Skip to content

Configuration Lifecycle Management as a Competitive Advantage w/Henrik Hulgaard

In this episode we will be talking with Henrik Hulgaard from Configit.

Henrik has a background in Electrical Engineering (MSc from Danish Technical University in 1990) and computer science (PhD from University of Washington 1995). As a researcher, he has worked with formal verification of circuits and systems. 

In 2000, Henrik founded Configit together with Henrik Reif Andersen and three of their students. Configit utilizes its patented VT technology to offer configuration solutions to companies that have complex configurable products such as Philips, Jaguar Land Rover, and ABB.

Subscribe and leave a review, would ya?

Links

Transcript

Myriam

Welcome to the Share PLM Podcast. We are very excited to have you here today. And you’re going to tell us a little bit more about product, about configuration management products, tools.

I was wondering if you could explain to our listeners what a configuration management tool is.

Henrik

It’s basically some sort of piece of software that you use when you have configurable products. So any product where there is a variant where you need to select certain choices, sometimes called options or features, before you can order and manufacture the product. So that is a classical example of that is a car. You need to select engine and colors and other aspect before you can order and manufacture the car. And the reason you cannot start manufacturing it before you have made the choices is that there are so many different combinations, literally billions and billions of combinations, typically.

So it makes no sense to try to guess what the customer wants. So you need to wait for the customer to have made the configuration, and then at that point, you can take the order, and then you know all the parts that go into that particular configuration and you can manufacture it. So that’s what you use a Configurator for that is to support those processes, which is quite essential for that type of product. There are two parts to a configuration solution. There is you could say the rule management, all the different dependencies between the choices that you need to express as rules.

That’s one part that’s a big part of what we do in Configit is to have very good tools for managing the rules and making sure that they’re right. And then the other part is the front-end application that should allow a customer or sales rep to basically build up a configuration. So select the choices. It sounds simple, but that can easily be many thousand choices to be made. For complex systems, you may have to build up a system with many components, maybe a whole plant or other very complex systems where there are many dependencies.

So you need a lot of support to help the person who has to build up the system, who doesn’t know all the details of the rules but help him guide him to find a good, basically, an optimal solution that best matches his requirements with the technical capabilities that are possible.

Myriam

Okay. I see. And I’m wondering while you’re talking, because I know that in PLM, for example, we can also manage variants and options. So how will you explain how a configuration tool differs from a PLM? And a company do have a PLM and they’re using it as an enterprise backbone? Do they still need a configuration management tool sometimes?

Henrik

Absolutely. And I think the reason is that very parts and bom-oriented. It’s very engineering-oriented. And these applications are sometimes or most of the times, also heavily involved in the commercial side of things. So one way of looking at it is that you’re right that PLM system often has a configurator with the ability to express variants and rules. It’s interesting enough, so do the sales tools in the sales area. You have what’s called CPQ configure, pricing coding. And there you also have a configurator. And in fact, if you look at the manufacturing side of the ERP system, they also have a configurator.

If you use SAP, they have the variant configurator and all of these configurators. They’re sort of there to help engineering express the engineering variance and in sales to express the commercial variance and in manufacturing and heap to express the variance over the bill of materials that goes into the manufacturing. But they each have their own separate perspective. And what we bring to the market is a view where the configuration part is centralized in our software, and then we integrate to those three enterprise systems. So in a way, we express the variance across all these different views.

So the PLM system is you could say the full view of the engineering capability, but it doesn’t have the information that links that to the customer requirements, for example. So if I want to again use a car as an example, because that’s so easy to understand. But if I want to get a car that say environmentally friendly cost at most $30,000 and can, let’s say, hold four people, that might be my overall requirements. And then there is a mapping from those down to, okay, what models can I actually get that will fit that, and that linkage is not to be found in Elm system.

Myriam

Okay, I understand. And in this case, so you say that configuration management is the main version of truth in terms of options. So how does it connect to other systems, such as PLM ERP, and how complex it is to pull the data from those other systems?

Henrik

So you absolutely write that the way we integrate is that we use these option codes or feature codes. That’s sort of the language, so to speak, that connects to different systems. And then you need to map that to a price in the sales area, and you need to map it to a bomb in the engineering space, and the integrations can be done in different ways. I would say we have some out of the box integrations, for example, to Rs and PDC when it comes to the PLM space and to SAP and Salesforce when it comes to the manufacturing and sales areas.

