00:00:08 Introduction and Eric Wilson’s background at IBF.
00:01:35 Challenges in producing useful numbers and insights from analytics.
00:03:45 The role of business decisions in utilizing analytics.
00:05:31 Importance of capable demand planners and the industry shortage.
00:07:25 Talent scarcity in the supply chain and analytics field.
00:10:58 The need for a new supply chain scientist role.
00:13:01 Probabilistic thinking in supply chain management.
00:14:41 The increasing importance of supply chain management in business.
00:15:47 Media’s growing interest in supply chain management.
00:17:18 Emergence of supply chain masters in elite universities.
00:18:27 Evolution of analytics and its role in the future.
00:19:59 Mature technology becoming invisible and blending into the background.
00:20:53 Example of anti-spam technology as a mature, invisible technology.
00:23:59 Definition of artificial intelligence and the value of asking the right questions.
00:24:57 The ideal silent supply chain and its implications for demand planners.
00:25:11 Eric’s hopes for his book and the skills demand planners can gain from it.
00:26:39 Closing remarks.
Summary
In an interview with Kieran Chandler, Joannes Vermorel, Lokad founder, and Eric Wilson, a certified business forecaster, discuss the increasing importance of analytics in modern organizations. They explore the challenges of using vast amounts of data for meaningful insights and the need for supply chain optimization. Both emphasize the importance of actionable analytics for better decision-making. The conversation highlights the growing demand for skilled professionals in supply chain management and the integration of technology with supply chain expertise. As analytics continue to evolve, organizations must adapt to become more agile, responsive, and predictive to remain competitive in a rapidly changing market.
Extended Summary
In this interview, host Kieran Chandler is joined by Joannes Vermorel, the founder of Lokad, and Eric Wilson, a Certified Professional Business Forecaster and host of IBF’s on-demand podcast. They discuss the role of analytics in modern-day organizations, with a focus on the challenges and benefits of using data to drive business decisions.
Vermorel holds a somewhat controversial view about analytics. He believes it’s easy to generate vast amounts of data, but much harder to produce a small number of valuable insights for human consumption. He suggests that the widespread use of analytics in companies often fails to produce useful information. Wilson, on the other hand, believes that while it’s crucial to sift through the vast amounts of data to find meaningful insights, the benefits of doing so far outweigh the costs.
Wilson emphasizes the importance of turning raw data into information and then into actionable insights. He recognizes that many companies struggle to reach the insight stage, but asserts that embarking on this journey is essential. Businesses should use data to gain a deeper understanding of their consumers and the economy, particularly in the face of challenges like COVID-19.
Both Vermorel and Wilson agree that the ultimate goal of analytics is to drive better business decisions. Vermorel points out that without a practical, tangible purpose, analytics can lead to unproductive lines of thought or action. He criticizes the widespread use of vanity metrics and a lack of focus on key performance indicators (KPIs). In contrast, he suggests that businesses should concentrate on actionable, automated decision-making based on data.
Wilson’s new book, “Predictive Analytics for Business Forecasting,” is aimed at demand planners. As data teams grow, there is increasing pressure on finding capable demand planners. Wilson believes that demand planners have the necessary skill set to grow into their positions, as they are wired to understand and communicate the factors impacting consumers and demand.
This interview highlights the challenges and benefits of using analytics in modern organizations. While Vermorel criticizes the overemphasis on data without clear actionable insights, Wilson stresses the importance of turning data into valuable information for decision-making. Both agree on the need for focused, actionable analytics to drive better business decisions.
The conversation begins with Eric Wilson acknowledging the current shortage of demand planners, as the demand for their skill set has significantly increased in recent years. Salaries have gone up 30 to 40 percent in the last five years, and job boards are constantly filled with listings for these roles. While qualified demand planners exist, there simply aren’t enough of them in the market.
Joannes Vermorel concurs, stating that talent is rare and high-quality supply chain scientists are difficult to find. He likens the situation to quantitative trading in banks, where a small number of traders generate the majority of the returns. He believes technology acts as a multiplier for human intelligence, allowing more capable individuals to operate faster and on a larger scale.
Vermorel points out that the growing popularity of data scientists in the last five years has led to an influx of professionals who are highly skilled in technology and programming languages, such as Python and machine learning tools like PyTorch, Keras, and TensorFlow. However, he argues that technical fluency alone is not a substitute for a deep understanding of supply chain intricacies.
