Given the host of disruptive forces that companies have endured over the last few years, it comes as no surprise that supply chain resilience is attracting a lot of attention. Yet creating resilient supply chains remains a difficult challenge for most enterprises. One of the main stumbling blocks is deciding how much to invest in resilience-building programs. Without a clear financial case, such investments often fail to materialize or companies compromise with half-hearted efforts that do not achieve their resilience objectives.
The challenges of trying to make the financial calculus work fall into three key areas: the deficiencies in being able to measure resilience, the uncertainties inherent in projecting future cash flows from investments in improving resilience, and problems obtaining the data required to estimate the discounted cash flow from investments. However, these shortcomings can be addressed by changing the way resilience investments are framed and evaluated.
When evaluating the level of investment required, companies need to know how much supply chain resilience they actually need. This may seem obvious, but arriving at such an estimate is far from easy when the methods currently used to measure resilience are imprecise and fragmented. There is no single, holistic measure of resilience available today. It’s very difficult to allocate an investment in something you can’t quantify with a reasonable degree of accuracy.
Two resilience measures in common use are time to recovery (TTR), which is the time it takes to recreate lost capacity after a disruption, and time to survive (TTS), which is how long the firm can continue to meet demand after a capacity loss due to disruption. Together, these indicate whether the firm can maintain service continuity after a disruption. If TTS is longer than TTR, there will be no disruption to the flow of goods. If TTS is shorter than TTR, there will be a break in the flow of goods equal to the difference between TTS and TTR.
On the surface, this form of measurement might seem useful. However, the calculation needs to be carried out at the value-stream and then the supply-chain-node level, and even if this is achieved (in itself a challenge, as we explore later), the figure arrived at does not give guidance on what corrective resilience actions need to be taken. Although some firms (e.g., Cisco) have used a resilience assessment employing TTR and TTS in the past, this approach remains mostly notional with limited adoption.
Another approach that some have proposed is to simulate the performance of the supply chain under different conditions and derive a measure of anticipated supply chain resilience — often called a stress test. This approach requires a dedicated digital model of each supply chain (a digital twin), but this is not scalable and is heavily dependent on human input and interpretation.
In lieu of a reliable resilience measure, most practitioners focus instead on measuring some kind of risk as a starting point to measuring resilience. Here are some examples:
- A high-tech company is attempting to measure risk by using natural language processing tools to analyze social media data for chatter about breaking political and natural disaster events that may give an early warning about imminent and emerging risks to its business.
- A fashion consumer products company is focusing on measuring supplier risks such as financial stability, proximity to regions prone to natural disasters, and geographic concentration of suppliers rather than evaluating the resilience of its supply chain.
- A consumer products company has created its own automated, data-based, risk-assessment tool that considers a variety of risks (e.g., geographic, supplier, manufacturing, delivery). It is informed by external sources of data (e.g., monitoring services and financial reports) and internally generated sources. The assessment identifies priorities for attention. Then the firm uses internally generated estimates for TTR and TTS to note potential outages, which are ranked by value at risk.
In these cases, an assessment of the company’s ability to respond is made using standard resilience actions (we call them archetypes), which assume that specific risks will result in specific capacity losses and that those lost capacities can be recreated by predetermined resilience actions. Some of these actions include qualifying an additional supplier in advance, contracting an additional supplier to provide a specific amount of capacity with a specific response time, procuring and maintaining a higher volume of raw material from the supplier to cover potential outage time, and redesigning the company’s product or packaging so it can to use a different material.
While the consumer product company’s approach is the most promising of the three, it is hardly a perfect solution. For one thing, it involves lots of assumptions — for instance, that future risks will be the same as the past risks experienced (e.g., that a hurricane will inflict the same kind of damage as previous hurricanes caused). For another, it’s a long journey from risk management to resilience action — at best the bridge between the two is shaky and rests on backward-looking assumptions.
The Cash Flow Conundrum
A second problem when evaluating resilience investments is how to estimate both the timing and the amount of the cash flows involved.
The timing is unknown because it is impossible to predict exactly when a disruption will occur that will interrupt production and delivery processes. Companies sometimes estimate the probability of a risk event (disruption due to natural disaster or labor action, for example) occurring over a period of 10 or 20 years. But the net present value (NPV) will vary dramatically depending on whether the disruption occurs in year two versus year 20.
