At AMR Research, we work with many companies across a wide range of industries on their approach to supply chain measurement. We've researched and written extensively about how to design a measurement strategy, identify what metrics to track, and implement a measurement program. We've also delved deeply into the tools and technologies that support those efforts.

An additional critical component of a company's measurement capability—and one which is often ignored—is the last step in the process: turning the data into action. This is the last mile of the journey, and often the most difficult one. It's where many companies flounder and where measurement efforts often die.

Here's the scenario: You've gone through the whole painful process of figuring out what supply chain metrics to collect—identifying which metrics are in the pot and which are out, getting the buy in, and standardizing across multiple businesses. You've also successfully accomplished two other difficult achievements: putting in place a measurement process and strategy (migration path defined, resources assigned), and collecting the data.

And then you find yourself staring at a spreadsheet full of metrics and numbers, and thinking: Now what? What do you do with them?

Whether your organization is working with a consultant or completing the task internally, comprehensive data analysis is critical to ensuring your organization can reap the benefits of all the invested effort. Most of the organizations that I've worked with assume that this is an innate talent; that if you know your business, you can somehow make sense of the data. As it turns out, this is not only untrue, but also misses a critical point: different individuals come to different interpretations and conclusions about what the data is telling them. To pull all those opinions together into one, cohesive, organizational interpretation that can then be acted upon, you need a process with rules. You also need the leadership to manage that process constructively.

The issues related to organizational change management and metrics programs are extensive, and are addressed in more detail elsewhere (for example, refer to Supply Chain Management Review's October 2006 article, “Getting to World Class Supply Chain Measurement”). Below I will share a framework that we've developed at AMR Research for successfully analyzing and interpreting measurement data to arrive at clear, actionable steps. The framework follows three steps, with the output of each leading naturally to the next one.

Step 1: Understanding Interdependencies—Using Themes and Patterns

A critical underpinning of the ability to interpret supply chain data is a clear understanding of the relationships among the metrics in a portfolio. Each metric does not sit in a vacuum. Similar to the supply chain processes they reflect, the metrics are interdependent and certain metrics drive others. Understanding the interdependencies serves to illuminate the path of root cause analysis. So rather than getting overwhelmed by a potpourri of numbers, you will have a structure in place with which to analyze results.

It's worth noting that we often find a disconnect here between concept and practice. Everyone conceptually understands that the metrics are interdependent. However, because companies are still largely organized by function, it's easy to develop tunnel vision and focus on just the metrics that relate to one function or another and lose sight of the interactions among them.

When faced with a large set of numbers, look for themes and patterns in the data. Think of the data as a narrative; your job is to reveal the final story. To do this, you need to start with a mental model of the logical groupings in the data. These groupings are the building blocks of your analysis.

There are some natural themes in a set of supply chain data. Three to start with are customer-facing metrics, supplier-facing metrics, and internal metrics. Notice that the metrics within each theme often cross organizational boundaries.

These are just three possible groupings. There are others as well, particularly further subsets of the internal metrics group. These include, for example, transportation-related metrics, warehousing-related metrics, and so on. The data can be sliced into different thematics depending on what makes sense in your particular situation and on what the specific data is revealing to you. It's helpful to visually map the metrics to the themes so you can more easily picture the inter-relationships.

Problems in any one or some combination of these metrics give you different clues about the supply chain picture, and where things might be breaking down. Looking at the metrics in this way allows you to more easily spot the interdependencies. One important aside here: to do this successfully, it is critical to let go of preconceived notions of what is going on. While it can be very difficult to do, it is important to approach the data from the perspective of curiosity, rather than being attached to a particular outcome.

Below we give examples of the three types of metrics: customer facing, supplier facing, and internal.

Customer Facing Metrics Example: Interdependencies

One of the most common patterns in the customer-facing metrics is the connection between demand forecast, perfect order, and inventory. Our research shows a strong correlation between demand forecast accuracy and the perfect order. We have further found in different analyses that companies that are better at demand forecasting have significantly lower inventories, stronger perfect order fulfillment, and shorter cash-to-cash cycle times.