And there are different ways to do this. And it’s always, of course, a project to get this established, not so much the technical integration, but more on the data side to make sure that all the data is aligned and consistent.

The approach would be to still maintain bill of materials, for example, with the variability within the PLM, but you then maintain the configuration logic outside. And then there is a test to make sure that anytime you have a configuration that’s valid, then it has a correct corresponding bill of material in the PLM system. And that’s an analysis that we do. And it’s called one validation, where you basically make sure that the data in those two systems are fully aligned before you before you release it out.

Myriam

Okay. I see from your experience, what would you say is one of the greatest challenges that companies typically face when they’re trying to configure complex products?

Henrik

There are a number of challenges, for sure. I mean once you do have complex products and they certainly are complex when you have even a car is a very complex product. There are maybe a thousand parameters you need to set for a car. You may think that’s a lot because you certainly don’t choose a thousand as a customer, but many of them are driven from just the differences in the different markets. So in some areas, the steering wheels on the left side of the car and other areas it’s on the right side of the car.

So there’s a variance there that has to be managed. So I think managing that complexity basically both from the rules perspective, making sure that all the constraints, the rules are correct, but also managing the bill of material and reducing the internal complexity. So reducing how many parts you have internally and still maintaining the variability on the commercial side. So those are the key challenges, and they’ve been around for a long time. I think more recently, we see some additional challenges that deals with with software, and particularly about how you can extend the capability of a product after it has been delivered to a car.

Again, typical example, you you bought a car, but then two years later, you are offered a an additional capability of the card that it didn’t have initially. And a lot of that would be a software upgrade and understanding the configuration of the car and whether you can actually have that upgrade and have it work in your specific configuration. That’s a pretty big challenge for many customers of ours.

Myriam

Yeah. I can imagine that the software component of it definitely is going to make things much more complex.

Henrik

It’s also a way to add variability to your product. I mean, a configurable product, you may think of it as the configuration is purely in the physical elements of the product, but you can push variability as part of the software, and we all do that with our phones. You could say our telephone is pretty standard product. It comes with a few sizes and a few colors, maybe. But besides that, all the configuration is done by the user in purely an software, depending on what apps they install.

And that is one extreme, of course. But even very standard, traditionally hardware-based products will increasingly get software and the configuration challenges of managing the software parameters is also tied into the overall configuration process. So as a customer, you want a certain capability, and you don’t really care whether that capability is delivered from a piece of hardware or whether it is just turned on by setting a software setting

Myriam

It does. And will you say that in general with an organization, you feel that this topic of product configuration is well known? Discuss mostly in the context of those challenges are becoming increasingly more real and present in terms of software and big data and all those things.

Henrik

I think a lot of I mean, some of our customers are very mature when it comes to the configuration challenges. They have done it for many years. Again, automotive is classical example, but others that we worked with are also have been working with configuration for a long time. But I think for some organizations, some manufacturing companies, it’s a pretty new topic for them. And they have been so used to being very parts oriented, where they basically sell parts, stand up parts with, you know, full fixed price description and everything.

And then they start moving into the area of selling not just individual parts, but maybe whole systems with combinations of parts still stand up parts. But once you take that step and want to sell more complete offerings, then you need to start having a configurator to help you about what parts will actually work with what other parts. And we saw this very clearly with one of our customers that were selling scanners for hospitals. So Mr Scanners and ultrasound scanners and that type of thing traditionally had been very partial oriented, but wanted to move to selling a whole sort of room with all this stuff in there and have the customer just select the requirements to the scanner itself.

And then they are immediately moving into the configuration space with the challenges that brings things.

Myriam

I see it’s interesting what you say about the scanner, because I was actually asking you, which industry do you think can benefit from a configuration management platform? Because obviously we mostly think about cars. Now you talked about scanner, which is not something I will have messily still about. Yeah. Do you think it’s relevant for most industries, or is it maybe more suited for certain industry? Can you give us some examples?

Henrik

Yes, certainly. So Configit has been doing this for almost 25 years now, and we have over the years focused entirely on automotive and on what we call industrial machinery and components. So those are companies that built complex machines that needs to be configured. It could be a printing machines from Hiddle Bar trick machine, or it can be components. So we also have a customer like signify that used to be Phillips Lighting. They do you could say, you know them from doing light bulbs, but they are moving towards doing entire lighting systems.

