<img src="https://secure.leadforensics.com/78049.png" style="display:none;">

The Connected Enterprise

PODCAST

ClearMetal Reflects on Data Science, Predictive Analytics, Machine Learning, and the Digital Transformation of the Global Supply Chain

Posted by Vision33 on Jul 31, 2019 3:40:39 PM

Subscribe to us on iTunes 

Full Transcript

Carl Lewis: Welcome to the Connected Enterprise podcast, where our guests share how they stay connected to their supply chain. I'm your host, Carl Lewis from Vision 33, and my guest is Adam Compain, the CEO of ClearMetal.

Carl Lewis: Adam, we met at Sapphire several months ago, and I was excited about what your company was doing. Please tell us about yourself, ClearMetal, and your role there.

Adam Compain: I’m Adam Compain, one of the founders of ClearMetal, and the CEO. ClearMetal is a predictive visibility company that exists within the supply chain, focusing on the transportation segment. Our SAS software sells to the largest retailers, manufacturers, and suppliers and helps them plan for their international freight transportation. It also helps them understand how their inventory and materials are moving around the world and collaborate with internal supply chain teams, finance, sales, and external customers. This ensures the inventory is moving optimally across the world.

Adam Compain: The problem ClearMetal is solving is that, for 30 years, supply chain operators solved their problems through physical economies of scale and brute force. That's a fantastic way to drive efficiency in a growing global trade, but in 2014, '15, and '16 we saw that trade growth slow down. More importantly, the Amazon Effect came into full swing. The heightened demands from consumers stemmed over to the heightened demands of corporate customers to where CEOs were saying, “Hey, we can’t just be bigger and tolerate what we considered a cost center supply chain. We have to make the supply chain a strategic and competitive advantage like Amazon.” Customers demand a lot more today, and what we thought of as ‘back office’ is really a driver for getting product across the world into customers’ hands.

Adam Compain: In 2015, '16, '17, the whole industry woke up this way and said we can’t just be bigger – we have to be smarter. We need advanced technology to give us visibility of what’s happening in our supply chain – how products are moving, how to make processes more efficient, how to improve to benefit customers.

Adam Compain: I spent over five years at Google in many roles before I got my MBA at Stanford. Then I went to Hong Kong to work at one of the largest container-shipping companies out of curiosity and fascination with the industry. That led to the founding of ClearMetal because I saw an industry that solved problems through scale, not data intelligence, software, and a shifting of macroeconomic trends for producing goods and meeting the heightened consumer demands a la the Amazon Effect.

Adam Compain: After Hong Kong, I met two of the top AI engineers from Stanford’s graduate school. The three of us – Will Harvey, Diego Canales, and I – founded the company, naming it ClearMetal after the premise that it's not about the metal. It's about seeing through the metal to master the underlying data. If we can master the data, we can provide a lot of value for supply chain operator businesses.

Carl Lewis: One thing that fascinates me is that you know where every ship is carrying containers on the ocean, within a reasonable distance. And that's just one example of how you're making automation and collaboration work together. People talk about it, but you guys are doing it. Can you be more specific about what ClearMetal is doing with artificial intelligence, the internet of things, and machine learning?

Adam Compain: I'll break it down into three pieces we see global logistics teams struggling with today and explain how we help them with AI.

Adam Compain: Consider a large company like Nike. Such companies must decide how far in advance to order product out of a factory in China (for example), ship it on trucks, trains, and boats to Chicago, and get it on the shelf. They're deciding that 40, 60, 80, 100 days in advance. And they're choosing the route, the carrier, and the mode of transportation upfront. To do this, they're using transit tables. Think of an old bus table that shows how long it takes to get from point A to point Z.

Adam Compain: The supply chain is dynamic and, for a particular shipment at a particular time going through a particular routing, things can fluctuate. Struggling to identify the right lead time doesn’t leave these companies with much buffer stock in the store or on the shelf. It's challenging to pick the right mode of transportation. Do I have to spend seven to ten times the cost to rush it by air freight or do I have time to ship it over the ocean? Which carrier should I use to ensure the product will arrive on time? All that is done manually or with a static Excel spreadsheet.

Adam Compain: We help these companies look at the characteristics of this day, this shipment, this routing, and what’s going on right now. What’s the most intelligent lead time, mode of transportation, and carrier to select? That brings a level of intelligence and granularity into the process that leaves them not short on the shelf and not long with excess buffer stock and inventory holding costs. We use data science and machine learning-based techniques to identify what the patterns are saying and what the probable outcome for a shipment is.

Adam Compain: The second example is a little bit more tangible where the data in the supply chain is flawed. A big brand you love can get information from its logistics providers – the ports, trucking companies, and shipping carriers – that says the shipment is out for delivery at the same time it's been delivered. If you're trying to staff labor in a warehouse or manage inbound shipments, how do you make that decision based on conflicting data? Or other information in the industry will say the ship left three times across two days. Which one was accurate?

