Visibility is a foundational element of any supply chain management process.
Alfred Sloan was chairman and CEO of General Motors for more than 30 years from the 1920s to the 1950s. His book, “My Years with General Motors,” explains how GM grew from a startup company in a startup industry to go on to become the largest company in the world. He explains how GM staved off bankruptcy early on by learning and adjusting. For its level of detail, insights, clarity, and quality of writing, Sloan’s book may quite possibly be the best business book ever written.
One such brush with bankruptcy, in the early 1920s, was brought on by the GM supply chain, in the form of over production. Divisions, which acted largely autonomously, ordered from factories. The divisions were notorious for over forecasting, with the belief that they could sell anything that was made. Furthermore, in those days, information time delays were measured in weeks or months. In other words, it might take weeks for orders from a division to be received by a factory. Over-production created an urgent situation that was only discovered months after the fact. As Sloan explains in his book,
“Comparison of unit sales and production figures for the four-month period October 1, 1923 to January 31, 1924, with the corresponding figures for the year before showed that our production had increased about 50 percent while our sales to the ultimate consumer had declined about 4 percent. Here the time lag entered. I did not get these figures until the first week in March 1924.”
In other words, more than a month after the close of four months of business, the CEO was finally able to compare sales to production. This situation, which created a cash flow crisis that threatened the existence of the enterprise, led to business process advances in the automotive industry. Instead of allowing individual divisions to make production decisions, GM put in place a central operation that managed production around actual consumption at the point of sale. They also limited the amount of overall inventory in their production-distribution system in an early example of a lean CONWIP (Constant Work In Process) strategy[i]. This is also an early example of a control tower (for a discussion on control towers see “What is a Control Tower?”). It’s useful to note, that while these lessons were known throughout the industry, many decades later the industry continued to push product into the sales channel, leading to similar crises.
This situation – from 100 years ago – is an example of what happens when a company operates without what today we call “visibility.” Sloan and his executives lacked timely visibility of sales, production, and inbound and channel inventories. It is also clear from this example that time is an important dimension of visibility. Visibility after the fact is not helpful in effectively managing operations.
It is interesting to note that nearly 100 years later during the early months of the COVID-19 pandemic, the US government experienced a similar lack of visibility as it tried to manage the flow of personal protection equipment, ventilators, and other critical supplies to hospitals. It started by managing flow based on demand from distribution centers and hospitals; this led to shortages and excesses. Only after it started managing flow based on actual end usage did the situation start to improve. In the retail distribution world this is known as “sell-in,” versus “sell-out.”
Over the course of the past thirty years, “visibility” has become a common term in supply chain management, and indeed, in business in general. When the Internet bubble burst in the early 2000s, it was common for CEOs to decline giving guidance, citing a lack of “visibility.” It was used so often that it became a bit of a running joke on television business channels. Certainly, in the early stages of the pandemic in 2020 many companies removed forward guidance citing a similar lack of visibility. In this case, the range of demand and supply possibilities became so large that it became prudent to suspend guidance until a future day when the fog lifted, and some semblance of “visibility” returned. During this time, CEOs became day-to-day operations managers, with their hands firmly on the steering wheel of the enterprise.
But what is visibility? It is a fungible term that is commonly used, but what is exactly meant by it? Here we apply first principles to provide clarity on the different forms of visibility and how they are applied to supply chain management.
In its broadest sense, visibility is access to information – either upon request (synchronously) or through an alert (asynchronously) - that is necessary to understand how the business is performing relative to a commitment – any commitment. Therefore, visibility is a foundational capability necessary for delivering on any commitment, and as such is a foundational capability of supply chain management.
For example, executives make commitments to budgets that include revenue and cost. These executives need timely updates on sales, production, shipments, costs, and the status of capital investments. What does timely mean? Timely means the data needs to be up to date to the extent necessary to provide some control over the process. (For a formal definition and discussion of “real-time,” see What is Real-Time?). Likewise at the operational level, a shipping manager or warehouse manager needs to know the status of delivery commitments made to customers. A production scheduler or plant manager needs to have visibility to inbound materials and outbound deliveries so that they can understand where they stand against production commitments. The operating committee of the company needs visibility to sales, production, distribution, and new product introductions. This is particularly true in today’s world, in which the sales and operations process (S&OP) has essentially evolved to a daily or continuous process.