Lighting up a football Stadium or parking area or maybe even taking over the lighting system of an entire city. So I think that a lot of companies are starting to also move in this direction of selling services instead of selling the products themselves and signifies a good example of that. They basically would like to instead of sell you a bunch of parts, they would like to sell you the capability of getting lights on the streets. And then we have some really sort of even though we try to focus on those two areas, we have other customers who are doing very different things.

We have a customer called Equal Label. They basically are it’s a company that provides a sort of a certification of products. It could be a shampoo. And if you get this certification, you can put their label on your product and then the consumer would recognize it and see, oh, I know this product doesn’t have any things in it that harm my skin or stuff like that. So they are basically just, you know, managing a lot of rules about what does it take to get a certification for a shampoo?

Or they can also certify a hotel or many other parts can be get this label. So I think there are many areas where you can use a configurator. Whenever you have some parameters and some rules and need some help in finding a valid combination of the parameters, then you can use a configurator and we have used it. We’ve had insurance companies that sell insurance where you configure the insurance basically, for how many years? What covers do you want that as outcomes, the price and the system guides you to find a valid combination.

We have done legal texts where you’re taking a law and turn that into a set of rules. This was the law for when you have a right to get a pension in Denmark, and you can then start to play with the law, you can say I want a pension, and the configuration can tell you that either you have to be a Danish citizen and you have to be more than 63 years old. And I can tell you all the other ways you can get a pension.

I think to summarize that the configurator is a very general piece of software that can be used in many different contexts. But it is, of course, also a complex piece of software and an expensive piece of software. So you have to use it only when the complexity of the rules is significant. If you only have five parameters and, you know, three choices for each, you can probably just do it in Excel and hard code it by hand. Right. But the moment you get to 20 parameters each with ten choices each, no human can really figure out what are the right combinations that it quickly grows into the many combinations.

Myriam

Yeah, I can imagine just a question. I’m wondering the front-end app, so that is used by the sales rep. And so I guess it’s quite user-friendly. It’s something that people can pick up quite easily, or how is it?

Henrik

Yeah. So we built them for different purposes, depending on the consumer. I mean, some of them are highly technical and very complex and would be difficult for you and I to use because they are focused on the expert in the domain. But they are very much built for the purpose. It’s not like a one-size-fits-all. Some are very graphical with beautiful pictures. And as you can see, when you go on to pick any car manufacturer and go to the website, you will find these nice looking customer-facing configurators.

Some are very visual. We have a collaboration with people doing very nice 3D visualizations, and you can then present your product in that way. But basically this is something that we built based on some standard components, but we build it for the individual requirements for each customer. The typical configurator is a little boring, with a list of dropdowns where you go through some taps and select values and dropdowns, but they can certainly be much more visual than that. We have examples where you build up a system by pulling in components from a list of different components.

I mean, you can think of it as building up a kitchen. You pull in different types of elements, and then you configure them. And there are rules that tie them all together, and it’s done in a 2D or 3D space where you can interact with it. That’s also how a configurator could look like.

Myriam

I see. Okay. And I was wondering, I mean, you give us some example, but what would you say is the most common kind of issue or challenge that your customer’s face? So what basically motivates them to come to you?

Henrik

That’s two different questions, actually.

Myriam

Okay. The first one is.

Henrik

I think, one challenge. We see some pretty significant challenges in the complexity of the system landscape, which means that you end up with a lot of integrations. And that is also tied to the other main challenge, which is basically data consistency. It is very common that since data is traditionally being maintained in separate systems siloed from each other, for example, the PLM system and the CPQ system, there is some connection points between them, but it’s not aligned, and that causes a ton of problems. It means that orders get accepted, and then when they hit the manufacturing their flagged as Invalid, you do not offer all your capabilities and stuff like that.

So I think this getting data aligned is a pretty fundamental challenge that has to be in place before you start to introduce a sort of an End-to-end solution, which we’re talking about here where what you do in your Palm system is fully exposed and transparent to your sales channels. So I think that was probably the first question. And then the second part of your question was, why do they do this?

Myriam

What motivation? Why do they come to you? What is the reason they’re going to come and ask for your help?

Henrik

And there are different reasons. But I would say the very typical one is to help streamline the whole sales process. So making sure that what is sold can also be manufactured and that everything you can actually manufacture also is exposed in the sales channels with the right data at the right pricing. That is the most fundamental challenge. But there are other reasons. And I would say one of the other reasons is to to help on the rule management. Many of these applications that we talked about up to this point, both PLM and oh.

Myriam

Sorry, no, you’re back. You were saying many of these applications, and then it gets for a second, right.

