"Smart Metals Podcast," hosted by Luke van Enkhuizen and Denis Gontcharov, offers a clear and practical look into the metals industry's journey through digital transformation, Industry 4.0, and the integration of the Unified Namespace. Listeners can expect in-depth discussions that break down these complex topics into understandable segments, actionable insights, and real-world applications. Luke and Denis bring their expertise to the table, guiding you through the evolving digital landscape with advice on leveraging technology for streamlined operations. Each episode aims to empower metal industry professionals with the knowledge needed to confidently adopt digital innovations and understand the impact of the Unified Namespace in creating a more connected and efficient production environment. Join us to navigate the future of the metals industry with clarity and confidence.
Denis: Hi, and welcome to
the Smart Metals podcast.
I'm your host Denis Gontcharov and
Luke: my name is Luke Van Enkhuizen.
Denis: this week we have a very
special guest, Wim Dijkgraaf
from Quotation Factory.
Wim, welcome to the show.
Wim: Thank you.
Thank you for inviting me.
Denis: So I think we'll begin
this episode by giving a short
introduction of Wim and Wim's company.
So go ahead, Wim, why don't
you talk a bit about yourself?
Wim: Yeah, well I'm Wim Dijkgraaf, founder
and the current CEO of Quotation Factory.
If you look at what I've done
in the past, I've always been
involved in change management,
ICT system implementations, etc.
A lot of done a lot in Brazil related
to healthcare and Bluetooth devices
and large population health management.
And what that was kind of my first
interest in the event driven architectures
and obtaining information from the
field, from people at long distance.
And by coincidence, I got
involved in metalworking I think
more than 10 years ago, started
quotation factory five years ago.
And what we actually try to accomplish
is to have state of the art technology
for metalworking businesses especially
in the field of high mix, low volume.
So we are not so much focused on large
series, but high mix low volume and
be able to become a supplier that
you can trust that they are able to do
what they promise and, and Important
part of that is being able to estimate
correctly the production time, the
lead times and the material usage.
So that's where we are focusing on
right now, with quotation factory.
Denis: Right, very good.
Can you delve a bit deeper into
the problem that Quotation Factory
attempts to solve for people who
are less familiar with the world of
high mix, low volume manufacturing?
Wim: Yeah, of course.
So high mix, low volume means in practice
that you on a daily basis, receive
requests for quotations or inquiries
that you have to analyze first to
understand what the customer wants.
So you have to analyze what it is that
he asks for, what are the features
in the product, what is written in
all kinds of engineering drawings.
For example, related to specific
production techniques or qualities to
be requested, or even, for example,
surface treatment that is necessary.
So, on a daily basis, they have to
analyze and be able to understand
what the customer asks for, then to
determine if they are the right supplier.
For that request and if they
are who else should be involved.
So which type of work should
be outsourced, et cetera.
And then of course, how
to manufacture that.
And then if you know how to manufacture
it, then you have to come up with a
realistic timelines and production
times and material usage, estimate that.
And try to respond as rapidly as
possible with the quotation in
order to have a high success rate.
Denis: Right, and this process
is, I'm assuming, mostly done
manually in the past and even today.
Wim: Yes, often we see current
metalworking companies doing
it manually, meaning of course
excel sheets are used a lot.
But for example, if you go a step further
in the end, if you have machines that work
on a CNC basis, so you can program the
machines that you use, that means that
you often have CAM systems to create.
Those programs and the CAM
systems are able to simulate
, that manufacturing step.
And a simulation means that also some
time estimation is done under the hood.
So often they integrate that
in their quotation process.
So it's a mixture of manual
work Excel sheets, CAM systems.
And then if it becomes a production order,
Often they have to manually do a lot of
yeah, data entry in the ERP system because
that's the central place where they
pick up the further production planning.
Denis: So from what I'm hearing is I
have already heard of Excel sheets,
MES systems, probably ERP systems.
Is it fair to say that the data
required to steer this process is very
spread out across multiple sources?
