An intellectually honest approach to DC design

More volume, more speed, more SKUs (stock-keeping units), more accuracy, more channels. … Supply chain leaders are facing increasing pressure from the business to deliver more from their distribution centers, and many are turning to highly automated solutions as a reasonable response to these demands. 

Afterall, automated solutions can provide many benefits such as:

• improving productivity and reducing cost,
• enabling safe social distancing by eliminating crowded picking and packing areas,
• enabling product segmentation strategies (such as essential vs. nonessential products), 
• reducing reliance on experienced labor (automation can help simplify tasks for employees and create a shorter learning curve), and
• helping to meet the expectations of more stringent service-level agreements (SLAs) and value-added processes.

But is automation the right answer for your particular operation? And if so, how much automation do you need? 

DC automation projects represent one of the largest investments that a company can incur, and, as a result, these projects can become highly visible both inside and outside of the organization. For this reason, it is imperative that the project not only have a business case with sound financials but also go through a thorough design process—one that evaluates alternative options, defines a realistic baseline scenario, considers holistic business impacts, and tests for resilience. We call this an “intellectually honest” approach to design.

What makes a design intellectually honest? 

During the approval process for a capital project, supply chain leaders are typically prepared to answer questions about scope, financials, and timeline. But they should also be ready to answer another set of questions to ascertain if the project can “hold water.” In other words, if circumstances change, will the project still be able to produce a return on investment? Essentially, this line of questioning involves determining if the design approach is intellectually honest.

Here are the four characteristics of an intellectually honest design approach:

• Incremental: Evaluating a solution solely against the current state might provide an appealing cost-benefit trade-off, but it could be deceiving if you don’t consider other intermediate solutions. Looking at alternative, more incremental solutions could reveal that higher investment levels would produce diminishing returns.
• Realistic: The baseline scenarios that you use to evaluate the benefits of any capital project should be representative of the DC’s current performance levels and its continuous improvement culture. It’s also important to consider the amount of change management required to achieve the intended results of a new technology. The shift from a manual operation to a highly automated one, for example, requires significantly more change management than the leap from a mechanized environment to fully automated one. In short, you need to be honest about whether your people are properly skilled, trained, and prepared to make the type of changes required for the proposed project.
• Holistic: The largest line items on a distribution facility’s profit and loss (P&L) statement are typically labor and building-related expenses. But that does not mean that the automation justification should be limited to headcount and footprint considerations. Such a move would ignore the broader impacts to the business, such as whether a more automated operation could help with customer retention or enable more profitable distribution channels.
• Flexible: It is important to test the boundaries of the design by changing specific parameters, like labor cost, growth, and order profile (for example, fewer single-unit orders), and assessing the impact of those changes to the business case. Is the design flexible enough to adapt to different business conditions?

To elaborate on this concept of an intellectually honest approach, let’s use a case study involving a food and beverage company that is looking to consolidate three existing manual DCs with limited capacity into one automated DC. This company distributes through three different channels: retailers, drop ship for its retailers’ e-commerce orders, and its own e-commerce operation. It is currently using a manual, cart-based picking process and is considering implementing a more automated solution to support its growth and improve productivity and service level. In the following sections, we will walk the case example through the four characteristics of an intellectually honest design approach.

Incremental

A simplistic justification approach to DC design would compare the benefits of a technology with the current or base scenario. An intellectually honest design approach, however, looks for the solution with the lowest investment possible that meets the company’s business goals, and then compares that solution to the incremental benefits and investments of subsequent, more expensive alternatives.

Using the intellectually honest approach, our food and beverage company is considering two alternative picking technologies, which can be seen in Figure 1. One solution involves using a highly automated goods-to-person system. The other would utilize goods-to-person for a subset of slower-moving SKUs and a pick module for faster-moving ones.

[Figure 1] Alternatives considered for picking at the new automated DC
Enlarge this image

When the highly automated Solution A is compared to a baseline manual operation, it has a very attractive payback (see Figure 2). The boost in productivity and much smaller footprint generates significant savings that results in a 4-year simple payback for the $11.5 million investment.