Figure 1– Integrated Commitments Allowing for Drill-Down
Thousands of commitments are executed every day; they are linked together and in the aggregate they represent the performance of the overall business (as illustrated in the figure above). That said, visibility is a necessary but insufficient capability for delivering a commitment. Visibility must be augmented with action, particularly when commitment delivery is at risk. This is part of the control loop that is integral to any supply chain management process. We now turn to a discussion of this topic.
Foundation of Supply Chain Management
The foundational DNA of supply chain management is a control loop. This first principle posits that managing a supply chain is largely about controlling its processes to effect an outcome. At the highest level of an enterprise, the outcome to be achieved is typically expressed in financial terms such as operating profit, cash flow, and return on invested capital (ROIC). These outcomes are converted into plans, which become the objective functions for the end-to-end supply chain and its functional departments. The plans are then turned into actions at the operational level, along with specific metrics, as shown in Figure 1.
At the operational level of an enterprise, this manifests itself in things like customer commitments; for example, a commitment to deliver an order at a specific time on a specific date to a specific location. This commitment is based on inventory being available, coupled with the lead time to get the inventory from its stocking location to the location of the customer. Here enters the control loop. Once a commitment has been made, we must ensure that processes execute to deliver that commitment.
The above diagram is a general control loop from control engineering, as applied to any process to be controlled. In the context of supply chain management, the process to be controlled might be a truck moving from point A to point B. The set point is the commitment to deliver the truck load to a certain location by a certain time and day. The sensor is the visibility component. Here, the progress of the truck is monitored along its route with location data. Intelligence is added to the location data to create a predicted ETA, which is updated as the truck moves along its route. This ETA is then compared with the commitment (set point). If there is a deviation, then the shipper, carrier, and customer get visibility to it and decide what to do about it. This is the control logic. If nothing can be done about the deviation, then the customer may choose to change its downstream schedules in order to meet its own objectives (another control loop).
Visibility provides us the ability to track the commitment and its physical movement from origin to destination. This visibility is used for supply chain operations and is furthermore provided to the customer. Visibility can even include the current process stage of the order (received, preparing for shipment, loaded, in-transit, et. al.). In a reasonable sized enterprise, there are too many such commitments for humans to monitor each one individually. (However, a particularly important commitment to an important customer might be monitored directly).
Typically, commitments are made and then exceptions are raised when the commitment is at risk. Thus, machines running software are being fed with visibility information; they add intelligence to this information and then serve up exceptions and graphic visualizations to humans. Humans can then decide what to do about the exceptions. This intelligence can furthermore be programmed into, or learned by software, with certain exceptions being automatically handled. This is the controlling part of the control loop.
For example, software might tell us that the latest ETA for an order is late. Humans or machines controlling the process might look for ways by which this can be fixed. If it cannot be fixed, the new ETA is communicated with the customer. The customer might choose to start a collaboration session with the supplier. The customer now starts its own control loop, in which it processes the new ETA to understand its impact. In the case of a manufacturer, this might impact its production schedule; in the case of a distributor, this might affect its dock schedule, and its downstream commitments to its customers. Therefore, visibility is feeding multiple control loops along the order’s lifecycle.
Levels of Visibility
There are multiple levels of visibility in an enterprise. The Sloan GM scenario is an example of executive-level visibility. The above discussion about shipment ETAs is an example of operational-level visibility. A lot of the discussion regarding visibility is at the operational level. Every consumer is familiar with this type of visibility when they order from Amazon, Walmart, Target, and other retailers with large on-line presences. Furthermore, an entire software market has evolved to address the operational level visibility problem, particularly in transportation. Gartner refers to this market as the “Real-Time Transportation Visibility Platform” market.
But the visibility problem in enterprises is broader than the transportation function. Here we use the traditional Strategic-Tactical-Operational time- and decision-level model to discuss different types of visibility.
