The most significant transformation opportunity in managing supply chains will be the ability to know the past, present, and future of every asset across the entire supply chain. And to do that, there three things must happen:
- We need to provide every asset in the supply chain with an identity.
- We need to enable connectivity and data transfer between the assets across the supply chain.
- Manage all of these assets automatically.
Simple, It Is Not.
This is far from easy. But it's also not entirely impossible. Here's why.
Imagine you are preparing for a trip around the world. You already have your routes planned out. You have sponsors to cover all of the upcoming expenses. Following intensive online social networking, you've made sure you will find a friend at every location to greet you as you arrive. There is only one thing missing: you don't have a car that will carry you in your journey.
And then, one magical morning, as you walk out of your home, a fully equipped 4X4 truck is parked in your driveway, waiting for you. You could practically throw in your already packed duffle bag and head off to the Sahara desert to start your journey.
But then some doubts sneak in. Who is the owner of this vehicle? Has it been serviced correctly? What is the condition of the engine and gearbox? Should you be concerned about components that may break under the hard conditions of your travel, and if so, which ones?
At first sight, this truck seemed like a gift from the heavens. But now you realize that you know nothing about its condition. Should you go into the Sahara, not knowing which of its components will fail in the future? Instead of throwing your duffle bag on the back seat and heading towards the golden dunes of the Sahara, you head off back to your house to look for a proper car.
The Past, Present, And Future of Assets
Are the coffee beans you just paid extra for are really "fair trade"? Were the flowers delivered to your friend's anniversary kept at a controlled temperature during delivery as promised? Should you sign off on activating the refurbished machine in your factory? Which parts should you ship off to an oil rig - the ones scheduled for maintenance or the ones that are most likely to fail? And which will fail in the coming weeks?
Understanding the past (source, condition, handling), present, and future state of objects, products, and assets across a supply chain is a critical piece of knowledge. With such information, we know that the part of the equipment in our hands is genuine. We can understand the state in which our equipment operates. We can have an insight as to the condition of assets across our supply chain. We can better plan for the future because we will know how to optimize operations, better utilize assets, and predict failures. On the flip side, without such knowledge, the trust we have in the many parts that constitute our supply chains diminishes. Our ability to optimize our operations and better utilize our assets wanes.
Providing us with such knowledge about every piece of a physical product is a tall order. However, IoT, Machine Intelligence, and Blockchain have advanced to a point at which we can begin to address this problem. But we also need to be aware of the limitations of these technologies. We can tackle the problem of knowing the past, present, and future of some assets, across certain supply chains. The critical thing is to understand what these technologies can and can't do, and for which use-cases we can solve. Here are a few of the complexities we need to consider:
- Identity: First off, we need to be able to provide every part and asset across a supply chain with a digital identity. As soon as we manufacture a component or pick a coffee bean, we will tag it with its unique digital identity. This identity will enable us to attach information about the asset throughout its life. Providing a tangible asset with its unique identity may sound like a simple enough problem to solve, but even this first step is a tall hurdle to clear. Can't we attach to a tangible asset a QR code or an RFID tag? Not so simple. Let's consider, for example, a fish. You're about to buy a filleted fish at your local supermarket, but because you are an environmentally conscious person, you want to know the origin of the fish. Was it fished sustainably and responsibly? To provide you with such information, we need to give the fish and its parts as it changes shape across the supply chain with an identity. If you had bought a whole fish, the problem would have been more straightforward. But you opted for a fillet. We now need to provide each part of the fish with an identity as it changes its form. And the same analogy holds for medicine. For example, providing a box of pills with an identity (simple), vs. providing each capsule in the box with an identity. For cases of fraud, providing each capsule with an identity is useful but complicated. Or consider the complexity of providing a drum of oil with an identity. A tag on the drum is easy enough, but how would you tag the oil within it? For some tangible assets, it is easy to provide identity as their shape or form does not change; others change shape or form, while others like liquid or gas are challenging to tag. And without a tag of some form, be it a printed label, a sensor, or some chemical signature, we can't provide an asset with an identity.
- Connectivity: We've gotten used to seeing four full bars on our phone at all times and locations. But what works for your mobile phone may not work for the fish from our previous example. The identity tags of physical objects can be passive or active. A QR code carrying some information, or even simple RFID tags (like the one that buzzes the security gate at a shoe store when someone leaves without paying) are cheap and straightforward. But they are also passive, which means that someone has to do the heavy lifting of scanning these tags or installing readers across the supply chain. Once we’ve captured any relevant information, that good samaritan will also need to send the data to some location where we manage the identity of the asset. When you consider billions of items that are moving across the globe, it's easy to see where such a model will fail. The more sophisticated option is to attach active tags to assets. Imagine a sensor that can capture information about the fish, such as its location and the temperature of its environment, and also send it to a preferred cloud service. From a technology perspective, such tags are feasible, but cost, battery life, and size mean only assets of particular value justify such a solution. An asset that has an identity but does not have connectivity can answer for some business use cases, but this is not the silver bullet solution that cuts across all assets and entire supply chains.
- Automation: due to the scale of supply chains worldwide and the quantity of moving parts in them, automation is a must. Providing identities, managing the data each asset captures, giving different users access to this data, and making sure nobody makes faulty changes to the information are all factors that need thinking about. Across this entire solution, automation is a must, or we will end up solving small parts of the supply chain for specific use cases.
From Vision To Reality
There are many grand visions, such as Industry 4.0, Smart Homes, or Smart Cities, in which providing identity, connectivity, and automation are a must for reaching scale and potential. But in these domains, as with supply chains, we are not yet at the point in which we can identify everything, provide connectivity to every piece of the puzzle, and manage the whole operation automatically.
When we consider the abilities, and, more importantly, the limitations of technologies such as IoT, Machine Intelligence, and Blockchain, we should be encouraged and cautious. With IoT we can tag and capture relevant data on a growing number of assets. Cheaper sensors and longer battery life increase the type of assets we can now tag and enrich with an identity. But there are still cost barriers, form factor limitations, and business viability constraints. We can provide more assets with an active identity, but not all assets.
Connectivity is also improving. A 5G-enabled world is getting closer; there are LPWAN options. More assets can become connected, but again, not all assets.
When it comes to trust and identities, the usual suspect is Blockchain. This technology has had its fair ups and downs, and by now you've got it - suitable for some use cases, but not others.
To answer the question of how IoT, MI, and Blockchain can help transform supply chains is too big of a problem. We need to break it down. I would start by asking, "Which parts of a supply chain can we transform considering the limitations of these technologies?" The good news is quite a few.
Back To Basics
We are not at the point at which we can solve for the past, present, and future state of every asset across any supply chain. Our efforts instead should go to the places where such an attempt is possible. To identify these opportunities, we need to align a few stars: The business problem we are trying to solve for, the specific use case, and the available technology capabilities. We should be looking for assets and supply chains in which:
- The value of the assets justifies the effort of physical-digital tags,
- Where connectivity is feasible,
- And where intelligent automation is possible.
If we find parts in supply chains that hold true to these three basic points, we have found an opportunity to begin and transform our supply chain with the power of IoT, AI, and possibly Blockchain.