[Figure 2] Example of payback comparison for alternative picking solutions
Enlarge this image

However, when compared to Solution B (an incremental alternative), the payback timeline of Solution A is almost twice as long (7.6 years). While Solution A offers higher productivity and footprint improvements, the blended Solution B approach does a better job of balancing desired business results and investment.

Even though the 4-year payback of Solution A when using the base approach is mathematically correct, it does not tell the whole story because it ignores the other viable and less expensive option in Solution B. 

Realistic 

Another common mistake that companies make when assessing automation projects is assuming that they will continue to operate at the same current “base” scenario. In today’s fluid environment, the status quo of a current state is hardly a representative reference point, so resulting comparisons would be flawed and hard to justify.

A flawed baseline scenario could misrepresent the business case in either direction. First, assuming that the current performance and profitability is sustainable without investment could be overly optimistic and under-represent the benefits of the proposed automation. On the other hand, assuming excessive performance declines or unnecessary investments in the baseline could potentially skew the numbers to make the business case look unjustifiably better than it is. 

Defining realistic baseline scenarios can be a difficult task, but there are certain elements that help:

• Representative: Consider whether the baseline scenario is truly representative of how your operation would respond to changing conditions. Most DC managers are very resourceful when it comes to adapting their operations to changing conditions. The baseline should reflect on-going initiatives that are consistent with the current continuous improvement culture, such as improving productivity through slotting changes or voice-directed picking.
• Change management: It’s also beneficial to consider the changes required to successfully adopt a new technology and achieve the intended results. This could include changes to IT infrastructure, reskilling maintenance technicians, and new training curriculum. 
• Viable: To make an “apples to apples” comparison, the baseline scenario should meet the business goals or include the cost of not meeting them.

If these three characteristics are met, it will make the business case credible and help to gain buy-in from stakeholders and decision makers.

Let’s consider what baseline scenario should be used for our case example involving the food and beverage company. Figure 3 presents three different baseline scenarios: 1) maintaining the current three DC network with the DCs possessing the same limited capacity, 2) expanding the distribution network by adding another manual DC, and 3) consolidating the three DCs into one DC that still operates manually. The figure also shows the key metrics that would need to be considered for each scenario. Different scenarios would require some differences in metrics.

[Figure 3] What is the most realistic baseline scenario?
Enlarge this image

The figure illustrates how the business case could vary depending on the baseline scenario that is considered. For example, while baseline scenario 1 represents current operations, it also creates some comparison challenges. This scenario limits the volume and revenue growth that the business case could consider, while it also assumes that current revenue and profitability levels will continue. Furthermore, the scenario is better suited for comparing three separate DCs to one consolidated DC than for comparing manual versus automated operations.

Baseline Scenario 2 does a better job of meeting expectations for volume and growth. But it still makes it challenging to accurately compare an automated solution versus a manual one, as it would be difficult to factor in the differences in cost structure and ability to meet other future requirements such as changes in the number of SKUs or to service level agreements.

Scenario number 3 provides the best reference point for most of the automation. Nevertheless, exploring different scenarios was important to determine the broader impact to the business that the project entails.

Holistic

It’s a common practice to use only the improvements to a DC’s own P&L for the financial justification for automation. As a result, you may hear statements like, “this equipment should pay for itself with better productivity.” While this may be a mathematically correct statement, it may not necessarily be representative of the full value that the automation could bring to the organization. Implementing an automated distribution center operation will have a significant impact on the business that goes beyond the “four walls” and will involve many different stakeholders.

A more holistic approach to the business case would include these outside impacts and assess the benefits to the entire business of adding a new or improved capability in the DC.

For our case study, there are many factors that should be considered. The two main “inside the four walls” impacts are:

• Labor costs: The business case should assess how the different types of automation would reduce the direct labor associated with the different fulfillment processes. It should also acknowledge how it is increasingly difficult to find and manage a large labor force.
• Building costs: The business case should also assess whether the automation will decrease the footprint required for the storage and fulfillment processes. You could argue that the building is a “sunk cost.” However, by freeing up space in the distribution center, the DC may be able to support more growth or fulfillment operations for other business units.

