00:00:03 Team roles’ importance in supply chains.
00:00:36 Supply chain executives’ roles in data management.
00:03:28 Need for data officer outside IT department.
00:05:39 Key deliverables from a data officer.
00:06:59 Role and tasks of supply chain scientist.
00:08:00 Supply chain scientist’s role in data prep.
00:09:31 Challenges in joining data sets, data prep.
00:11:13 Pros and cons of internal supply chain scientists.
00:12:51 Balancing internal, external roles for data officers.
00:14:53 Supply chain decisions: challenges and experience role.
00:16:02 Need for a dedicated supply chain coordinator.
00:17:02 Considering physical realities in supply chain decisions.
00:19:28 Challenges in unconventional supply chain roles.
00:22:17 Strategies to attract engineers to supply chains.
00:24:42 Success factors for cohesive supply chain teams.
Summary
Kieran Chandler and Joannes Vermorel are discussing the key roles in successful supply chain initiatives, which include supply chain executives, data officers, supply chain scientists, and coordinators. Vermorel emphasizes the role of executives in aligning the company’s vision, the data officers operating independently from the IT department to extract and architect data, and the scientists’ responsibility to generate optimized decisions based on that data. He introduces the concept of a “product kaneto” or coordinator, who communicates management’s vision while collecting ground-level feedback. Vermorel also tackles the challenges associated with hiring for these roles and the crucial role of team cohesion in resolving complex supply chain issues.
Extended Summary
Kieran Chandler, the host, initiates the interview, emphasizing the significance of a robust team in the successful implementation of supply chain initiatives. He then invites Joannes Vermorel, the Lokad founder, to impart his knowledge about different job roles crucial for this success.
Vermorel elucidates the role of supply chain executives. He explains that these individuals primarily take charge of conveying and aligning everyone with the company’s vision of optimizing the supply chain. They face the challenge of fragmentation in the supply chain, marked by vertical and horizontal segmentation. The executives strive to transition the organization from this fragmented ‘matrix’ structure to a more interconnected one. This shift isn’t for change’s sake but to tackle the inefficiencies of the matrix approach, which can limit flexibility and hinder the optimal function of the supply chain.
The conversation then proceeds to the role of the data officer. In contrast to the prevalent practice of situating the data officer within the IT department, Vermorel proposes that this role should function independently. This autonomy is essential because the data officer’s task is to extract data from multiple systems, a task that could be compromised if they were intertwined with the daily urgencies of the IT department. The data officer’s priority is to design a consistent data representation for all things pertinent to the supply chain.
The key deliverable from the data officer, Vermorel details, is a ‘production-grade’ data pipeline that extracts data daily from all relevant systems and presents it in a way that can be programmatically exploited. This data is typically consolidated into a ‘data lake’ - a storage repository that stores a massive amount of raw data in its native format until it’s needed. Alongside this, the data officer also supplies comprehensive documentation of the data lake so that others in the organization can comprehend and utilize the data effectively.
Shifting focus to the role of the supply chain scientist, Vermorel portrays them as the individuals in charge of crafting mathematical models based on the data supplied by the data officer. This involves preparing the production data so it’s appropriate for statistical analysis and forecasting, as well as generating models that yield optimized decisions. The supply chain scientist also offers Key Performance Indicators (KPIs) to assure the rest of the organization that these decisions are effective and under control.
In reaction to Chandler’s suggestion about separating data preparation and modelization tasks, Vermorel concurs that some tasks can be shared with the data officer. He underscores, however, that the supply chain scientist plays a vital role in making the data appropriate for statistical analysis and creating models based on this data.
Vermorel conveys the challenges that emerge when handling data, particularly when attempting to combine disparate datasets. He clarifies that while a data officer can streamline this process somewhat, there are still substantial complications. For instance, aligning sales and return data can be complex and the methods to resolve this can differ depending on whether the issue is approached from a supply chain, marketing, or internal audit viewpoint. Therefore, Vermorel proposes that data should not be overly prepared; instead, it should be kept as close to the production systems as possible, while eradicating as many IT-related complexities as feasible.