In Vermorel’s view, the challenge lies in supply chain scientists being able to apply their skills to real-world situations beyond the scope of tech giants like Google and Facebook. The difficulty is in fine-tuning supply chain analysis, which is a different kind of challenge compared to handling large amounts of data.
Wilson envisions a future where both skill sets merge to create a demand planning role capable of handling daily model changes and incorporating analytics as an additional capability. Vermorel jokingly suggests the term “supply chain scientist” for this type of position, though he admits it’s a made-up term used by Lokad to differentiate their roles in the market.
The interview highlights the increasing demand for skilled professionals in supply chain optimization and the challenges in finding qualified individuals who can handle both the technical and business aspects of the field. While technology continues to advance, the need for a deep understanding of supply chain complexities and effective collaboration between demand planners and data scientists remains critical.
The conversation highlights the importance of probabilistic thinking in supply chain management, moving away from deterministic approaches. As the industry evolves, demand planning and supply chain have become increasingly important to companies, with executives focusing on these areas more than ever.
The rise in popularity of data science has led to an influx of buzzwords in the industry, but the interviewees emphasize the need for a deeper understanding of the actual processes. They discuss how supply chain management has gradually gained more attention from prestigious universities, with top-notch professors and students exploring the field. This shift is helping to bring more talent into the industry, which is becoming increasingly complex due to factors like compliance and globalization.
Looking ahead, analytics will play a crucial role in the evolution of supply chain management. Companies will need to be more agile, responsive, and predictive to keep up with changing consumer behavior. This will involve a greater reliance on demand planning and supply chain optimization to support targeted marketing efforts. The democratization of data and analytics will continue to drive changes in the industry, emphasizing the importance of supply chain optimization for businesses.
Wilson sees organizations becoming flatter and more dependent on analytics to drive decision-making. Vermorel believes that as technology matures, it becomes invisible, blending into the background and working seamlessly. He cites anti-spam technology as an example of a mature technology that operates unobtrusively yet effectively.
Vermorel envisions the future of supply chain analytics as largely invisible, driving mundane decisions without capturing the attention of top executives. However, he acknowledges that supply chains are diverse, and no single company or technology can capture the entire market. Despite its unassuming nature, Vermorel believes that advanced analytics will become more important than ever for maintaining competitiveness.
When asked about his book on predictive analytics, Wilson explains that it is not a math-heavy text, but rather an introduction to machine learning, artificial intelligence, and predictive analytics for demand planners. The book covers people, processes, analytics, and technology, with a focus on building data-driven organizations and understanding how to utilize data effectively within the organization.
Full Transcript
Kieran Chandler: Hey, today we’re delighted to be joined by Eric Wilson, the host of IBF’s On Demand podcasts. We’re going to discuss with him the role of analytics in modern day organizations and what we can learn from his new book entitled Predictive Analytics for Business Forecasting. So Eric, thanks very much for joining us live today from the States. Perhaps just to start off, you could tell us a little bit more about yourself and also your role at IBF.
Eric Wilson: I’m excited to be part of this and to be part of your cast as well. My name is Eric Wilson, I’m the thought leader for the Institute of Business Forecasting. It’s actually a global organization with over 50,000 members worldwide. We’re specifically about fostering the growth of demand planning, forecasting, predictive analytics, SNOP, and related fields. That’s what we do as an organization. One of the things we do is share knowledge, and that’s where I come in. I write articles, and I host a bi-weekly podcast, IBF On Demand, which you can find on YouTube or wherever you find your podcasts. So that’s a little bit about me. I have about 30 years of experience in too many industries and too many different positions, but it allowed me to get to where I am right now.
Kieran Chandler: Brilliant! Today, Joannes, our topic is all about analytics, particularly in modern day organizations. I think when we discussed this, you had a bit of a controversial view about analytics and what their actual role is. What’s your initial overview?
Joannes Vermorel: My view, in a nutshell, is that it’s very easy to produce one million numbers per second with a computer, but it’s actually very hard to produce five numbers per day that are worth being read by human beings. The biggest challenge with analytics is how to produce anything that is worth the attention of a human being. My casual observation is that what is currently widespread in companies, especially on the supply chain segment, but not only, doesn’t pass this test.
Kieran Chandler: What are your thoughts on this, Eric? Data has grown an awful amount over the last 20 years or so. Would you say that we’re now producing too many numbers and not looking at what’s really important?
