From ship to shore. From field to fork. From boardroom to dinner table. Supply chain awareness is now firmly embedded in the public consciousness. Product scarcity has touched everyone’s life in the last few years, from semiconductors to baby formula and toilet paper, and in that most private of spaces, everyone fears the sight of a bare cardboard tube.
As businesses, globally, endeavor to shore up their supply chains, extreme weather events and record temperatures are amplifying the challenges faced by an already disrupted global infrastructure. It’s time to dive into the deep waters of shipping and logistics to find out how one of the world’s biggest firms is harnessing new technologies to adapt and change their impact on a climate that is already changing.
Anne-Sophie Zerlang Karlsen is Head of Customer Delivery, Maersk Asia Pacific. She and her colleagues in the APA region are now dealing directly with the emerging challenge of climate change, with their boots on the ground, or perhaps in the mud…
“Last year we had record low water levels in the Yangtze river, which severely hampered the ability to transport cargo from inland locations to the main ports. Heavy typhoon seasons that impact the ability to operate ports are also something we have seen in recent years. It is difficult to say exactly what is related to climate change of course, but the increase in meteorological extremes is definitely impacting global logistics as it is wider societies.”
It transpires that operating Maersk’s vast network of ports, warehouses, ships and containers during a pandemic did at least yield some unexpected benefits.
“Most [shipping] terminals had to move away from fixed berthing windows – simply because of the heavy delays across all global trades. This was a paradigm shift in liner shipping which hampered reliability significantly, requiring increased flexibility in the way we operate our business. The lessons we have learnt from the pandemic are, in reality, enabling us to handle some of these climate change challenges.”
“Most people are familiar with the Ever Given that got stuck in the Suez Canal in March 2021, but there was less media attention on the closure of Yantian port in China due to COVID cases in the terminal in May 2021. An estimated 250,000 containers were impacted by the Ever Given incident, while more than 600,000 FFEs [Forty Foot Equipment] were impacted by the port closure in China.”
Anne-Sophie’s team is working hard to return operations back to pre-pandemic efficiency levels, but as she acknowledges, it is “something that is yet to be fully restored.” Maersk’s efforts to boost supply chain reliability now require them to monitor events beyond their own operational needs, assessing potential shortages across their customers’ many other business sectors.
“Every aspect of the full end-to-end and the supply chain shifted overnight,” reflects Holly Landry, Maersk’s Chief Data Officer. “Maersk are obviously more in logistic transportation, we’re not manufacturers, but when you think about the great toilet paper shortage, that was sort of our first inkling of needing to see those insights and the data of where those shortages are.”
Economy of scale. Energy efficiency. Environmental impact improvement. The three Es of a ‘Triple E’, the 400 meter long behemoth cargo ships that are the backbone of Maersk’s global logistics operation do, of course, have a carbon footprint of their own…
“Maersk has, for over a decade, been working hard to reduce our emissions – from slower sailing speeds to more efficient vessels and cleaner fuel (low-sulfur fuels). Further to this, we have teams that constantly look at how we can improve our operations to be more efficient on CO2 footprints. It is not a few large things, but a long list of small things, like how we berth our vessels, or minimising port waiting time that, combined with the new green methanol powered vessels, will decarbonise the industry.”
At the other end of the network, on the other side of the globe is Holly Landry, Maersk’s Chief Data Officer, a Californian AI pioneer now based in Denmark at the heart of the company’s operations. It is from here that Holly is leading the charge, using big data to find those small changes. She begins to explain Maersk’s drive of multiple industries to a digitally native strategy that could revolutionize the efficiency and carbon footprint of global logistics.
“Aviation and healthcare are industries that have had three or four digital transformations over a long period of time. In global logistics, we're in our first for many facets of the supply chain – truckers outside of the US aren’t using apps in most of Europe, warehouses in South America are just beginning advanced robotics, or much of Africa where customs automation is just beginning. About a decade ago we put in the IoT on the vessels so that we could slow them down to save the emissions. From an optimization perspective, digital capabilities worked better because we know immediately when the schedule is off and can mitigate downstream disruption.”
“Maersk offers our customers an ‘Emissions Dashboard’, which allows them to get the full visibility of their global logistics emissions, across carriers and modes of transportation. In operations, we are actively involved in collecting the data for this, from the systems we use in real-time. With the help of this solution, customers can identify their emission hotspots, and the source of those hotspots - this allows them to make strategic decisions on where to start their decarbonization initiatives, and the scope of them. When using our low emission solutions, we further offer 3rd party verified certificates, stating the emission savings achieved compared to the conventional fuels/technologies.”