Adam Compain: The problem stems from different dialects and pseudo standards of data that have crept up in the industry to a point where even the largest logistics companies and the biggest brands can’t make sense of the information and make intelligent decisions.

Adam Compain: We’ve found the source of truth in transportation data by leveraging the best machine learning and data science techniques so we can aggregate the information and make sense of it. We understand where a shipment is despite conflicting data. We call it canonicalized. And with information that's clean and reliable, we provide insights, analytics, and predictions to brands about when shipments will arrive, when they’ll be delayed, etc. so they can decide: Do I need to order more? Should I have left more time? Should I notify my customer that I’ll be late? Should I source product from a different factory or warehouse to make sure consumers aren’t disappointed?

Adam Compain: Kind of a long answer. Let me pause and have you guide me here.

Carl Lewis: No, that was great. It helps people understand what you're dealing with and how important it is in today's marketplace. I like that you talk about the Amazon Effect because it's like businesses are trying to give their customers, which are other businesses, the same experience a typical consumer has with Amazon. That's remarkable. And it would take this data analysis and intelligence to accomplish that.

Carl Lewis: I'm sure many businesses are looking at this, and you've probably dealt with many since you talked about 2014, '15, '16, and on. What are some of the biggest challenges those businesses have in making use of these new technologies?

Adam Compain: There are a few inherent challenges. First, you have some of the biggest, most established, and mature organizations in the world coming into an environment where they need to "digitally transform." What’s interesting is that these companies have become so mature and successful because they build great apparel or produce specialty chemicals – not because they build software and solve logistics problems, etc.

Adam Compain: One of the biggest challenges is having many companies around the world specializing in one thing – but they're faced with a new set of challenges regarding data and dynamic environments that are best handled with software. These companies don't know how to use that. There's just a natural ‘innovator’s challenge’ some of the biggest companies in the world have.

Adam Compain: Second is the challenge of making highly complex software elegant and easy to use. Think about iPhones or Google Maps. They’re sophisticated, but the user interface and experience are simple. Many companies are intimidated by terms like AI and how technology works. They're thinking about the application of technology as Star Wars versus the iPhone and Google Maps products they use daily. And there's a change management aspect to this of, “I've been using my gut instinct for 30 years. Can I trust that machine learning-based prediction? I need to see a lot of proof.”

Adam Compain: The oddest thing is we trust this technology in our daily lives, but in the corporate environment there's reluctance. I'm not sure why.

Carl Lewis: On a personal basis, we trust simplicity without asking questions of that beautiful interface and the simplicity of the software, like on our iPhones. But in business, we don't trust simplicity – we believe it must be complex.

Adam Compain: Exactly. Part of it comes from B2B marketing. There's so much value to provide and extract in an industry like this that software companies need to tout how sophisticated they are, especially when people are looking for AI and the cool, sleek stuff. I think that's the problem – companies talking about what they do, how sophisticated they are, and the machine intelligence that goes into it is pointing business customers in the wrong direction. The vocabulary is less about solving tangible problems – which ClearMetal focuses on – and more about who can say “AI!” the loudest.

Carl Lewis: When we met, you mentioned your background at Stanford, and there were others in the company and some guys from MIT. You're smart guys. What's the next big thing?

Adam Compain: In terms of the supply chain?

Carl Lewis: Yes.

Adam Compain: Google talked about AI being a fantastic copilot. It’s not meant to replace people, it's meant to complement them, and that's how we're thinking about it. The supply chain is naturally a human and collaborative task, physically moving product around the world. And the best complement to that difficult task is AI as an assistant or copilot. We're thinking about that on a philosophical level.

Adam Compain: In terms of the supply chain, what we're building toward, and where we believe the industry is going, is a dynamic, intelligent supply chain. The days of pre-laid, static plans made months in advance that you blindly follow even though they never matched up with the dynamics of where the supply chain needs to go. Instead, at every SKU level, we’ll ask, “What’s the context of what's going on right now? What’s the most intelligent decision in that context? The supply chain is naturally complex; you need a system and a capability to respond and adapt as things shift.

Adam Compain: To make that tangible: I have an order to distribute from my factory in Shenzhen to my customer in Atlanta. I want to decide – on the fly and unrelated to the previous time I did it – how should I route this cargo? What carrier should I use? What’s the best place to distribute this product? Does it make sense to source from this place or another? That kind of thing, on a true continuous learning and in-dynamic basis.

Adam Compain: There's a fantastic article that maps this well. McKinsey wrote about supply chain 4.0, which articulates much of what I'm saying.

Carl Lewis: You're not just following processes and standard procedure – you're solving problems in real time, or as close as possible.