Say, for example, you have high demand forecast error, high inventories, and poor perfect order performance. What does that tell you? Low perfect order, despite holding lots of inventory, means the organization either has the wrong inventory or lacks visibility into the inventory it has. This indicates an opportunity to improve demand and/or inventory visibility.

Consider another example: What if a company has poor perfect order rates but does have low inventories and low (i.e., good) demand forecast error? It's the same metrics relating with each other, but the different results tell a different story. Poor perfect order despite a good demand forecast, combined with low inventory, suggests that the company either doesn't have enough inventory, has the wrong inventory or lacks visibility into inventory. The opportunity in this company's situation lies in better translation of the good demand forecast back through the organization.

Supplier Facing Metrics Example: Interdependencies

Similar patterns can be found in the supplier-facing metrics. Consider the financial connection between supplier performance, days payables outstanding, and direct material costs. One example we often see is poor supplier performance combined with high direct material costs and short days payables outstanding. A company in this situation is receiving poor service from its suppliers, despite the fact that it is paying them a lot and quickly. Clearly, this represents an opportunity to revisit the terms and agreements of those supplier relationships.

Poor supplier performance has significant downstream implications. A company that doesn't receive materials on-time or in acceptable quality can find itself with high inbound transportation and purchasing operating costs This can be due to such factors as expediting, higher-than-normal raw material inventory as a buffer, high production schedule variance as production has to be interrupted when materials are not available as planned, and lower plant/asset utilization as a result.

Warehouse Related Metrics Example: Interdependencies

Turning to internal metrics, what about the case of a company with a high level of inaccurate shipments (a sub-component of the perfect order), and also high warehousing costs, long order-entry-to-ship time (a sub-component of order cycle time), and high inter-facility transportation costs?

One possible reason for lots of inaccurate shipments can be pick/pack issues in the warehouse. The high inter-facility transportation cost suggests the company might not have finished goods in the right location at the right time, pointing to potential issues with its distribution network design and the need for better inventory management and deployment planning. All of the associated churning could be leading to bloated warehousing costs and extra time showing up in the order-entry-to-ship metric.

We often work with companies that are outsourcing some portion of warehousing to a third-party logistics provider. In these cases, the higher costs accompanied by extra time and too many inaccurate shipments suggest the company is not getting the service it is paying for.

Note that in the discussion of these interdependency examples, we have already begun to touch on possible root causes. In fact, looking for interdependencies leads naturally to root causes, which is discussed further in Step 3.

As noted above, there are other groupings and thematics that are possible. These provide the building blocks for an effective analysis. In real-life supply chains, we rarely see just one of these building blocks. Just as all the metrics within each building block are inter-dependent, so are the building blocks themselves, reflecting the complexities of the business realities they are measuring. For example, high production schedule variance can be the result of poor supplier performance, but it can also be due to high demand variability or inherent manufacturing variability. Variability in the demand forecast can contribute to problems with internal performance such as production schedule variance. But it can also contribute to supplier and logistics provider performance issues. In the next section, we'll look at examples of how these building blocks come together.

There is one last building block to note before we move on: the use of enablers. By “enablers” I mean the applications and best practices that enable or constrain a company's business performance, as evidenced in its operational metrics. These include, for example, the use of a sales and operations planning process (S&OP); demand planning and forecasting methods and their supporting technologies; collaborative practices with customers, suppliers and logistics providers; and so on. Each of these supports different areas of the business, and their use should result in measurable business improvement. If a particular enabler yields no sign of improved operational performance, that's a pretty good indication that it is not being leveraged appropriately and thus requires immediate attention.

To identify the tradeoffs being made and analyze the root cause, it's very useful to have a visual depiction of the data. At AMR Research, we use the Hierarchy of Supply Chain Metrics (see “Getting to World Class Supply Chain Measurement”). This is a 3-tiered model of metrics based on our supply chain benchmarking research data that helps highlight interdependencies. The demand forecast is at the top of the hierarchy due to its impact on performance all the way through the supply chain, as noted earlier. Together with the perfect order, it tells you how responsive a supply chain is. To understand the overall health of a supply chain, however, we need to also look at total supply chain costs. This is where almost all of the companies we've worked with make a tradeoff: they either have a great perfect order and high costs, or vice versa. True supply chain excellence is about finding the balance between the two — a profitable perfect order.