He then examines the roles within a quantitative supply chain initiative, focusing on the balance between having these roles internally or externally. For the executive part, it can be outsourced to strategic consultants for validation. However, for vision and leadership, it needs to be internal. Similarly, a data officer could be external, but there is a need for familiarity with the company’s IT landscape, which makes an internal position more efficient.
Vermorel underscores the role of a supply chain scientist, who creates models that guide decisions. However, the impact of these decisions may not be realized for several months and can carry substantial costs if
errors occur, such as halting a manufacturing plant due to stock outs. Given these high stakes, he suggests starting with an experienced external supply chain scientist who has worked across multiple companies, before nurturing this competency internally.
The first role Vermorel mentions is that of a supply chain scientist. This individual takes responsibility for analyzing data, ensuring the economic modernization aligns with the company’s strategy, and understanding the constraints of the supply chain. However, Vermorel acknowledges that such individuals often lack the time to examine all the necessary workflows.
To bridge this gap, Vermorel introduces the concept of a “product kaneto” or coordinator role. This person acts as a mediator, communicating the management’s vision to the workforce while also gathering essential feedback from the ground level. This information exchange ensures that the automation generated aligns with the realities of the supply chain’s execution.
Next, the conversation veers towards filling these newly defined roles. Vermorel acknowledges that locating suitable candidates for these positions can be challenging due to their unconventional nature. For the data officer role, which requires a blend of IT skills and a readiness to work outside of a traditional IT environment, he suggests searching for individuals with experience as data architects or administrators.
The role of a supply chain scientist is generally filled by engineers, but attracting talented engineers to a field that might not be perceived as “cool” or cutting-edge can be a challenge. Vermorel suggests looking for capable individuals who may not necessarily be the “rock stars” of their fields, but who are nonetheless competent and skilled.
The coordinator role, Vermorel proposes, would be best filled by individuals with an MBA or a similar background, ideally those who exhibit an entrepreneurial or “intrapreneurial” mindset. This role requires a high degree of organization, stamina, and clear communication skills due to the need for constant interaction with various stakeholders within the organization.
Discussing the ingredients for a successful supply chain initiative, Vermorel highlights the need for a team capable of addressing tough problems without resorting to personal disputes. He underscores that issues in supply chain management are complex and require time to solve. Maintaining cohesion and patience within the team is crucial to prevent premature abandonment of a project that might have been on the right track, only needing more time for solution realization.
Full Transcript
Kieran Chandler: Today, we’re going to be discussing some of the job roles behind that team and understanding how their individual skill sets can contribute to either the success or the failure of a data management initiative. So Joannes, let’s start off by introducing a few of these different job roles. Let’s begin with the supply chain executives. They’re always going to be at a higher level, not involved in day-to-day operations. What is their main role in all of this?
Joannes Vermorel: The supply chain executives play a crucial role in initiatives that are probably of primary interest to Lokad, such as quantitative supply chain initiatives. Their role is fundamentally different in terms of how they optimize the supply chain. The executive’s responsibility is essentially to bring everyone on board with this vision and to align the efforts. This is no small thing. In our previous episode, we discussed the fragmentation of the supply chain, characterized by both vertical and horizontal segmentation. For example, you might have various departments representing ranges of products, managed by different people. You also have another dimension to this matrix, with different roles responsible for forecasting, planning, and so on. However, even if every single cell of this matrix is optimized, the supply chain as a whole can be inefficient. Anything that happens between two different cells of the same matrix can’t be handled with a divide and conquer approach. At Lokad, we challenge this matrix vision for the supply chain. Our goal is to push the supply chain executive to think beyond the matrix into a more interconnected system. The performance of your supply chain is typically constrained by a bottleneck that can be anywhere. There’s no point in micro-optimizing everything locally if you’re just moving a problem from one place to another.