Henrik

The rule management is another reason why customers come to us. Because even though, as you mentioned, M Health rule authoring environment, so do CPQ solutions, they’re not really focused on the rule management, meaning that you can validate your rules, make sure that they’re right understand why they behave the way they do, create different types of reporting about what are the valid combinations. So we more and more focus on on that part. And that brings a lot of value, because that gives customers the the basis for doing these complexity optimizations that we also talked about to reduce the internal complexity and still maintain the external complexity.

Myriam

Okay, let me see. And what would you say are some of the maybe the best practice to integrate a configuration management tool to an enterprise? And mostly in terms of we’re talking a bit at the beginning about connecting it with the other tools. So we had the ERP, CRM and so on. So if some company wanted to move towards that direction, how can they best prepare?

Henrik

Yes. I think you, of course, have to address the areas where the business the biggest challenge initially. And what we see there is that you often, as I said, start in the sales area. So what we normally would start with would be to improve on the sales processes by importing and combining rules from many different sources. So we will reuse what’s already there. So if you already have rules in your PLM system, we will just keep them there and extract them out. If you already have rules in ASAP, in the variant configurator, we can extract them out and import that into our tool and reuse them.

You can keep maintaining these rules in those systems. And then we will combine all the rules and expose in a single source of truth when it comes to configuration. And from that, you will build your different applications. It could be a website, it could be a CPQ application. It could be a mobile application that will use that data. And then from there you could decide whether you want to stay there or whether you want to, for example, start leveraging some of the capabilities of our rule management.

So we have a number of our customers who want to move out of using SAP for rule management and want to use our software for that. And that you can take step by step because the consuming applications will all use the same API. They really won’t notice that you change that on the back inside. But then you get much better the performance, you get better guidance, you get better analysis, you get better quality, all those things. You then benefit. Once you start to use our rule management tools.

Myriam

Okay. And can you tell us a bit what differentiates Configit as a configuration management company than other companies that are doing similar kinds of services?

Henrik

So we are based out of a way to solve a configuration problem, which is which is really what started the company. We found a unique way to solve this challenge of having a computer guide a user to find the correct choices for the different parameters. Most of our competitors is using it as approach called constraint propagation or similar. And we have a way where we have a two-step process where we take all the rules and we compile them a little bit, like how you compile source code to binary code before you can run your program.

We compile all the rules into what we call a BT reputation T stand for virtual table. So what we do is that we calculate all the possible combinations and people say, oh, but there are a lot of them. Yes, there are bazillions of them. And still we have a way where we can we can save that. And sometimes we say we save it in a VT file. You can think of it as an output of the compilation process, but it is really that decoupling, because then the Configurator only needs to know about that VT file in order to guide the user.

So that is from a technology perspective, how we started, and that’s still our unique benefit. It has a number of technical benefits. It’s very easy to build. For example, offline configurators, you have excellent performance. So the response time is guaranteed and very, very short. It scales very well. There are many benefits from a technical perspective of that approach. But I would also say that we are also different in that we are much more focused on the rule management and the data integrity. So making sure that these different systems are integrated and aligned, where I think many of our competitors are more focused on.

Yes, they have a configurator, but then they’re very focused on the CPQ. So the sales perspective and they tend to ignore that the data has to be aligned with, for example, the data in the PLM system. So I would say that that’s just you would say more of a focus of our projects tend to be highly integrated, also very fairly long lived and complex. They can easily take a year or longer before from initial point to we have a running solution because we do really do a digital transformation of these companies to make sure that the data can flow consistently from a change being done in Pen, so that it, without any manual intervention, gets reflected in all the different sales channels I see.

Myriam

Yeah. Wow. It sounds that you’re really taking it is a holistic approach, and I think it’s important. What you were saying about the data consistency definitely in the management of data is definitely the hardest part of most projects.

Henrik

Yeah. We’ve seen some crazy stuff out there because of technical limitations. For example, some legacy systems they can only handle see 500 codes. And then they start using weird things like the voltage of the battery to determine whether it’s US or Europe-based truck. For example. Just imagine a rule like saying 8 voltage batteries, 48 vaults, then something, something. It really means the truck is driving in the US, simple things like that, but different codes in different products, that means the same thing. We all know this, try to search for a hard disk of a certain size or voltage.