And if this is true, in how far is
this a challenge for Quotation Factory?
Wim: I think that's true.
So many systems are involved all of course
with their own specific functionality or
specialties and and knowledge and still
a lot of users are involved because you
had one user has how do you say that?
Is specialized in that specific
manufacturing step and programming
the cam system, et cetera.
So a lot of roles, a lot of manual data
entry and a lot of data dispersed all
over the place on hard disks and in
individual systems, absolutely correct.
Denis: Luke, this sounds
pretty familiar, doesn't it?
Luke: Yes, absolutely.
This is indeed the industry that
I focus on, the metal fabrication.
And that's how I also know Wim.
It's, it's always a challenge to,
to find the data at the right place
and respond fast to requests and
get the order ready for production.
A challenge.
I always, I find to to get to the
actual data that you need to, to
make a good quote, to make a good
price, to make a good project.
Wim: Absolutely.
And it's especially what you actually do
is you determine how to manufacture it.
Then you'd start using all the systems
involved and use the outcome of those
systems in order to create a quote.
So two processes happen at the same time.
You look at a sales process as well as
a work preparation and planning process.
And often that's also
organized in different teams.
So there's also a lot of
team communication going on.
Denis: Just for my understanding, how is
the final price of an order determined?
Does it depend on the lead time or
does the customer always will get
the same price depending on the part?
,
Wim: At contract manufacturing, then often
it's determined based on a cost price
and then with an additional surcharge
and discount, that's more or less it,
but it's mainly cost price driven.
In this market segment instead
of , for example, value based pricing,
that is what we encounter way less.
Luke: Yeah.
And even if it's value based, I would
say it's just the surcharge is determined
on other factors such as the urgency.
So basically there are,
I think, two key flows.
Like one is Version where the price
is determined up front and that is
also the price that you will pay
when the product is delivered.
And then there is the flow of a post
calculation method, and then in between,
there is all kinds of things like
staggered prices of fixed prices of
contracts agreements for prices for parts
or even entire assemblies and products.
So I think it's important just to
introduce the listener here a little
bit of where we are in the spectrum of
the metals industry, because what we're
talking about here is a manufacturer
that takes a sheet or a tube or any
raw material and transforms that
into a semi finished or end products.
So for a manufacturer, for example,
of a car, this would be them, the
metal components that are used to
mount a wheel, for example, or this
could be a component of a train.
This could be the frame of a,
of a new transportation vehicle
or a metal end piece that is
used in an industrial kitchen.
And so this is traditionally called
discrete manufacturing and the subset
of that is called a job shop environment
to be specific in the terminology.
So a job shop company works by job to job.
They don't have a fixed schedule.
They don't have a fixed set of articles.
Everything is almost custom for a client.
And this is a thing where.
You really find this
high mixed low volume.
There are companies that are doing
basically they have a fixed set of
articles and they are combining this
in the manufacturing room, but I don't
change the article inventory that often.
This is also happening, but you
see everywhere in the industry
that this is being changed and
favored by a more custom production.
The customer wants everything custom these
nowadays ,series are getting smaller.
And that's, therefore you have to
be more adaptive to those needs.
So just to clarify a little bit of,
from my experience of what we're
talking about here and what makes this
such a unique part of the business.
Wim: Yeah.
Great addition.
That's that's exactly the context
that we are talking about right now.
Yeah.
Maybe the only thing I can add is
that it, because you never know
the exact details of a product, you
need the data from the past in order
to be able to estimate correctly.
Luke: Yeah, that's huge, right?
Cause you need to make an estimation
of how long something will take because
you haven't, you haven't made that exact
product with the exact same circumstances
before, you cannot say this batch
will be the same as previous batch.
And it must be the same
as a previous batch.
You cannot really tell.
Cause it's a lot of manual labor as
well involved in many fabricators.
Like welding, for example,
and like even bending parts
is still often done manually.
Like, so this makes it even harder.