Kieran Chandler: Let’s talk about one of those cells in the matrix that we’re discussing here. If we look at the role of a data officer, often, we’re talking about that data officer not being part of an IT department and acting on their own. Why is that?
Joannes Vermorel: If you want to carry out a quantitative supply chain initiative, you need to extract data from many systems. You’ll need purchase orders, product lists, supplier information, and more. For any sizable company, you’ll end up having to extract data from a couple of systems, potentially a dozen if the company is large. If the person doing this job depends on the IT department, their priority is to keep production running smoothly. Anything else is a distant secondary priority. Therefore, you need a data officer who has all the necessary IT skills to extract large amounts of data but is not entangled with the IT department’s priorities.
Kieran Chandler: Could you explain the day-to-day urgency of the IT department and why it requires separate attention?
Joannes Vermorel: Yes, the IT department primarily focuses on maintaining the production systems to keep them up and running. Hence, they need separate attention. It’s also crucial to have a dedicated person or function because you want to build a consistent vision of the data. If you rely on a matrix, the risk is you end up with as many representations of, say, the inventory valuation as you have departments. What you really need is someone who can architect a consistent data representation of all things relevant for the supply chain.
Kieran Chandler: So what’s the key deliverable that the data officer is providing then?
Joannes Vermorel: The key deliverable is typically a production-grade data pipeline. This will run daily, extracting data from all relevant systems and presenting them in a way that can be programmatically exploited. Essentially, the data has to be consolidated in something like a data lake, which are databases specialized in serving the data in bulk. The goal isn’t to serve the data line-by-line; it’s more about providing the entire sales history for the last few years, for instance. This is typically referred to as a data lake, and you can find plenty of data lake solutions on all major cloud computing platforms. Another deliverable is having consistent documentation of the data, so people know how to consume the output.
Kieran Chandler: Okay, so we talked about that data lake. The person who has to deal with it is the supply chain scientist, who seems to be incredibly busy juggling a lot of responsibilities. What are they doing on a daily basis?
Joannes Vermorel: The supply chain scientist is responsible for generating a model, often a mathematical one. Sometimes it involves mundane tasks like preparing the production data so it’s ready and suitable for statistical analysis. One of the most obvious tasks is demand forecasting, but you also have lead time forecasting and other uncertainties. However, to do this kind of forecasting, you can’t just use the data directly extracted from the production system. There are plenty of artifacts that require careful thinking. For example, sales do not equate to demand. If you have a stock out, your sales might drop while the demand is actually increasing. The supply chain scientist generates a model based on these data, ultimately delivering optimized decisions. They also provide KPIs to prove to the rest of the organization that those decisions are under control and improving things.
Kieran Chandler: In terms of modernization and data preparation, wouldn’t it be more efficient to split those two tasks? Have one team responsible for modernization and another for preparation?
Joannes Vermorel: To some extent, yes. The data officer can indeed make the task easier for the supply chain scientist. There are a lot of things he or she can do. In the data lake, for example, data can be made as consistent as possible, such as having the same format for numbers or dates. If you have master data in place, you can ensure there’s a consistent way to identify the products that you buy or sell, or anything else across the organization.
There are tons of things that can be done by the data officer to level the field so that datasets are more immediately ready to be processed. However, the problem arises when you want to join datasets. Suddenly, you encounter an unbounded amount of complications. For example, if you want to attach statistically meaningful data where sales and returns are involved, you can face a lot of complications.
The way you want to do these attachments might differ depending on whether you’re approaching the problem from a supply chain perspective, a marketing perspective, or even an internal audit perspective. This is where the business angle comes into play. Therefore, we cut the line by making sure that those preparing the data remain relatively business agnostic.
The problem is, if you don’t do this, you risk premature optimization where the data has already been reframed in a certain way that prevents certain classes of optimization from happening at a later stage. Essentially, as a supply chain scientist, the data should not be overly prepared. It should be as close as possible to the production systems, while removing as much as possible all the IT artifacts and IT-related issues.