Real-time tracking is a common customer service experience as part of last mile e-commerce logistics, but it has its limitations when it comes to adaptive routing, as Holly explains.
“Can you make the vessel move faster? Not really. Can I move that truck faster? Probably not. A storm may be moving at five knots an hour, but the vessel is moving at maybe one. Triple Es are orders of magnitude larger than the Titanic. It's not a car. You're not turning on a dime. You're not going to reroute instantly.”
Granted, changing course on a vessel the size of a town is a tall order, but does real-time come into play once the cargo is back on shore?
“We are focused heavily on visibility for ‘exception handling’ or contingency planning. A lot of the food bound for Europe now comes from South America, so if you're shipping fruit from Peru, real-time knowledge of events at the Panama Canal gives you the ability to adjust to whatever situation is happening downstream for trucks and railroad to ensure the bananas aren’t ripe on arrival.”
Holly continues, starting to reveal the real goal of her data driven endeavors. “What has become the biggest benefit is predictability. Everyone focuses on ‘faster, faster, faster’, but then what happens if we turn around a vessel faster and the shipment gets to Copenhagen early? What if Copenhagen is not ready? It’s like an airport, what happens when your flight lands if they don't have a gate? We don't want our ports to turn it around faster and we don't want them to turn it around slower. What we want is a very predictable end-to-end, getting the container from the vessel to the railroad, to the truck at the time when they are actually ready for it. Predictability is the thing we care about the most.”
With so much data being gathered from across Maersk’s global network, Holly and her team have pooled this vast array of information into a ‘data lake’ and it is this immense resource that is providing the building blocks for the latest phase of Holly’s digitally native mission, digital twin.
“Only a quarter of my team is working on digital twins or data lake instrumentation. The other three quarters are working on advanced data products and KPIs [Key Performance Indicators] to drive even better data-driven decisions.We have seven thousand users making better decisions using our data lake.”
The data lake is already capable of providing incredible optimizations.
“Generative AI is incredibly powerful. We took years of case management, salesforce, and transaction data. We threw everything and the kitchen sink into that thing and started asking questions. Without any fine tuning it could answer accurately, immediately. An estimate that would have taken a frontline team two weeks to calculate, with consultations and reports from tender teams and contract teams can be generated in minutes.”
As Holly explains, however, real-time data has a carbon footprint. “Even as a chief data officer, I don't always believe in generating that much data. It's bad for the environment, requiring huge data centers. A lot of our data has to be transmitted by satellites, and that's hugely expensive. We've got to be really clear on what is the real-time use case.”
For Holly, this data, in combination with AI, has the power to “super-charge” our humans, provided it is the right data.
“I think we went way overboard on this data democratization idea. Less is more. We have 110,000 employees and I'm much more inclined to put the data in the right hands, with the right tools and the right automation. I think we'll start to see enterprises and companies follow suit as generative technologies give us much easier ways to interact with data and get the answers you need without having to have extra layers of chassis APIs [Application Programming Interface], reports and dashboards...We don’t want to build unnecessary software”
“Any kind of big data project is hypothesis driven. These scenarios are well beyond what the human brain can calculate. The real power of the digital twin is being able to visualize and understand the trade-offs of a given decision. You can put many different variables together and test without having to pull the cranes out in a live operation. We started with a standard digital twin simulation where we replicated a physical environment to simulate different scenarios on physical equipment.”
One might be forgiven for thinking that Holly has the monopoly on digital innovation and implementation at Maersk, but not so. A quick spin of the globe back to APA reveals how Anne-Sophie is harnessing the educational power of virtual reality to make operations that much safer.
“We are using Virtual Reality for specific task training as well as for general safety. ‘Lashing’ is the process of securing the container to the ship's structure and also to the container placed below. This procedure, using lashing rods, turnbuckles, twist-locks, etc, prevents the containers from moving from their places or falling off into the sea during rough weather or heavy winds. Carrying out this operation can be quite dangerous if not carried out correctly.”
As it turns out, Vitual Reality is incredibly effective not only for the training of individual workers, but as a method of developing wider understanding of operational procedures across entire organizations.