Eric Wilson: I don’t think you can have too many numbers, but there is some credibility in what Joannes is saying about finding the right information. Data in all forms is just that – raw data, the building blocks you can start building something with. Turning that data into information and then into insights is where companies are struggling. However, the benefits of getting there far outweigh the cost of the journey to get there. Companies need to start on that journey even if they’re struggling now, because developing that raw data into useful insights for an organization is crucial.
Kieran Chandler: We need to stop living in the past and start looking forward, finding new insights into consumers and the economy, especially during times like we’re facing right now with COVID. We need to start opening up those insights.
Joannes Vermorel: One of the key insights we focus on at Lokad is the idea of business decisions. Where can these decisions really change the way a business operates? With analytics, there are several paths that can lead to non-productive lines of activity or thought. You need a practical, mundane purpose that drives what you’re doing with your analysis, the numbers, the display, and everything. A decision is something that has a physical, tangible impact on supply chains, like a purchasing decision, a stock movement, or a price change. If you’re looking at numbers with the direct intent to improve a decision, it can be good. What I usually see are oceans of vanity metrics, where you end up with so many KPIs that it’s almost an insult to call them “key.” They lack focus and a built-in mechanism or intent to turn them into something actionable at scale and in ways that are completely automated.
Kieran Chandler: Eric, in your new book, “Predictive Analytics for Business Forecasting,” it’s very much focused on the demand planner. We’ve noticed in the industry that as data teams grow, there’s more impetus on looking for a capable demand planner. Is there a shortcoming in the industry regarding that?
Eric Wilson: There is both yes and no to that. The demand planners are capable of growing into the position. They’re wired to look at what’s impacting consumers, what’s impacting demand, to understand and make connections with different variables. They’re wired to communicate that into the supply chain, into finance, and into other parts of the organization, making them useful insights that other parts of the organization can use. They have the skillsets to do it. That said, right now, we’re seeing a shortage of demand planners because there’s such a demand for them. Salaries have gone up anywhere from 30 to 40 percent over the last five years. We’re seeing job boards being filled even during these times with people looking for those with qualitative, quantitative, and communication skills to bring analytics and business acumen together for an organization. So, are they qualified? Yes, they can do it. But is there enough of them out there? The answer would be no.
Kieran Chandler: And I think that’s something we’d probably echo here. We’re always on the hunt for well-qualified supply chain scientists, and it’s always something that can be fairly challenging to find. Why is it so hard from your perspective, Joannes?
Joannes Vermorel: Talent is rare by definition. Every single company says they only hire the best, but the reality is the market just hires the average. These sorts of jobs are where people who are better at it get disproportionally good results. We are entering a realm similar to quantitative trading in banks, where a few traders make the bulk of the return. Technology is a demultiplier for human intelligence, so if you have somebody who is smarter, more capable, and has better business insights, they will just do it faster at a larger scale for their organization. This is becoming very true for supply chain as well, not just trading for banking and finance.
What makes it harder than it should be, I believe, is the idea of the data scientist. This has become very popular over the last five years, but the problem is that you end up with people who have been told at universities that their focus should be the technology itself. They need to become very good at Python, PyTorch, Keras, TensorFlow, or whatever the popular open source toolkit of the day is for machine learning. While it is certainly a requirement to have a certain degree of fluency with technical tooling, it’s not a substitute for a very patient understanding of what makes the supply chain tick, including the minute details that drive an organization. If you miss them, you’re completely off. Thus, there’s a bit of a struggle for supply chain scientists who have done tons of exercises on models tested and rolled out at companies like Facebook and Google. When they arrive at an actual regular company that is not Google, it feels non-ambitious compared to what they’ve seen on the Google side. The reality is that the difficulty is of a different kind; it’s not about having a massive cluster of GPUs where you’re going to crunch petabytes of data, but rather getting the fine print of your supply chain analysis very right, which is a different kind of difficulty.
Kieran Chandler: What are your thoughts on this, Eric? Your book obviously covers a wide range of different analytical techniques. Would you say that grounding and that overview is something that’s a bit more difficult to achieve?
Eric Wilson: I would agree.
Kieran Chandler: There’s different skill sets in that data scientist and the demand planner. There’s also a lot they can learn from each other as well, and I think that’s a great overview of what exactly you talked about, some of the struggles that we’re seeing. I mean, the demand planner needs to be more based in science. They need to look at things and look at external variables, look at new technologies, modeling those things that are really the data scientist’s world. They, the demand planners, need to come out of their comfort zone and do more of that. At the same time though, that collaboration, that being comfortable with ambiguity, those type of situations, communication, all those things that demand planners have as strong skill sets, those are things that help them out, and that’s where the data scientists need to come as well. So there’s really going to be that meshing of the two type skill sets going forward for a demand planning role.
