’ll have some self-service data. any business unit can take information, use it, run it, make automatic analysis, maybe ⁓ using AI or any sort of thing. yeah, the things is like connected.

data from manufacturing, from supply chain as well to improve and accelerate the business excellences.

Speaker 2 (36:23)
Nice, Hey, and before we close, and this is a totally different topic, but I saw that you actually wrote a book as well. ⁓ It’s a French one, it’s titled FRANCHIR L’IMPOSSIBLE, which I guess translates more or less to overcoming the impossible.

Speaker 1 (36:34)
It’s late.

Exactly.

Impossible exactly, yes.

Speaker 2 (36:46)
Nice,

so what made you write the book and what is it all about?

Speaker 1 (36:54)
So basically in 2020 during the COVID-19, so the world stopped. It’s from our company as well. ⁓ we stopped to work as well. So we are really off, let’s say. And it’s the first time ⁓ from a long time, from my school to 2020. It’s the first time that I do nothing during one month.

And it’s really strange. And, and, uh, yeah. So I asked questions. Okay. So what, so what’s next? What, what can I do during this? And by, uh, by seeing as well, a lot of the bad news on the TV and so on and so that, yeah. So, uh, it can be me as well. So I said, like, I have a special, uh, path, let’s say, because I am come from Madagascar.

I grew up there and just come to France after my baccalaureate and continue my university school here. yeah, from my poor, so my, from a modest family. So my path is not really easy. So I cannot imagine where I am today if I think back about my initial, ⁓ where I come from.

I said to my wife, yeah, I would like to write something on paper for my two kids, my two boys. If, I don’t know if the COVID take me, so at least they have something. So I started writing some memories like this. And ⁓ after two, three days, I had a lot of things to share. And after this, said, ⁓ maybe I can turn this in term of what is the problem.

how can I overcome this? What is the result and what is an unexpected decision? I create, let’s say, next chapter for its story in the book. And I discussed with another teacher I had at the Pari Dauphin University. So he said, yeah, it can be interesting to write a book on this. And this is why I started to write the book. And I did it in three weeks.

Speaker 2 (39:16)
Wow. ⁓ I think just the title, I’m going to say it in English again, but overcoming the impossible. ⁓ I really like the title. ⁓ But, I think it also is perfectly applicable to many problems we face in our professional life. ⁓ And also bridging the ITOT gap is something which is possible.

Speaker 1 (39:40)
Exactly,

yeah.

Speaker 2 (39:42)
All right. All right. ⁓ Thank you for joining us. ⁓ This is a a wrap for this episode of the IT/OT insider podcast. Also a big thank you to Litmus for sponsoring this episode. ⁓ Yeah, thank you again. I really hope I would say I wish you all the best. It’s a super, super interesting journey. And I’m sure you’re going to do ⁓ lots of very, very good stuff for your company. Thank you.

Speaker 1 (40:10)
Thank you very much. Thank you, David, for having me. See you.

Speaker 2 (40:13)
See you and to our listeners, if you enjoyed the conversation, don’t forget to subscribe at itotinsider.com, discover our trainings on itot.academy and see you next time for more insights on how to bridge IT & OT, it is possible. Until then, take care, bye bye.