People sometimes use 220 volts. Sometimes it’s eight to ten to turn 30. And and this happens also within organizations. So one of our customers app, they’re doing electrical stuff, and sure enough, they don’t have a standard enterprise code set for voltage. I mean, you would think so, but now they each division have their own sort of code for that.

Myriam

Wow. So I can imagine that’s actually because you were saying that one of the first part is to kind of compile all the rules that exist. But I can imagine there must be so big work of cleaning a lot of those rules that might be either outdated from legacy system or that might not make sense or that are, like, not unified. Is that something that you run into? That?

Henrik

Absolutely. So you’re absolutely right. And it’s interesting because very many of the legacy systems that we deal with and also PLM and others, you can write the rules and you can run a configurator, but that’s about it. It doesn’t help you very much in understanding the full set of valid combinations. So we often find that when we import the rules and run our analysis on it, and once we have the full set of valid combinations, we can make all kinds of reports. And, for example, we can tell you if you have parts, that network can be selected in your bomb, and this happens regularly, nobody dares to remove it because there is a part in your bomb structure.

It has a selection condition that says this part is included. If this and this and this, and nobody can figure out that that selection condition will never be true. And that’s part of what we do. We can do these types of analyses that really can reduce and simplify your bill material structures.

Myriam

Yeah.

Henrik

I can imagine and make sure that they’re right also, because that also happens that incomes of configuration and then you build a car and there’s no engine in it. I know that’s probably a little extreme, but but that really happens then when they have to put it together, suddenly they’re like, wait a minute. We are missing some parts here.

Myriam

Well, yeah, I can imagine that that must be something that you run into quite often. Keeping the data clean is really hard, and it’s really hard normally legacy system, but how people are going to input the data and there’s always those mistakes happening. So I guess, in a way, going to a project of integrating a configuration management is also a way to make sure that your data and new rules are still up to date and make sense.

Henrik

It is. And having this I mean, everybody likes this single source of truth. And we provide that when it comes to configuration, that there is one place there’s a service you can call it, and it will always give you the accurate details. And it has all the information from both the commercial side and the engineering side. So as an engineer, you can ignore all the commercial stuff about packaging. And yeah, all the decisions made by marketing. But look only at the technical aspects. But you can also, from a marketing perspective, just run the commercial configurator, not dealing with all the technical aspect, but it’s all in there.

So the configuration engine knows about all of them and always ensures that it that the whole set is consistent.

Myriam

Okay. That’s very interesting. Yeah. It’s so many things that you don’t think about, but actually all the things that come out okay. Was there anything else you want to talk about? Is there a question I didn’t ask you about yourself?

Henrik

I can keep going for a while.

Myriam

That is something you think will be important for a listener to know about configure about configuration management

Henrik

I think the best advice I can give in general would be to start thinking about features rather than parts and maybe start establishing as a central sort of global feature dictionary where these features get managed so that you use the same code across different products. That would be definitely the starting point for engaging with us and also just to get the data line as we talked about. And then, of course, once you have that, you can decide whether you want to improve on the PM size, to get a better rule authoring environment to optimize the variance or where they want to focus on the sales aspect to get accurate quotes and offering, you know, better sales tools to your salespeople and customers.

Myriam

Absolutely.

Henrik

That’s what we do. So it was a lot to cover in the short period of time.

Myriam

But I think it gave us a good overview. At least it really clarified a lot of things for me, so that’s very helpful. And can you also tell us where they can find more information about Configit or about you?

Henrik

Well, we have, of course, a website, Configit.com. That would be the starting point. I think we are so quite active on LinkedIn and have several groups there where people can with an interest in what we call configuration lifecycle management, which is really what I’ve been talking about up to now.

We call it, and that is a good place to learn. We used to have a yearly conference called the CLM Summit, where our customers presented their approaches and their solutions for other customers. And that has been very, very well received. And we have had some fantastic events with many hundred participants. And unfortunately, of course, we had to cancel that. And this year we did a virtual version of that, and I believe that is still out there and can be accessed to or you can see the recordings.

Myriam

If you use the link, I can put them on the podcast webpage so people can find it.

Henrik

Absolutely no problem.

Myriam

Awesome. Okay. And are you going to do another conference maybe next year, or is that still…?

Henrik

I certainly believe we will. Absolutely. Whether it will be physical or virtual, I don’t know.

Myriam

Good to know. Well, Henirk, thank you so much for your time. Thank you for sharing all this information with us. It was very insightful, and I hope that our listeners will appreciate it as much as I did.

Henrik

Thank you, Myriam.