“In the last year we have invested in a program that simulates an environment for training on the procedures for ‘lashing’ on board a container vessel. We use the VR application both to let our seafarers train, but also for our employees on land to understand the risks and hazards at sea, which we need to take into consideration when we direct our operations. To my mind, there is a lot of potential in expanding Virtual Reality, especially in the training environment!”
It is one thing to embed sensor technology into a product during manufacture, but what does implementation look like in an already live, 24/7 operational environment? “For the first terminal that we digitally twinned, it was a year-long process to put in the IoT and get all the signals. Then we could really start tweaking because you've got to remove the variance and standardize before you can optimize or you're constantly optimizing the wrong thing.”
So how does Holly avoid optimizing the wrong thing or, indeed, risk destabilizing the existing infrastructure?
“AI is the cool new thing… except for those of us that have been in it for decades. I really hate technology window shopping. First we start with the business problem. It's a four step process. Hack it. Prove it. Ship it. Scale it. We test the viability of the tech and assess the quality of the data.
“So first we hack it and we just ask, ‘is this a thing?’ In the case of digital twins, we started with a need to increase capacity and the throughput at our hubs. Then we prove it. We put it in the hands of actual users in advance. Then we ship it. We start to move it from just one terminal and one twin to multiple terminals or warehouses. The last step is scaling…”
As Holly explains, patience is key, as scaling too early often leads to missteps.
“A lot of companies and a lot of teams move right to scale. They build a model and start running it, giving it to thousands of users. That's a terrible way of rolling something out. Between the ‘hack it’ and the ‘scale it’ is probably a year and a half long journey. This is enterprise software – building production code is not magic - with thousands of users, not a startup with a grassroots growth to scale measured in the hundreds. You have to be really patient, and give the users the time to adapt, adopt and bring them along on that journey. We run a lot of ‘a-b’ tests - experimenting with a control group using the old system and another group with the new system. You really have to take a stepwise approach and build incrementally, because at each stage you're going to get feedback you didn't expect. You want that before you try to scale it.
“If you don't go through the steps and just switch to the new system, it fails every time. It's like organ rejection.”
With the digital twin established, what was its impact on predictability? “We really set out to see if we could improve the throughput. Can we move more containers at this given terminal if we can plan better? It used to be that you would plan the day before a vessel arrives. Now you can plan three weeks in advance and you can know exactly what the labor and equipment is going to be.”
Were there any unpredictable outcomes? Holly breaks into a wry smile…
“For one of our Triple Es, there was a hypothesis that if we added more equipment we could load and unload faster. However, what was proved in the digital twin simulations is that we could actually take out equipment. We used to have 12 cranes operating around each other, each with their own upstream and downstream dependencies, so we took out four and proved that it was actually faster, because you have fewer ‘equipment clashes’.”
There are many nautical miles ahead for Holly, rolling out digital twin technology across all regions. As she continues to champion Maersk’s voyage towards digitally native, what lies beyond the horizon for the ever expanding data lake?
“I think linear supply chains are increasingly brittle, making it more complex to serve our global Customers. We have multiple modular building blocks. We have first, middle mile and last mile - ocean, customs, warehouses, rail and inland transportation. We need to start connecting those building blocks into not just a linear supply chain, but one with a circularity where we have dual sourcing and reverse global logistics to deliver what we call ‘zero touch logistics’. Customers need flexibility.
“The big dream is to execute end-to-end supply chains through the modularity of a ‘network of networks’ that can only be done by ‘the brain’. The brain is an ensemble model, mapping all the different parts of the supply chain networks and applying that to modular products. At the end of the day, this data and these capabilities need to make our employees’ and our customers’ lives easier.”
Maersk is on its journey towards a digitally native future, to “super-charge” their humans, translating into our teams’ operational predictability, to meet the challenges of climate change. As Holly attests, “before there was ESG, we were worried about emissions. We don't see it as an ‘or’, it’s an ‘and’. It's just assumed, it's like oxygen. We’re going to do all that we can. At Maersk, the impact that we can have on the planet, for everyone, it’s huge!”
For every organization to play their part to counteract climate change they will need the support of global supply chains that are, themselves, served by decarbonized shipping and logistics networks. As Holly and her team continue to roll out their AI driven digital twin technologies, delivering operational insights without interruption, colleagues such as AnSo will be able to make those small changes that have big impacts.