It is difficult to quantify the cash flow effects of a disruption in part because its duration (e.g., two weeks or two months), impact on the organization, and responses from competitors are unknown. The resilience investment will enable the business to continue to produce and sell when it would otherwise have been shut down. However, quantifying this cash flow benefit is tricky without knowing exactly how long a future crisis will endure,
This conundrum goes to a core problem with resilience investment: Nothing happens (i.e., there’s no loss of continuity) if a resilience investment is successful. It follows that investments to make a supply chain resilient generally don’t yield a significant benefit until they are used. For example, if the firm invests in additional inventory, that investment only provides a financial return when the inventory is used. Up to that point, the extra stock serves as a security blanket for the commercial side of the business, but it is an expensive one since it is tying up working capital.
Further complicating this cash flow uncertainty is that there are many options for increasing resilience and each one has a different cost and impact profile. For example, if the firm invests in six weeks of inventory, it ties up working capital and secures six weeks of immediate coverage. However, if it invests in securing an additional supplier, that will involve supplier-qualification costs, an upfront option payment to secure capacity (operating expenses), and likely higher unit costs (increased cost of goods sold) for the benefit of potentially unlimited coverage after ramp-up time. Which cost and impact profile is preferrable? And which ones will deliver the most bang for the resilience-investment buck?
As a result of these uncertainties, senior executives are often conflicted to sanction significant investments in measures designed to mitigate the effects of a future situation they regard as largely theoretical and, as a result, lack a clear ROI.
Holes in the Data
The key data required to develop the discounted cash flow of a supply-chain-resilience investment is difficult to access and often inaccurate. A primary problem is deriving high-quality TTR data.
Collecting this data requires a detailed analysis of each supplier, not just Tier 1 suppliers. When a firm has hundreds, if not thousands, of suppliers in its network, this is an impractical task. Moreover, a supplier’s TTR likely changes over time based on the supply and demand dynamics of materials and systems required to restore lost capacities. This variability has been extreme over recent years as suppliers have struggled to cope with huge swings in demand and the vagaries of the bullwhip effect (where a customer over-orders to compensate for supply uncertainty and this practice becomes more exaggerated as one progresses along the supply chain).
As an alternative to collecting the data from each supplier, one company has used its own subject-matter experts to estimate the TTR for each supplier (likely informed by historical experience). This method provides more control, but it is subject to potential wide variation as it may be difficult to calibrate across the experts.
How to Handle the Deficiencies
There are no miracle remedies for the three problems areas discussed above. Some work — including our own (not yet published), Cisco’s, and Resilinc’s (a firm in which one of us, Jim Rice, invested) — is going on to improve the measurement of resilience, but no clear holistic solution is on the horizon. It remains impossible to know for sure when a large-scale disruption will hit and hence to calculate the cash flow implications with precision. And the data challenges outlined above will persist for the foreseeable future.
However, there are ways to mitigate these problems and develop more effective approaches to building the financial case for resilient supply chains.
Involve other functions.
Since the options for improving resilience go beyond the supply-chain-management function, other functions need to be involved. For example, sales and operations planning and integrated business planning are often responsible for working capital investments, the finance function typically develops investment proposals in collaboration with various other functions as part of the capital-expenditure-budgeting process, and operations budgeting concentrates on operating expenses.
Recognize that investment decisions may have a strategic element.
By this we mean that the calculus for making decisions to invest in making your supply chain more resilient may not be clear enough to support an investment, but the business imperative to have response capacity warrants continuity.
We saw this in the investment that Toyota made in inventories of semiconductors after the Great Tōhoku Earthquake and subsequent tsunami devastated portions of Japan, disrupting the company’s supply chains. It recognized that a semiconductor shortage in the future would likely threaten the franchise. However, that investment required a lot of courage.
Get clear on the objectives.
For instance, are you trying to reduce the probability of losing a core capacity such as supply, transportation, or labor? Or are you trying to enable the business to respond when a crisis disrupts your supply chain? These very different objectives characterize the problem of conflating risk and resilience; and because they are different objectives, they require different actions. To address this, you must gain a deeper understanding of the need for resilience, what the different methods for building it achieve, and the resources required to put them in place.
Don’t seek perfection.
Accept that resilience-building in supply chains is not an exact science. Avoid analysis paralysis by waiting for perfect information and recognize that actions to make supply chains more resilient can be directionally on target and may not pay off for some time.
Toyota’s investment in extra semiconductor inventories didn’t pay off until nine years later, when semiconductor shortages related to the Covid-19 pandemic hit the automotive industry. It has enabled the company to ride out the storm much better than other auto makers.
Not every resilience investment will deliver a return in this way. But in today’s uncertain world, this type of investment is part of the cost of doing business.