Figure 2 – Decision Hierarchy Model
Using the decision hierarchy model, visibility within each level can be summarized as follows:
Strategic-level visibility is about knowing where the company stands relative to monthly, quarterly, and yearly commitments for revenue, cost, margin, cash flow, and investment.
Tactical-level visibility is about knowing where the company stands relative to daily, weekly, and monthly commitments to plans, including production, distribution, cost, inventory, and customer service.
Operational-level visibility is about knowing where the company stands relative to hourly, daily, and weekly commitments to orders, deliveries, yields, production schedules, material schedules, distribution schedules and associated efficiencies.
A brief discussion of each level follows.
The strategic level is typically involved with growth and operational objectives, as dictated by company level budgets. This includes revenue, gross margin, operating profit, capital expenditure (CAPEX), R&D investment, cash flow, and ROI measure such as ROIC, ROCE, or economic profit. It is vital that executives have visibility to these measures through timely rollups. What is timely in this context? This could be daily or weekly, and in certain situations hourly. In retail, for example, it is critical to get continuous sales updates on days like black Friday, or in the case of Amazon, prime days. Certainly, advanced visibility to revenue and cost shortfalls is critical so that adjustments can be made when targets are at risk.
The tactical level is the purview of the operating committee of the company. What is the operating committee? The operating committee is a cross-functional body that makes key operational decisions that impact revenue, cost of goods sold, operating profit, new product introduction, capital expenditure, and cash flow; these decisions involve materials, inventory, distribution, production, labor, and assets. The committee typically includes the COO, procurement, supply chain, finance, sales, R&D, production, distribution, and customer support. (Note: a lot of these functions are typically under the supply chain organization). This is also the area where intelligent “bets” are placed. Should we carry more or less raw material based on what we are seeing? What is the probability profile for demand across our product portfolio and how do we capture the upsides and avoid the downsides?
This operating committee executes the sales and operations planning process (S&OP) (also called the integrated business planning process). Historically, this has been a monthly process, but has evolved to more of a continuous process over the course of the past five to ten years. This has been enabled by the cloud, in-memory computing, rapid scenario management, IOT, collaboration tools, and the ubiquity of smartphones.
At this level, visibility to threats to plans is critical. Plans in this context are time-phased, typically out to 12 to 18 months. Here it is possible to get advanced visibility to problems such that scenarios can be run to understand how to mitigate their impact. What is advanced visibility? Advanced visibility is notification of problems or potential problems with an amount of time that allows for some mitigation maneuverability. For example, knowing that a hurricane is going to hit the eastern part of the US in seven days allows for planned downtimes and rerouting and prepositioning of inventory.
Time-phasing also means that end-products are connected to their components through a bill of materials or an equivalent thereof. This makes visibility in the tactical category fundamentally different from operational visibility, in which you are typically dealing with the movement of a discrete part, product, material, or SKU from one place to another.
The operational level of decision making is where physical product movement takes place in the physical supply chain. Material is moved to production lines, production takes place, final products are staged and transported to distribution centers and then transported to customers. This is the level where much of the supply chain visibility discussion has been taking place. Visibility here is typically involved with location, status, arrival times, and order compliance. The primary commitments in this area are associated with orders delivered on time in the right quantity to the right location.
It is important to note that decision levels and time horizons are a useful model for discussion; in reality, they represent more of a continuum between the strategic level and the operational level. There is a lot of overlap in business processes; businesses today operate with more seamless flow of information from the operational level to the strategic level. Today it is possible to drill down from a high-level view of revenue to specific operational problems that are preventing attainment of objectives, as shown previously in Figure 1. Drill down capability is a top-to-bottom form of visibility in which executives can start with high level commitments (revenue, costs, cash flow) and drill down to a detailed level to understand how and where there are risks to achievement.
Figure 3 – Some Example Forms of Visibility
Evolution of Visibility Solutions
Most discussions on visibility are focused on its applicability to short term operational decisions – specifically the location and status of orders, and in large part their location and status in the transportation network, including over the road, sea, rail, and air. Today, there is an entire category of software solutions in this area.