Kieran Chandler: A lot of the time, supply chain scientists are acting externally, working with us here at Lokad. Why does that make sense? Wouldn’t they have a better understanding of business processes and a better understanding of the business on a day-to-day basis if they were acting internally?
Joannes Vermorel: That’s a very interesting point. For all roles in a quantitative supply initiative, there’s a balance between having it done internally or externally. For the executive part, the way it’s often delegated externally is by having strategic consultants validate management decisions. This is typically what well-known strategic consulting groups do.
However, when it comes to providing leadership and getting people to follow the vision you’ve presented, that has to be done internally. The same goes for the data officer. You can rely to a large extent on an external IT company, and many companies do that. The only limit is the familiarity with the company’s application landscape.
Over the years, internal staff become extremely familiar with your IT landscape, and this becomes a kind of capitalistic factor. The data officer’s role is probably the easiest to externalize, but if you don’t internalize it, you end up with something more expensive on an ongoing basis. This is because people will just be less efficient due to a lack of familiarity and experience with your IT landscape.
When it comes to the supply chain scientist, we also have a balance between internal and external. But there’s a subtle twist. This person is going to produce a model that generates decisions. With the data officer, it’s pretty easy to see if they can access the data and if the work is done suitably or not.
However, with a supply chain scientist, it’s also possible and fairly straightforward to assess the quality of the work, but it comes with a small twist. When it comes to a supply chain decision, you can typically only assess its correctness, or the fact that it was a bad decision, about six months down the road.
Kieran Chandler: You end up with a very specific challenge, typically decisions that have a very asymmetrical cost reward ratio. This means you might be able to save a bit by reducing inventory, but if you face a situation where you have dramatic stock outs, you could have a whole manufacturing plant halt just because a few things are missing. The cost can be highly asymmetrical and it takes a few months to get there.
Joannes Vermorel: Yes, exactly. It’s the sort of situation where you don’t want a supply chain scientist tackling it for the first time in your company. It’s better to start with someone from a team that has done this for many companies, and then gradually try to ramp up your internal competency on this topic. Over the long run, you can build this competency internally, but for the kickoff, it’s better to find people who have already done this elsewhere. That’s the balance, and that’s why at Lokad, we provide the supply chain scientist as part of the package, at least initially.
Kieran Chandler: That makes sense. The final piece of the jigsaw, so to speak, is a project manager. We’re talking about a very small team here, so why do you actually need a project manager?
Joannes Vermorel: You typically need a project manager or coordinator on quantitative supply chain initiatives because the vision laid out by the supply chain management needs to be relayed to many people, which takes a lot of time. It’s beneficial if the supply chain director isn’t the one having to speak individually to tons of parties. As part of management, this person is probably already spending a large portion of their time speaking to many parties. It really helps if there’s someone whose function can coordinate a lot of people and do all the mundane work of making sure everyone is on board. It’s very time-consuming.
The project coordinator also probes the fine print of all the supply chain workflows. A decision generated by the model produced by the supply chain scientist, like inventory movement decisions, are only optimized if they are compliant with the hard reality of the supply chain. For example, if you decide to put this quantity of inventory on a shelf but it physically doesn’t fit, it doesn’t matter if your model tells you that it’s a good decision. The reality is you’re exceeding the shelf capacity, so it cannot be a good decision to move this much inventory on the shelf.
Your mathematical or statistical model may fail if it violates something that’s a fundamental reality. There can be tons of very subtle constraints that emerge from the workflow. Sometimes it’s literally the way the supply chain is physically organized. It takes a lot of time to assess the fine print of that and the supply chain scientist is already very busy crunching the data, making sure that the analysis makes sense, and that the economic modernization is really aligned with the strategy of the company.