So yeah, it's a, it's a big challenge,
I think,, to get this right from before
you ever made it in the past, and I think
this is where I think if we talked in
a previous episode, Denis, about tribal
knowledge, like a lot of companies have
somewhat of an Excel sheet where they
write down like how these things are
happening and how long is the lead time?
Those kinds of things.
They have a feeling about it.
They say, okay, for example, if I do sheet
cutting and bending, that takes me about
three days, but it's just a guesstimation.
It's not actually something that
they really know in my experience.
Denis: And I assume that by decreasing
the volumes and increasing the complexity,
At some point, these manual process
will no longer work, will be less
optimal in the future, which is why we
want to go to a more digital process.
But then we face the
challenges we highlighted here.
We raised the concept of an
event driven architecture
earlier in the interview, Wim.
Can you explain, this architecture,
, how it could potentially address the
issues we raised about the data being
spread out across multiple sources?
Wim: Well, first of all, in the
introduction, it already became
clear that these kinds of companies
work with a lot of systems and
that it all has to be integrated.
So first of all, you have a
systems integration issue.
What you often see still in this
industry is that they use the, not
the old fashioned, but the traditional
ISO 95 model as a starting point.
The ISA 95 model shows the ERP system
at the top of the pyramid with the PFCs
down and then in the middle of somewhere
you have the manufacturing execution
system and the SCADA system, et cetera.
But they have the traditional way of
looking at systems and integration where
the ERP system is the central point
of departure or the single source of
truth, if you want to use that word.
That means that all systems have
some kind of an integration.
So for example, the cam systems or
the time registration system, or
whatever had the scanners, the barcode
scanners, et cetera, all have a point to
point connection with the ERP system.
That's the traditional way to look
at how these companies approach the
digitalization, how would you call it?
Strategy.
Denis: So what is the
challenge to the ERP?
It's a pretty high level system.
And we talk about PLC data.
It's usually a very large volume.
Are modern ERP systems designed
to process that amounts of data?
Wim: First of all I think not.
So there's a big mismatch.
I call that an impedance mismatch between
the values that you record or measure
on the shop floor and the abstraction
level that you find in an ERP system.
So that's one.
Another one is that those integrations in
itself are kind of complicated to manage
from a project management perspective,
because all kinds of parties are involved.
And if you create point to point
connections, then both parties should
understand the world on the other side.
And that creates a lot of complexity
and the risk that it takes longer
and it's more expensive to realize
these kinds of integrations.
So from a systems integrations
perspective, that's I think the
first really added value of applying
a more event driven architecture.
Denis: Can you describe in a bit
more detail how this event driven
architecture would look like?
If it's not the ERP, then what would
be the central component of this
system or most of the data will reside?
Wim: Yeah, so the term data
is an interesting terminology.
You have a lot of definitions about what
is data, what is information, etc, etc.
A lot of metalworking companies
say, but I have a lot of data
because it's all in my ERP.
I think the essence of becoming a
smart factory or an with a focus
on, for example, industry 4.
0 aspects, then the critical type
of information that you have to
focus on obtaining it and storing
it is, is what we call an event.
So what you try to do is to be able
to observe everything that happens in
your company that can be on the shop
floor or that can be contact points
with a customer supplier in terms
of document interchange, et cetera.
But those moments in time that
an relevant event happens.
You want to be able to, to observe the
fact that the event happens and you want
to be able to store that event, the moment
in time when the event happened and the
necessary data related to that event.
And that's actually when we say we
need, a good data strategy, I think
that that's what we are talking about.
Mainly the events.
Luke: If I can comment on that one
indeed, because commonly indeed, I think
the, the way that companies proceeds
with their digital transformation is
that they reach out to their vendor
of their, for example, ERP system.
And then those providers will then
suggest to expand that solution with
additional modules perhaps upgraded
or even move it to the cloud.
And then which then should result
in some new APIs, and perhaps
then you can interact with your
software through those connectors.
And then you could basically
subtract information from it.
Now, of course, that promise makes sense
from a, I think, from a perspective
to just do some basic analytics on
the current situation on the past.