Eric Wilson: There is something unique in a supply chain. There’s something unique in being able to have a model change daily and being able to adapt to it. There’s something unique in supply chain that you need to be able to offer from a demand planning role, adding in analytics as another capability inside that as well. And that’s really what you’re looking to be able to do.
Kieran Chandler: Did you want to dive in, Joannes? I’m wondering if we would not need, I don’t know, supply chain scientist or something?
Joannes Vermorel: No, just kidding. It’s literally the made-up terminology of Lokad for this type of position. It’s a bit made up, but it was a way for Lokad to literally signal to the market because we received a lot of applicants, especially on the data science side, because it’s mostly what universities produce. I would say probably universities produce probably 10 data scientists for every single demand planner nowadays. It has become a big trend. And it was just to kind of put the candidates, the applicants, in the right mind that they will be first and foremost doing supply chain, not fancy advanced deep learning modeling.
Eric Wilson: That’s a good way to look at it. I mean, it’s those basic things you need in supply chain, but adding in that probabilistic thinking. Because a lot of people in supply chain, I mean the old way, were very deterministic. “I’m going to sell X number next month and I’m going to plan my whole supply chain around that.” We all know that that’s not what’s going to occur. We need to start thinking more on probability, start thinking more in ranges, think of more of those risks and opportunities. That’s where a supply chain scientist would come in, that’s where demand planning helps enable, those are the things that companies need to go to. So when you’re talking about analytics, it’s just the very beginning. When analytics becomes that buzzword in certain organizations, you can utilize it with the right thinking, right culture inside of an organization, and start changing the mindsets of supply chain, start changing the mindsets of an organization to utilize that analytics more, say, than the systems, the probabilistic thinking, things of that sort.
Kieran Chandler: Eric, you sort of said there that data science is something that’s becoming a little bit more fashionable, and that’s something I think we definitely agree with. It’s something that we’re hearing more and more about. How about the supply chain industry itself? I mean, there’s so much complexity there. Would you say that’s intimidating for someone starting out?
Eric Wilson: Personally, I think demand planning is sexy, and I believe it’s going to be the next sexy career going forward. But to that point, in the most recent surveys due to COVID, when CFOs and CEOs were interviewed, their top concerns were cash flow, when the pandemic will end, and demand planning and supply chain. So, we’ve gone from the cubicle to the boardroom. There’s a lot of attention on the supply chain now. You see newspapers and TV shows discussing supply chain, which didn’t happen a few years ago. The importance of it has elevated, and with that comes people wanting to get into the position and grow that position as well. Is it intimidating? No, I just think it was a back-office function that was done, and people didn’t understand it. People are beginning to not only understand it but also understand the importance of it now.
Kieran Chandler: We’re seeing a lot of media reporting on supply chain using different buzzwords, but there’s not so much understanding behind them. Would you say that’s also something that’s intimidating for someone starting out, Joannes?
Joannes Vermorel: As far as journalists are concerned, their complete lack of understanding of a subject has never prevented them from writing tons of content about it. However, joking aside, I’ve noticed a change in perception. My father, who used to run industrial companies, once told me that if someone was very reliable and square in their way of thinking, they would be put on the production side. If they were energetic and action-oriented, they’d be placed on the sales side. But if they were neither energetic nor reliable, they’d be put on the supply chain side. That was the mindset back then.
Fortunately, over the last two decades, many universities have started offering supply chain master’s programs that are not a joke, with top-notch professors and students. There’s more talent in the industry now than ever, but things have also become more complicated for various reasons, including compliance and globalization, which have made it even more challenging.
Kieran Chandler: How much would you say over the next couple of decades the role of analytics is going to change and how can you see that evolving?
Eric Wilson: Oh wow, I mean obviously there’s going to be a huge evolution or revolution over just the next few years. We’re seeing the necessity to become more agile, more responsive, and more predictive inside organizations. So with that, organizations are going to have to start catching up. They’re going to start having to rely on more micro-targeting consumers now. They can’t just blanket the airwaves and websites with material. They have to start targeting more specifically, and that is going to rely on good demand planning and supply chain to be able to help support those things. What we’re going to see is the democratization of data, and we already see the democratization of analytics and that function of supply chain really becoming a core support across the organization, supporting all types of functions going forward. I truly think organizations are going to become a lot flatter, and they’re going to be more dependent on analytics as the driving force of those organizations going forward.