Operational visibility has evolved from boundary limited and locked up in specific applications to a horizontal capability that can both be used by itself and served up to other applications (as depicted in Figure 4). For example, warehouse management software historically might provide some upstream and downstream visibility. These visibility solutions were typically “thrown in” as add-ons to the warehouse or other functional solution. However, this visibility has typically been designed to only support the warehouse management function; it has not typically been available through APIs or other means to other applications.
Figure 4 – Evolution of Operational Visibility from Function-Bound to End-to-End
Previous efforts to provide such end-to-end visibility involved expensive integration efforts that were specific to individual companies. Visibility platforms are a significant breakthrough for supply chain management that has accelerated in the past five years as IOT location and status information has become affordable and ubiquitous. It has been further enabled by smartphones, which allow virtually every employee to connect to enterprise applications for information that is pertinent to their respective jobs. The cloud has also been a key enabler; its enterprise-agnostic nature has made it particularly suitable for problem sets such as transportation, which reside between enterprises.
As shown in Figure 4, operational visibility has now been abstracted into a horizontal layer in the application architecture. This type of capability, which crosses functional, organizational, and company boundaries, is particularly well suited for the cloud.
Customer demands for higher levels of precision – in the products they buy and in how, where, and when they are delivered – has driven the need for higher levels of precision in all areas of the supply chain, including delivery time windows, production schedules, and materials. Higher levels of precision require, first and foremost, higher levels of visibility. These two forces have moved in lockstep – the need for higher levels of precision has driven the need for higher levels of visibility; visibility, in turn, has enabled higher precision, causing customers to ask for more in a self-reinforcing ratcheting up cycle.
Architecture Evolution from Database-Driven to Event-Driven
The 1980s saw the advent of packaged enterprise application software. At the center of all application software was a relational database management system (RDBMS). At that time, there was a vision that all data could be normalized and integrated through an RDBMS. This was also the dawn of enterprise resource planning systems (ERP); in this architectural paradigm, the RDBMS was at the center of the application universe and all applications (finance, HR, manufacturing, customer relationship management, supply chain) would act as satellites to this center.
In this historical approach, databases were the center of the enterprise architecture. Visibility data was entrenched in function-bound applications, tethered to databases. Getting a visibility picture across functional applications typically required custom integrations. Furthermore, data integrations were typically done through batch extract, load, and transform (ETL) functions. Therefore, visibility data typically had significant latency associated with it.
Now, visibility platforms have become an application-agnostic capability that is forming a foundational infrastructure layer in the supply chain management software stack. The location and timing of shipments along a transportation path from point A to point B or from point A to their final destination have become available through APIs, which are open and available not just to transportation applications, but to all other supply chain management applications that might have the need for such information. In the past, such information was entered manually or transported through earlier generation technologies and loaded into a database from which it cascaded to using applications through batch processes. For example, an S&OP application today can get this information directly in real-time versus getting it from a database snapshot from an ERP system or through ETL from a data warehouse. These snapshots typically have time latencies associated with them.
Visibility platforms are part of a more general architectural movement away from database centricity towards event-centricity. In this new approach, databases are simply another source of events. This new architecture is based on events and streaming through open-source software such as Apache Kafka and Pulsar. Several companies such as Confluent have built businesses on this technology.
Real-time visibility platforms are simply an SCM-specific application of this longer-term trend towards event-driven, API-based, streaming enterprise architectures. So far, end-to-end visibility solutions have been limited to the operational realm of decision making, and mostly to the transportation function (e.g., tracking and predicting the movement of goods). It remains to be seen if these platforms will take on other forms of visibility data (across the strategic-tactical-operational decision spectrum), or if other forms of pure-play visibility platforms will emerge. Today such visibility is provided by control tower, S&OP, and IBP applications.
What is End-to-End?
“End-to-end” is a term that has been used in supply chain management for at least a couple of decades. It is fashionable to describe “end-to-end” SCM solutions or “end-to-end” visibility. But what is meant by “end-to-end?”
“End-to-end” visibility typically is used to describe the ability to know the exact location and ETA of a shipment from point A to point B. In this case, point A is one end and point B is another end. In a narrow sense, point A might be a distribution center and point B might be retailer. In a broader sense, point A might be a raw material supplier or product manufacturer and point B might be the retailer. In this case, the retailer and brand manufacturer might want to know the type of materials and labor used and the factory conditions under which workers work, along with the complete buildup of CO2 emissions as the product traverses its way to the point of consumption.