But I think at the moment you want to use
this data for something more than that.
For example, you want to run
large computations on this, like
check the history over, over time.
And very importantly, I think that
to make the factory more interactive,
that I think it's important to mention
here that if something happens in the
factory, an event occurs, that you can
create this if then then this structure.
Like, if you see something happening
in a different system, what can
that mean for another system?
And that you're more flexible in defining,
you know the various connections between
all the systems you have in your factory.
So I do understand where most
vendors come from in a certain way.
But over the last years, I've
really seen this huge challenge
with it's where it comes like.
For example, a product order
has a an issue in manufacturing.
There is a standstill.
There is a quality concern.
Now this creates some kind of
notification, but then you need
information from three or four
different systems to clarify what
actually is the root cause of that.
And if you want to do into real
time and actually analyze on that
information, it will get lost if
you only send this to the ERP layer.
So in my experience, it's, it's more
about having more depth of information
and having more flexibility to
reach the goals you need to reach.
I'm not sure if you agree with him,
but I'm a bit more conservative
perhaps in this regard, but I'm
curious to hear your thoughts on this.
Wim: I think you are completely right.
And in the end, what you say is you are
going to react to things that happen.
So, first of all, you need to be
able to to observe that something
happens in a digital way.
That you can connect your processes
and all your, your knowledge
in systems in order to make the
right decisions on top of that.
Yeah.
And then actually what you then can
see, and that's also so important for
high mix, low volume focused businesses
is that they can adjust based on what
happens instead of trying to remove
variability and risk In order that
you can pre plan everything upfront.
We have, for example, planning
and scheduling systems, et cetera,
have a huge role right now.
But in reality, all kinds of other things
happen that you didn't know would happen.
And you can organize your factory in a way
that is way more reactive or responsive
to what actually happens, which makes you
in the end way more reliable., And agile.
Denis: One thing I really like about
the event driven architecture is, as you
mentioned, everything that happens in your
business is an event with a timestamp.
And this implies that your entire
company has one data model.
It doesn't matter if the data
comes from level two from a sensor.
Or from something happening
in the MES system.
It's essentially an event with
a timestamp, and this allows you
to unify the data in one place.
Whether you want to call this place
a single source of truth or a unified
namespace, for me, it doesn't matter
too much at this point, but I think it's
very powerful to have one single data
model for all of the data, regardless
from which level of the pyramid it comes.
Wim: While the term data model is a
little bit of tricky terminology for ICT
people, I mean, because ICT people see
a data model like a relational database
and how things are connected, et cetera.
But you are completely right, you
can have just a standardized event
and it doesn't matter from which
system it comes or whatsoever that
is the standard description that that
event happens with that collection
of data, which is really powerful.
And it gives you a lot of
flexibility, not only integrating it.
But for example, decouple the information
from the system where it originated.
So you are no longer dependent on how
the data is structured in the ERP or a
CAM system or a time registration system.
The event is the event.
And if you want to implement or switch
to another system, you can just do that.
Keep the event the same
and it still works.
So it decouples everything
in your business.
And from an IT perspective,
it makes you way more agile.
It also doesn't matter which
brand of machine you have.
If you make that machine observable
and you transform the information
to a standardized event.
It will just work and, and things
become actually plug and play.
So there are a lot of advantages, I think.
Denis: So when would you say that the
unified namespace is a variant of event
driven architecture, that was designed
specifically for manufacturing and in how
far do you see the unified namespace as
proposed by Walker Reynolds as a potential
solution or implementation of an EDA?
Wim: I have a software background,
so I'm aware, or I have experience
in applying all kinds of.
Event driven architecture types.
Also, if you look at quotation factory,
we use events all over the place
and event streams and distribution
of events through hubs, et cetera.
So events is it's an essential element
of software development in general,
especially if you want to create things
that can react independently in an
asynchronous way, so you are correct.
The U.
N.
S.
is actually an architecture.
Walker Reynolds described it and explained
it in various videos, and I think
currently is writing a book about it.