Kieran Chandler: Joannes, we already talked about the idea of micro-targeting when we were speaking about using loyalty card data, and it’s definitely an interesting concept. How about you, what do you see as the future for the landscape of tech?
Joannes Vermorel: My perception is that when technology matures, it tends to become invisible, blending into the background. When it’s really perfected, you barely notice it anymore, although it’s more present than ever. I think the archetype for that would be the anti-spam. You have a piece of advanced machine learning that is sorting your mail all the time, and it’s very accurate. If you check your spam box, 99% of it is spam, well classified, and you don’t do anything; it just works. If you’ve been using Google Mail, Outlook, or whatever, it’s like that. When done right, very mature technology, especially on the machine learning side, disappears and does its stuff quietly and reliably, without fuss. You kind of forget it’s there, but you can keep working on improving them, doing a lot thanks to that. I would say the future of supply chain analytics tech for many organizations will probably be like that. It will be something that is driving tons of very mundane decisions, and it will not naturally capture the bandwidth of the CEO.
Kieran Chandler: Can you tell us how you see the role of artificial intelligence in supply chain optimization?
Joannes Vermorel: Direct targeting will be done automatically, just working smoothly. The workload of your warehouses, plants, stores, whatever will also be done in the background. Nobody will pay attention to those sorts of things on a day-to-day basis, except for a few specialists. Nonetheless, it will become an art of having people who are very good to remain competitive and keep improving the overall system. Lacking this sort of technology will make you uncompetitive, just like having a mailbox without an anti-spam would require you to spend your entire day sorting out your spam. Obviously, without that, it would be almost impossible to use the email. Now email is not a completely adequate analogy because you can have an anti-spam technology that is the same for millions of companies using Gmail or Outlook. Supply chain is way more diverse, so I don’t see this as a realistic market position to have one company capture the entire market because it’s just too diverse. There will still be plenty of technology, but if I had to guess one thing, it will be much more like anti-spam, mostly invisible, and impressive yet more important than ever.
Eric Wilson: To that point, you’re talking about the definition of artificial intelligence, which is an ambiguous term, but it’s anything that automates or augments a process or output. That’s what we’re pretty much talking about, where it’s going to be less of the modeling, the analytics, the technical side of it, and more of the soft sides of it.
Kieran Chandler: Anything becomes commoditized something else becomes a premium. So when your data becomes a commodity, when your modeling even becomes commoditized inside organizations because technology can help provide that almost hands-off, then what questions to ask becomes the premium or, you know, how to translate that into you know talking to the CEO, those become the premiums. And the that’s where you’re looking at being a supply chain demand planning, those people fitting those goals going forward. So great insights there.
Joannes Vermorel: Yeah, we sort of spoke about it here, this kind of idea of a supply chain being kind of completely silent, and that’s kind of the dream.
Kieran Chandler: Eric, we’ll kind of leave the last word up to you. As a final word, what are your hopes for your book, and what are kind of the skills you’d like a demand planner to gain from reading it?
Eric Wilson: Yeah, my hope for the book is it’s not a mathy book. It really gives you the introductions to machine learning, artificial intelligence, predictive analytics. It’s in the name predictive analytics for business forecasting, for supply chain, for the demand planner to help them go from you know, worry that internal data sets looking back, to more of a forward-looking, looking at external data sets, looking at new ways they can look at data, look at models that they may not have looked at before. It gives them those introductory pieces. It’s separated out between a people process analytics and technology is the way the book’s laid out. So, it doesn’t just focus on, you know, here’s an ensemble, here’s the, you know, a decision tree, here’s, you know, these models and how you do it. It gives you a little bit of taste of that, but it starts you with the people side of it, how to build a data-driven or analytical-driven organization. It looks at the technology side, what kind of companies help support, what do you need to start building, how do you get that visualization now? And it looks at that data side as well, explaining exactly what data is and how you can start utilizing inside your organization, instead of, you know, just, you know, swimming in the data lake. It really lets you understand how to bring pieces into your organization today that you can use today, and that’s really what I hope to be to do. All right.
Kieran Chandler: Brilliant. Well, we’re gonna have to wrap it up there, but thank you both for your time today.
Joannes Vermorel: Thank you.
Kieran Chandler: That’s everything for this week. Thanks very much for tuning in, and we’ll see you again in the next episode. Thanks for watching.