There were also four “outside of the four walls” impacts that should be considered:

• New distribution channels: The business case should assess whether the automation would allow the DC to serve new distribution channels. Different distribution channels often need completely different fulfillment processes. For example, direct-to-consumer distribution often requires companies to pick the same product that they sell in retail stores but in different units of measure—“eaches” instead of cartons or pallets. The business case should assess whether the automation will help the business better handle these new processes.
• Service responsiveness: Next-day or same-day order fulfillment is becoming the norm to compete. However, to execute to these standards, companies often have to double their required throughput capacity. Automation can help companies increase their responsiveness and keep up with these demands. This ability should be included in any automation business case presentation.
• Inventory reduction: Will the new automation allow the company to centralize inventory from several facilities as well as handle more frequent and smaller shipments? If so, the implementation could yield significant working capital reductions, which should be noted in the business case.
• Entry barrier/customer retention: Does this automation project ensure that your company will have a reliable distribution infrastructure with best-in-class technology? Such a capability will help you attract and secure business in the future and should be an important consideration when weighing the pros and cons of a large-scale automation project.

Figure 4 outlines the specific metrics and business impact for each of the factors listed above.

[Figure 4] Some factors to consider for the business case
Enlarge this image

Flexibility

This past year and a half have shown us an extreme example of changing conditions. The shift from retail store sales to online orders during the pandemic meant DCs had to go from shipping cartons with large quantities of product to bags with single units. While the pandemic was an unusual situation, changing conditions are not. 

An important part of an intellectually honest design process, then, is to test the boundaries of the proposed solution and understand how resilient it is to change. In other words, if conditions change, will the proposed solution remain viable, or will the anticipated savings and benefits disappear?

Let’s look at three conditions to consider when assessing how flexible an automation solution may be.

Technology thresholds: It’s important to check the throughput rates that the technology can handle and how much it can be scaled up. Knowing where this threshold lies can help you determine a point at which the technology may not work or may become a hindrance to the operation. In this respect, autonomous robotic solutions often have the advantage over fixed material handling equipment (MHE). That’s because robot solutions can often be scaled up simply by adding more units, whereas scaling up MHE solutions might require more disruptive installations.

Labor availability and cost: If labor becomes scarce or wages increase beyond a certain threshold, automation may be better justified. Understanding those tipping points should be a key part of making decisions about the use of automation.

Change in growth and/or order profiles: Growth projection will always have variations, particularly when there are several channels. You need to factor in that certain channels or order profiles (such as small, direct-to-consumer orders) are more labor intensive than others. 

By performing a sensitivity analysis over certain metrics, companies can determine how sensitive the business case is to changing conditions. The sensitivity analysis can identify “guardrails” that define the conditions under which the business case is favorable. In our case study (see Figure 5), we did a sensitivity analysis for labor cost and overall growth to see if the more automated Solution A would make more sense than Solution B’s blended approach under different conditions. The analysis helped to show that:

• The payback for Solution A is always greater than five years no matter what the change in labor cost and/or growth. 
• If growth slows, the payback for the more automated solution would stretch to more than 10 years.
• Even if labor costs increase, the payback for Solution A would still only range from six to nine years.

[Figure 5] Sensitivity analysis for labor cost and overall growth
Enlarge this image

The sensitivity analysis also showed that Solution B’s more targeted automation approach is more resilient to changes in labor cost and growth than Solution A. Even if growth and/or labor costs came in below current levels, Solution B still showed a four-year payback period versus manual operations (not depicted in the figure). Therefore, the sensitivity analysis helped to determine that Solution B was the right choice even under a reasonable uncertainty over labor cost and growth.

An honest answer

Frankly, it’s an exciting time to consider automated solutions for your distribution operations. Both mature and emerging technologies can provide attractive alternatives to address different operational needs.

But it is important to make sure that this excitement and enthusiasm for the power of technology does not run away with itself. Many factors go into a distribution center automation project, and there is no silver bullet to meet a company's business goals. A strong business case with sound financials is a must for any new capital expenditure project, but it is also important to take it to the next step and submit a proposal with an intellectually honest design approach.