It's quite a popular way to look
at an event driven architecture in
manufacturing contexts, and I think
it's an excellent starting point.
And it gives you a really clear
view at what the ingredients are
that you need in an event driven.
Architecture for a manufacturing business.
What you have, I think, to be aware
of is that he and his business comes
from a systems integration perspective,
looking at large businesses, with
a lot of machines with a potential
of a lot of sensors, et cetera.
And then, for example, the concept
of having an hierarchy and being able
and make that hierarchy as intuitive
to understand is way more important
than, for example, a relatively
small metalworking company with
only just a couple of machines.
So the concept is the same.
It can be applied if you understand it.
It's easy, I think, to
implement and to apply.
But concepts like, for example, the
semantic hierarchy, et cetera, become
a little bit less relevant if you're,
if you have a relatively small company.
Company, but that's just, yeah, it
can be overwhelming for for someone
that is new in digitalization and
even thinking about ICT architectures,
but in essence, it's quite simple.
But also it can, you know, You can
read it a little bit of overkill for a
relatively small,, metal working company.
So that that's our role, I think, to to
explain that in a way that it is relevant
for the smaller companies as well.
Luke: Yeah, I think this is
a very good comment, Wim.
Thank you for that.
And I think we should emphasize this more
often that even if you just have one site.
And one fabrication line, or
you have perhaps two sites and
you do two different things.
That is enough to just get started
and saying, okay, we just take this
concept as it is, we don't overthink
okay, the enterprise that we're having
is our company group name, whatever,
what you feel like calling it.
We have one side that is our factory.
And from there on, you just start to.
Group things on a way
that makes sense to you.
I feel that the more I'm working
with this and the way that,
that you approach this is really
focusing on the IOT infrastructure.
So the industrial internet of
things infrastructure to have
the technology behind it right.
And I feel that the technology to make
that shift to the right technology
is more important than to really.
Focus on the effective
hierarchy that it should be in.
I think that if a company starts using
like, for example, MQTT or PCOA, and
they start working with a broker, if
they learn those concepts and they start
using tools like Node RED to visually
connect some endpoints to another, I think
that is already a major leap forward.
And I think that's where I think
everybody should start with just one
big, one good use case and one solution.
So for example, I am implementing it
now on my clients on a software level.
It's not a single machine involved.
And we are just creating a way to monitor
a shop for operations software and the
things that happen there and collect
those messages that are coming from
there in a unified way, so that we can
then connect a dashboard to that, that
monitors The current state of each order.
We can send it to the ERP with
information that we think necessary.
Perhaps if there's rework, send it to
another system to make the rework order.
So there's a bit more flexibility.
And I think that's also a very
important use case to implement it.
Wim: Yep, absolutely.
Maybe one thing I can add that's
what in the UNS, but also in other
architectures is named the historian.
So that's actually where
you store all your events.
It's one of the most simple
databases that you can imagine.
It's just you add events
that happen over time.
And why that is so important is ,you can
use those events to create projections.
So events are things that happened
and will never change because they
happened and under those conditions.
So, events and your event database,
also named historian, is something
really easy to set up and you
just dump your events there.
And then you can reconstruct from
those events what the state was of
your business at a certain moment
in time, because that result is just
a sum of the events that happened.
And you can name that projections, so
you can create all kinds of projections,
also named queries sometimes.
Or based on actually static data
because the events never change anymore,
which also means that you can create
events in a stream over time, which
makes your whole business like you
can replay what happened in the past.
And if you change parameters in
those events, because for example,
you expect a process improvement.
And lead times can become smaller, for
example, that means that you can use
those same events, play them back with
different timestamps, and then see what
the results will be in your factory.
So it's so incredibly powerful
Denis: I think it's an important remark
because I remember when I was first
studying the Unified Namespace, it
took me a while to be able to answer
the question where the data is stored.
Because in essence, the UNS just gives you
a real time snapshot of the current state
of your business and is the historian
that's often responsible for capturing
the previous states and keep history.
Luke: I think it's interesting.