Achieving end-to-end visibility increases in difficulty with the number of organizations and physical handoffs involved. Providing visibility of shipment location between a distribution center and a retailer when the distribution center and transportation function are owned by one company is much easier than a situation in which the DC is owned by one company, and the transportation function is owned by multiple companies (e.g., in a multi-modal situation). Likewise, providing visibility of a product from its origins as material to its final delivery to the ultimate end customer might involve dozens of companies, and multiple functions within those companies.
Physical and data handoffs can result in visibility “dead spots.” These spots are common in plant and warehouse yards, where trucks drop off trailers before they reach a dock; they are likewise common at railyards and ports. Operational visibility solutions are currently adding capabilities to eliminate these dead spots.
The complexity of providing visibility also depends on whether it’s being provided for a finished good or further upstream to include the components and materials of the finished goods. The latter requires a bill of materials, which adds significantly to the complexity of providing visibility. A finished good like an automobile contains 20,000 detailed parts with 2000 subassemblies that come together at the point of assembly. For products of reasonable complexity, providing visibility down to the detailed part level is significantly more challenging than simply providing visibility for a part from point A to point B. Complex products may have supplier networks that are 5 to 10 levels deep. This type of visibility requires a digital twin that comprehends a bill of materials that maps to a supply chain model incorporating supply chain operational policies.
What About Top-to-Bottom Visibility?
Of equal importance to end-to-end visibility is “top-to-bottom” visibility. For example, when there is a revenue shortfall, it is important to quickly identify the causes. The cause might be lack of inventory, a supplier shortage, a late shipment, production downtime, demand shortfall, or inability to realize expected prices. Reasonable sized businesses typically have a large portfolio of products distributed through multiple channels and sold in multiple regions. Being able to drill-down from high level financial measures to specific operational metrics is critical to maintaining alignment across the organization and to steering and controlling the business.
Companies may employ different business strategies at the intersection of product-channel-customer-geography-time. For example, for the next six months the company may decide to go after market share for Product A in the southeast region of the US through its retail network. Based on the competitive environment, it may employ a different strategy for Product A in western Europe. There may be hundreds or thousands of such intersection points in a large business. Having visibility to the performance of these strategies is critical.
Furthermore, it is critical to ensure operational policies are aligned to these segmented business strategies. For example, if the strategy for Product A is to go after market share in a certain segment or geography, then it is critical that inventory policies are updated to align with that policy. For example, having an inventory policy that minimizes cost may be at odds with the market share business strategy. Therefore, drill-down visibility is critical to guiding and managing the business.
Visibility is a foundational input into the control loop that is the DNA of any supply chain management process. As such, it is the foundational input into supply chain decision making. Knowing that a shipment is going to be on time may require no new decision making; however, knowing that a shipment is going to be late may require changes to plans in a downstream warehouse or production facility.
But visibility is more than about order status and shipment location. Visibility is about being able to see exactly what is happening in the business at a strategic, tactical, and operational level, simultaneously. Visibility has always been important in supply chain management. It has become increasingly important because of customer requirements for ever higher levels of precision in everything – from product specifications to delivery windows.
Supply chain surveys that ask practitioners their top challenges have invariably placed “visibility” among the top three. This is true whether the survey was taken in 1995, 2005, 2015 or today in 2021. Does this mean we have made no progress? No. The reason for this is that the customer keeps ratcheting up the need for greater precision. Furthermore, supply chains keep shifting, therefore the problem to be solved keeps shifting. For example, when retailers started shipping inventory between stores to “save” sales, they lost visibility to this inventory; their systems never accounted for such moves. In other words, we are solving for “X,” but “X” keeps changing; that’s why visibility remains among the top challenges cited by supply chain practitioners.
The good news is that today’s solutions for visibility are more powerful and flexible than ever and will continue to evolve towards greater levels of precision, and to account for the changing needs of customers.
[i] Wallace Hopp, Mark Spearman, “Factory Physics, Foundations of Manufacturing Management,” McGraw Hill, 1996