Carl Lewis: Adam, let me ask some questions on a personal note. You're a young entrepreneur. Business communication in the last decade has primarily been through email, but that trend may be changing, especially in younger companies where other technologies, like social media, are becoming ways people communicate frequently. Do you see any modifications in your business life on a personal level?

Adam Compain: Yes, although it's not so much about social media but about blending communication mediums. I text with customers, which I find funny, and I interact with colleagues, customers, and partners in a combination of Slack, text, email, and phone. I’m amused when a 100-year-old logistics company and I are including emojis in our communication. But I like it. I think it builds relationships, and it's casual.

Carl Lewis: I think that's key for me: it's more casual, but it’s a contemporary way to tighten the relationship with customers, vendors, and other people we work with.

Carl Lewis: What are challenging parts of collaborating and staying connected with critical people outside your company? Things have sped up and continue to get faster; what things are a challenge to keep up with?

Adam Compain: I've noticed the whole Silicon Valley agile innovation thing has really gotten into business overall. We're worsening in this industry, though. We sense it. I think it's fantastic because companies get out there, experiment, and learn and big companies are working with startups, etc. There are positive elements to that.

Adam Compain: The challenge is that such behavior lends itself to the millennial easily-distracted-flakiness that’s like a cultural phenomenon right now. There's so much pressure or hype around innovation, and executives and companies are leaning into it, which we approve of. But they’re doing it without a strategy, getting distracted, and not approaching it cohesively. So those are at opposite ends of the spectrum, but both are always going to exist. I notice that the bigger companies that aren't as practiced at innovating are a bit challenged there.

Adam Compain: We've had executives at our customers say, “We're trying to innovate and build networks within Silicon Valley so we can learn, but we don't know how.” And they need help.

Carl Lewis: As I talk to people on the podcast and other places, I notice the stress of doing creative things. Businesses create automated responses and automated alerts almost for the sake of automation itself, without much thought into, “How is this perceived by the receiver of these messages?” We're trying to accomplish digitalization goals as quickly as we can, but we're not always doing the right thing on our way to success, whatever that might be.

Carl Lewis: Adam, are there parts of your business’s relationship with customers and vendors that use automated transactions?

Adam Compain: A ton. One of our largest customers was giving their customers promised delivery dates based on a basic table that was often wrong. What they've started to do is use our more dynamic predictions about how long journeys will take and, on an automated basis, give that to their customers as their own company's promised delivery dates. It’s very powerful. A supplier that can deliver product to its customer with a tighter window of delivery and on time makes that customer the supplier of choice.

Adam Compain: A second example is that thousands of people around the world spend all day trying to understand where products are and how they’re moving and where they’re delayed. Doing that used to be manual detective work – going to websites, making phone calls, receiving faxes, scouring through email. We can take a whole breadth of data, pull it into one view, and serve it intelligently to the right person at the right time at the click of a button. That’s a high-leverage output solution we've built to give logistics operators one-view search like Google and views into their supply chain.

Adam Compain: A third example is in the finance office. Our customers use our information to automate the invoicing process. Because the data is so bad in the industry, it was challenging for companies to know when their product arrived and when to invoice their customers, so there was a tie-up of cashflow. The invoice-to-payment cycle was too long. Now they say, “ClearMetal is automatically making sense of when our product arrives, so we can automate our invoicing processes off of ClearMetal's canonicalized data on when product arrives.”

Adam Compain: I could give you 10 more examples of how this is helping automate the supply chain.

Carl Lewis: No, those are excellent. I rarely get specific examples, so thank you.

Carl Lewis: Are you doing anything to help customers measure and track the effectiveness of the automation?

Adam Compain: That's the beauty of all this data. We can see how much more accurate and complete the information is. What we do with our customers is build out a very tangible ROI calculator that translates AI into solutions, problems solved, and tangible business impact.

Adam Compain: One of our competitive advantages is how well we understand the dynamics of the problem – how well the pipeline flows – so we can easily quantify tens of millions of dollars with each customer on how does something affect cash flow, customer intimacy, personnel productivity, reduced transportation expenses, holding costs, and other things.

Adam Compain: Now the comedy of that is that when you give customers new information about how they're performing – where they’ve never had information before – suddenly they ask so many more questions. I saw this in advertising, where brands were spending millions of dollars on TV advertising with no understanding of how it performed. Then they got AdWords with all the data, and suddenly they're unhappy. Really, the attribution is much better, and it's more effective, but people can see it now. I think that's a good thing.

Carl Lewis: Definitely. Well, Adam, I appreciate you spending time with me, the Connected Enterprise podcast, and our guests. I was at a conference last weekend, and several people gave the podcast kudos – the most important one being that it was short and to the point. I want to keep it that way. I look forward to checking back with you guys in a few months to see how things are going.

Carl Lewis: So, thank you very much, and thanks everyone for tuning in to the Connected Enterprise podcast. One piece of advice before we go: until next time, stay connected.

Tags: Podcast