I am still in the exploratory phase
of really learning what the technology
can mean for a fabrication marker.
And for me, I think it's interesting.
There's a few concepts, right?
So first of one of them is that MQTT
allows you to keep a memory in record.
So basically to hold it and keep that
one for everybody that subscribes to
that system to receive the latest value.
So that's basically the point
that is this right now, like.
Where no changes were like since,
since you logged in, this is the
latest moment is where we stand right
now for everything you want, but it
doesn't, it doesn't keep the history.
Right.
You need to write it away
in a different place.
If you want to keep history and
then you are not limited to using a
historian per se, you can use also
a traditional database if you don't
really need everything you need to
be a bit selective of what you need
to keep in your historian, right?
And then you're still free to use all
kinds of means and you can even create
this real time overview as a layer
on top of your company, just to see
where everything stands now in all the
systems combined in one unified way.
And I think that's also where a lot of
value can be found because how often
have, I think, especially fabricators have
there been searching to find a specific
part, where in the factory it is exactly
in which station, which logistical place?
Or the latest contact point with a
customer, or perhaps you need to find
which part have been reworked and
which one is reworked and when was that
done, and what was the latest answer?
Now this can be in all kinds of
different places and systems.
And to do this all through one ERP
system is going to really because
the very complicated, you might have
to create all kinds of extra tables.
And perhaps you don't need to have
the full history immediately, but just
to having a real time, or at least
like the latest state of the entire
business in every single view that you
need to have in one unified structure.
I think that's so valuable
to have that insight.
And I think that's, for me,
the major learning I've been
doing for the last year or two.
So to see that, how that can work.
Wim: And maybe that's an excellent point.
Look, and maybe for those listeners
who still think in ERP and think about
their ERP as a single source of truth.
I prefer the, the, the term single
source of latest state or current state.
So it's your current state and the
UNS is the single source for that.
The single source is actually
a, if events never change.
That means that you can distribute
events in various databases
because events are just events
and is what happened in the past.
So you don't need a single database that
has the data truth, because the set of
events is actually the truth and you can
replicate it as many times as you want.
The events are just the
events and will never change.
So instead of thinking about
the centralized database, for
example, the ERP, that's no longer
relevant if you think in terms of
events and can play back events.
You see, so the whole concept of
a single source, yeah, your set of
events is the source and it doesn't
matter whether you replicate that many
times, for different purposes anymore.
Luke: that's deep has always been.
Yeah.
Well, but I think you
may make a good point!
Because you don't have to have one source
because there are many places, but just
the information that is most relevant,
that is be used to trigger other events in
it, as I said, to, to create workflows and
statement statements and do some scripting
and automate more parts of your factory.
You wouldn't even know about that.
You could do before.
I think that's what you need.
Right?
Like, so there's so many examples
here that you can think about,
but , just, just as I said, like
the latest state, how everything's
standing right now in this moment, in
this very second where everything is.
That's going to be a really important
thing to know and to collect that
from all kinds of ways that you're
no longer limited to what, for
example, one vendor can provide.
I think if I can add one part here is
that if we talk about the different
layers in the stack a common problem
is I think for manufacturing companies,
particularly those in fabrication, that
they have reached a certain point of
saturation, where indeed they have to
ERP now they have this machines running,
they have the software with the machines,
but then they want to include sensors.
And then this big question comes like,
are they going to buy sensors from a
machine vendor and their software and
expand that MES or CAM software further.
Or do they want to bring in their own low
cost, like open source technology sensors?
And if they do the latter, where
did I store that information and how
do they interact with the company?
And that's, I think this is this,
this huge fork in the road that
is coming at them every time.
Because.
It's so easy to get like a Raspberry
or Arduino and a sensor and like detect
something, but what do we do with the
information and how can we use that
and combine it with other systems?
And I think that's the, where the
major value is unlocked that it allows
you to be more flexible with your
technology choices and connectivity.
Wim: So we are in an era in which AI
will become more and more important.
You already mentioned that Luke.
