Have you ever heard someone tell a joke where everyone laughed except for you – because you didn’t get it?
That’s how I felt when I first heard about digital transformation. Everyone seemed to know what it was. Everyone was excited about it. In supply chain management, it was all anyone could talk about. Every article discussed the importance of digitally transforming the supply chain. And every conference showcased the latest technologies available to help you do so. And yet, when I first heard the term, I wasn’t even sure what it meant.
In essence, the question I had was: If “transformation” implies a dramatic or radical change, then how does your supply chain “dramatically change” with digital transformation?
To try and answer the question, I did some research, but the results were not very clear. McKinsey, Cap Gemini and the Boston Consulting Group all suggest digital transformation is about applying digital technologies [such as Artificial Intelligence (AI), Machine Learning (ML), the Internet of Things (IoT) and Blockchain] to operational processes and creating improvements. The trouble with this definition is that it doesn’t explain what exactly changes in supply chain management—that is, what gets “transformed”—when digital technologies are adopted. The analogous conclusion I reached is that, if I wear an IoT device like a Fitbit, then I’ve “digitally transformed” my health. Unfortunately, wearing a Fitbit doesn’t actually guarantee any of your behaviors will really change. More importantly, it doesn’t mean your health will get transformed for the better.
Since there didn’t seem to be a clear definition of digital supply chain transformation available, I developed one from first principles.
Using First Principles to Understand Digital Supply Chain Transformation
The Holy Grail of supply chain management is to, as efficiently as possible, exactly match supply to real demand wherever it occurs. If digital transformation is to “transform” supply chain management, then it must do so in a way that significantly improves this primary objective.
In supply chain management, there are three key factors that impact the ability to match supply to demand:
- Demand uncertainty and the inability to accurately forecast demand
- Production uncertainties leading to changes in supply
- Lack of synchronization among supply chain partners
Behind each of these factors is a common root cause—information gaps in the supply chain. The inability to forecast demand is really an information gap between what customers want or will want and what businesses think they will want. Production uncertainties are caused by unexpected occurrances, such as yields that are different from forecasts, or factory machines that break down. These are also information gaps between what you expect to happen in a factory and what actually happens. Similarly, poor synchronization among supply chain partners is often due to partners lacking information they need, when they need it. Consequently, if digital transformation can close these supply chain information gaps, then it really can “transform” the performance of supply chains to improve supply-demand matching and therefore achieve the primary objective of supply chain management. The question is: can it close those gaps?
Closing the Demand Uncertainty Information Gap
Demand uncertainty in supply chains is traditionally managed using statistical forecasting techniques. While there are many possible algorithms, the summary of how they work looks at historical sales patterns, and then uses these to predict future demand. For example, if sales typically spike in December, then the expectation is that sales will again spike in the coming December.
The trouble with traditional forecasting methods is they can be impacted by one-time events such as economic changes, special promotions, fashion trends, a spike in social chatter or extreme weather, that affect the stability of historical sales patterns. Under these circumstances, statistical forecasting methods will struggle to provide accurate predictions.
Digital transformation can improve traditional forecasting methods in two ways. The first is to gather new data, such as sentiment information from social channels, weather inputs, economic performance or information from new IoT or Fog Computing sensors that can provide insights into customer demand. The second is to use ML to continuously “learn” from this data to determine the contributions of these factors in predicting demand. For example, ML algorithms may show that once temperatures drop below a certain number, demand significantly drops. Once these demand drivers are understood, its possible to improve forecast demand, by monitoring external weather temperatures.
Closing the Production Uncertainty Information Gap
Manufacturing production is a strong adherent to Murphy’s Law. What can go wrong on the factory floor will go wrong. Machines will break down. Inputs will not be of the quantity or quality required. Yield rates will vary. Throughput will often be different from what was planned.
In today’s supply chain, these variances are accepted as statistical anomalies that are simply managed. For example, if yields are down, then more lots are started to get the throughput required. The challenge, however, occurs when production variances are unexpected. This can lead to less than required throughput, leading to delays in shipment to customers and disruptions in the supply chain.
Digital transformation can use IoT to continuously monitor machines on the shop floor, track key performance metrics and then use predictive analytics to understand what these performance metrics mean for yield, quality or the likelihood of machine failure. On the shop floor, it can close information gaps and help you take preventative action on machines before they fail.
Closing the Supply Chain Synchronization Gap
In the kids game, Telephone, a phrase is whispered from one child to the next. Yet by the time the last child hears the phrase, it’s often very different from the starting phrase! Supply chains, interestingly, operate similarly to a game of Telephone. At one end of the supply chain, a retailer may determine a particular demand based on what end consumers are buying. This demand signals the next tier in the supply chain, which sends its own demand signal to the next tier and so on. The end result is a view of demand a few tiers into the supply chain that is very different from the original demand requirement from the retailer. The supply chain, in effect, becomes unsynchronized.
Digital transformation offers an especially useful solution to the synchronization issue with blockchain. Blockchain is a distributed ledger, with information instantly visible to all parties of the blockchain and ensures a single version of the truth – such as a single understanding of true end-customer demand - in the supply chain. This is what synchronizes all supply chain partners. Blockchain also immediately removes information latency for all supply chain partners on the same blockchain. This speeds up information flow throughout the supply chaint
The history of supply chain management, has always been about information—getting more of it, managing it better and building new capabilities with it. Supply chain management saw its first leap in efficiency with the introduction of materials requirements planning (MRP) software that managed the component information needed to manufacture products. It saw another leap with enterprise resource planning (ERP) software that created information visibility throughout the enterprise. A third leap in efficiency occurred when planning and optimization capabilities were introduced to create insights on top of the new information available from MRP and ERP systems.
Digital supply chain transformation is following this same path, creating a fourth leap in efficiency. Emerging technologies are enabling both new sources of data (from IoT and blockchain) and insights (from AI and ML) to close information gaps and better match supply to demand and, in the process, transform supply chain management.
Rahul Asthana is senior principal software engineer, Oracle.