The person who knows most about these constraints is also the person who needs to be informed of the new vision laid out by the management. That’s why it’s very beneficial to have this role of a project coordinator. This person can both carry the vision of the management and also gather all the information necessary.
Kieran Chandler: We have received feedback suggesting that there is a need for a new role in supply chain management to optimize automation and maximize return on investment. However, these are unique roles that don’t classically exist. Where should we look to fill these roles?
Joannes Vermorel: Within supply chain management, there is a long-standing tradition of having people who are instrumental in leading change. So, I think this aspect is relatively covered and fits into the traditional picture. However, for roles such as the data officer, it’s typically filled by people who have experience being data architects or administrators.
Kieran Chandler: Could you elaborate more on the nature of the data officer role?
Joannes Vermorel: The novelty of the data officer role is that it requires IT skills but also needs to function outside of the traditional IT sphere. This person essentially becomes the IT representative outside of the IT department, which is a unique twist. However, this can complicate securing talent for this role.
Kieran Chandler: Why would this role be challenging to fill?
Joannes Vermorel: The challenge lies in the career path. In a large company, the IT department is like a mini-organization, with a clear progression for its members. But if you’re the IT person outside IT, the career path may not be as clearly defined, and that uncertainty can be daunting.
Kieran Chandler: What about the supply chain scientist role? What challenges lie there?
Joannes Vermorel: For supply chain scientists, the typical background is engineering. Attracting talented engineers to a field that might not be seen as the most glamorous can be a challenge. Young engineers might aspire to work for high-profile companies like Apple or Airbnb.
Kieran Chandler: Is it just about the prestige of these companies or is there more to it?
Joannes Vermorel: It’s not just about the prestige. These companies are seen as cool because their management is highly capable and leads by example. A supply chain scientist wants to be able to look up to someone more experienced and capable than they are, and aspire to be like them in the future. For traditional companies that don’t have this in-house competency, this can be a challenge.
Kieran Chandler: So, how can these companies overcome this challenge?
Joannes Vermorel: They have to take small steps at a time. Perhaps they may not be able to hire engineers who would have otherwise worked for Google, but they can still find fairly good talent with a slightly different profile.
Kieran Chandler: So, I would say, it’s obviously good that they can work anywhere they choose. For the coordinator here, it’s typically a profile that you get from people who have done an MBA. The best profiles are people who are almost wannabe entrepreneurs, young things like that. I think they’re called intrapreneurs when you have an entrepreneurial mindset within the company.
Joannes Vermorel: Yes, typically, they are the leaders. One of the qualities you need is a huge amount of stamina because you will be relentlessly talking to tons of people. You need to be very clear in your communication so that you don’t scare everyone and you don’t betray the management’s vision. That takes a lot of energy. You also need people who are very organized. If you’re in contact with many people in the organization and you’re not very organized, you can actually generate a lot of disorder. It’s very important to ensure you don’t have a net negative impact on the entire initiative by causing chaos across the organization.
Kieran Chandler: Okay, and to wrap things up, what would be the key ingredient that this team needs to ensure that a supply chain initiative succeeds?
Joannes Vermorel: Probably the key thing is to be tough on the problem rather than being tough on the people. That’s a general mindset because the problems being dealt with are very difficult and they will challenge the egos of many people. One of the biggest challenges is to maintain something that really acts as a team with a lot of cohesion so that the problems, no matter how difficult, are addressed. When a problem proves difficult, it also means that it will need more time to be solved. If you lose cohesion just because a problem suddenly requires a few more months to get fully addressed, then the whole thing can fall apart even if it was on a good track. It’s just that it needed more time to deliver a satisfying solution instead of rolling out a quick fix where people say, “This machine is insane, let’s stop this madness and stop even trying to optimize the supply chain.”
Kieran Chandler: Great, we’re going to have to leave it there for today, but thanks for taking out the time.
Joannes Vermorel: Thank you.
Kieran Chandler: Okay, so that’s everything for this week. We’ll be back again next week with another episode. Until then, thanks for watching.