If you look at where we are right
now, we are, the head of the world is
thinking in terms of AI agents, which
means that you have little, yeah,
sets of logic that, that get agency.
So they are able to make choices and
decisions in an autonomous way, because
in the end we want to have our factory
work as much autonomous as possible.
The idea is to have agents for that.
Agents have knowledge, but need experience
to base their decisions on and the events
is actually the digital representation
of, of your experience over the past.
So, these ingredients, like the, the
unified namespace and event driven
architecture, a critical component
to get right in order to to become
a smart factory, with in the end
agents being the autonomous , logical
units that run your business..
Denis: And I think for one important
concept we highlighted, but haven't really
delved into was the semantic hierarchy,
which is the solution that allows you
to easily retrieve your data and make it
understandable by both human and machine.
So it's, in my opinion, the piece
that was often missing from solutions,
such as cloud data lakes or cloud data
warehouses, where data is copied too.
But then it's very hard to navigate those
giants databases to find what you need.
So, the unified namespace
with its semantic hierarchy
,
Wim, would you say that that's an
important aspect of it as well?
Wim: Absolutely.
.
Especially.
If you look at the traditional way to look
at hierarchy in UNS, it's a little bit
the, and the examples you see a lot ISA
95 model from enterprise to what is it?
A workstation.
Now there's another term
for that cell or what is the
Denis: So you need, I think,
Wim: but that's just an example.
What I understand from communicating
with, with other people about
this is that first of all, that's
not the only possible structure.
So theoretically you can have various
structures at the same time, for example,
a structure based on where things are
located or a more functional structure.
So I'm very curious , over the next
couple of years, what the, the, the
default or the out of the box box
structures are that we need in order
not only to, to look at assets, but
for example, also to, to product flow.
So especially in the high mix, low
volume, it's essential that you, that you
look at the flow of the product through
your factory and look at touch time
and and idle time and things like that.
So I think, we will figure out many
ways to structure the namespace
according to what we want to
like to accomplish and to solve.
Luke: That's let's maybe take a little
second here and then summarize what
we have discussed today and some
takeaways and perhaps opportunities for
the listeners, when they are curious
to start implementing these ideas.
So let's start with that.
Wim, if you could perhaps answer the
question of the, why should companies
consider changing to this approach
instead of the legacy way of doing things?
What it can bring them?
From your perspective as quotation
factory and your journey so far,
Wim: Yeah.
So two answers to that.
First of all, I think it makes
systems integration way more
easy and way more flexible.
Another part is that by storing
events,, you are storing the most
critical piece of data that you need.
That is the really gold
it's that you have to store.
So don't think in terms of systems
with their databases, Have your
events stored in a standardized way.
That's how you build up experience
and that's how you make
your business play backable.
If that's an it's an existing term and
that is needed in order to simulate
future improvements or changes.
That's one.
If you look at quotation factory,
why we are helping our customers with
the notion of events , is that in
quotation factory, we want to help our
customers to create closed loop systems.
And a closed loop system means that if
you estimate, for example, production
time or lead time or whatever, that, that
you don't only estimate it, but you are
able to compare it with the real values.
And that you use that comparison in order
to optimize your estimation logic, at
first, a human being will be in the loop.
But we have, of course, nowadays,
machine learning techniques such that
we can do the comparison automatically
and train machine learning models.
So, in the event, machine learning
will be able to substitute all kind of
formulas that we use now for estimations,
but in order to create a closed
loop system ,or a feedback we need to
be able to obtain those events from
systems as well as from the shop floor.
And we are a cloud based platform.
We have currently over 50 customers
with all their own systems, and it's
undoable to create all kinds of point
to point solutions or integrations.
With all those CAM systems,
ERP systems, et cetera.
So by standardizing events and
being able to subscribe to those
events, we create an a platform,
That is able to obtain with relative
low costs, all the necessary
information to really improve on
estimating and therefore becoming a
more reliable metalworking company.
Denis: that's my experience.
Companies often ask me, well, why
should we need a unified namespace?
And I think it's companies like quotation
factory that solve an interesting use
case or the reason to go for it, because
you need the solid foundation in order
to fully benefit from the solutions that
are currently available on the market.
Wim: Yeah, exactly.
Denis.
Yeah.
Otherwise the UNS or an event driven
architecture, it looks like it's just
an IT problem, but it's far from that.
It's actually the, the essence of being
able in the future to really realize the
digital transformation of your business.
Denis: I don't think I
could have said it a better!
Luke, is there anything that you
would like to add from your side?
Luke: I think it's very well set.
I do think that there is more
to it than only being able to
implement certain solutions.
I think if we talk about the structure
that we talked about today, as I
repeated a couple of times can you
tell right now in your factory?
Exactly where everything is from a
second ago, or like the latest movement
made from every single, and what I
mean with a thing, it could be an
order, it could be a physical product.
Can you see right now how the factory
is standing in this moment right now?
And if that answer is no, then.
You probably have to still manually check
things and that is costing you time.
And if you want to see that during a
traditionally, you need to make a lot
of integration work over various systems
to legacy databases to get that answer.
So I think that's where a lot of
value can be unlocked right now for
anybody that wants to start with
this, but that's just my experience.
And I think it's just, we building on
top of what Vim is saying with a specific
use case to quote better, to have better
ideas of your lead times and get better
overview of how things went and how you
think they will go, but there's also
the other side of the coin is like.
How do we stand right now?
Where is something right now?
Like, you know, what, how is it going?
And so I think there's two, two sides
here that you could try it from.
And both are valuable.
Of course.
Wim: Absolutely.
One use case that I see more and more
often, by the way, is also if you
have plans to add a machine or two.
To add an complete new capability.
For example, you always did sheet cutting
and now you want to do tube cutting.
If you have an event driven
architecture, you can just plug in a
simulator of that machine, build the
whole infrastructure, such that that
machine will be able to connect to
the network and it, and the rest of
all the integrations are already done.
And that's totally the
opposite of what happens now.
They buy a machine, it will be
installed, and then the whole
IT integration problem starts.
And it takes them months in
order to get that up and running.
And that's a waste of time, so you
can really start doing the integration
upfront, even before the machine is
actually there on your shop floor.
I think it's also a really
interesting Use case.
Denis: Yeah, that does sound like a dream.
I think there is enough material
here for a future episode on the
potential of a UNS once it's installed!
I think we haven't touched this subject
yet, but I think the UNS or an EDA, if you
want, opens a whole range of new markets.
Wim: I agreDenisnis.
And by the way, it, maybe the
audience thinks it's new, but all
the ingredients, everything, it
already exists for over 20 years.
And if you look at the oil
industry, et cetera, all these
things are already in place, but
then for example, from a security
perspective or other perspectives.
So it's really this moment
in time where industry 4.
0 is, it looks like
that's the way forward.
There's that these type of
architectures become really critical
to implement in order to move
from industry 3 to industry 4.
0.
Yeah.
Great times.
Luke: Excellent.
Yeah.
Exciting times to never stop learning
and experimenting and finding ways
to do the same better and faster and
cheaper, even new things, I must say.
Yeah.
Denis: Yeah.
Right.
Wim, where can people go to learn
more about you or quotation factory?
Wim: Well, you can find me on LinkedIn.
That's interesting.
I think to follow because I mentioned
a lot of things outside of what's
specifically for quotation factory.
If you are interested in our platform,
go to our website, quotation factory.
com and you will find lots of information
and you can contact our customer
support team almost immediately.
And nowadays we can get you up and
running in less than five minutes if
you want to play with the platform.
So just let us know.
Denis: Great.
We will put the links below.
I think that's it for this show, right?
Look.
Luke: Yes.
Thank you so much for joining us.
Wim: It was a pleasure.
Thank you for inviting me.
Denis: Yes.
And thank you to listen for listening
and we will see you in the next
episode of the smart metals podcast.
Luke: Bye
Denis: Thank you.
And bye bye.