Let's go back to the customer-facing pattern noted earlier between demand forecast, perfect order, and inventory. Consider the two companies in Exhibit 2. Both have poor demand forecast accuracy, but that is where the similarity ends. Each has made a different choice. One has chosen to keep its perfect order up at the expense of cost, and the other has chosen to keep costs down at the expense of customer responsiveness. How does that play out in additional associated metrics? To keep the perfect order high, the first company is holding a lot of inventory, which further impacts cash-to-cash. The second company, having made a different tradeoff, has a favorable working capital position. The same question applies to both companies: how can they improve on the metric that needs improvement, while holding steady in the area where they have a strength?

Consider Company A, a consumer products discrete manufacturer. Exhibit 3 shows the results of its measurement exercise. At first glance, this looks like an overwhelmingly negative picture, with lots of red all over the hierarchy. What's going on here? The company clearly has an issue on both the demand and supply sides of the equation, which requires heroics internally to produce even a reasonable perfect order to its customers.

The cost of maintaining a good perfect order despite poor demand visibility is not limited to inventory, of course. A problem with demand visibility can also require, for example, expediting to customers, which shows up in high outbound transportation costs. It can require increased movement of goods between warehouses, which leads to high warehouse operating cost, longer order entry to ship time and late shipments. It can also contribute to the low plant utilization we see in their hierarchy resulting from constant changeovers.

On the supplier side, Company A is taking longer to pay suppliers than the norm (note the good accounts payable time), and is receiving poor service in return, as evidenced by its poor supplier quality and on-time metrics. This, in turn, is affecting various downstream metrics — high raw material inventory that serves as a buffer, low plant utilization, and high purchasing operating costs due to the constant need to expedite. And because Company A was using contract manufacturers for close to half of its revenue, it was purchasing not only raw materials but finished goods as well. When the supplies are finished goods, poor supplier on-time and quality can directly affect a company's perfect order to its own customers, and can require holding added finished goods inventory as a buffer.

With some further digging, the major underlying factor contributing to the poor demand forecast came to light. Company A had grown through acquisition, and was facing greater-than-usual complexity in its supply chain — 90 percent of its revenue was coming from less than 1 percent of its customers. It was suffering from extensive item proliferation, with twice the average number of items in its product portfolio. Also, upon measuring its use of enablers, we saw that the company was engaged in less collaborative practices with its suppliers than the norm among its peers.

In contrast, consider Company B, a chemical manufacturer (see Exhibit 4). On the surface, this hierarchy doesn't look as negative, certainly in its outcomes at the top of the hierarchy. Notice the good demand forecast accuracy, low costs, and reasonable perfect order and cash-to-cash. But there's clearly something going on under the surface, as evidenced by the red at the bottom of the hierarchy.

Similar to Company A, Company B is receiving poor service from its suppliers, in this case despite high direct material costs. That poor service is having a significant downstream impact—high raw material and work-in-process inventories plus high production schedule variance due to material interruptions.

On the demand side of the picture, it looks as if the high demand forecast accuracy is helping the company keep inventories low and cash-to-cash relatively short, all of which is positive. But when we look at the details of its perfect order, we see that the stockout component is high. We also see at the bottom of the hierarchy that its order cycle time is long. While demand forecast is strong, it is not translating into as strong a perfect order as it should. And while finished goods inventory is low, the company is stocking out too often. This could mean that while Company B is able to hold less inventory, it may not be the right inventory. The bottom line: this company has a good demand signal but it's not being translated back through the supply chain.

Looking at Company B's enablers revealed almost no use of a S&OP process. On the other hand, the company was engaged in extensive collaborative practices with suppliers.

What a successful analysis of themes, interdependencies and tradeoffs reveals is where the levers are in your supply chain. The appropriate action items and their priorities then naturally emerge.

At AMR Research, we define three categories of prioritization:

• Immediate Focus: These areas present the greatest opportunity and potential benefit from improvements. These are often root cause levers which, if fixed, will have a significant positive impact on other performance areas. Examples here include improvement to the demand forecast or to supplier on-time.

• Raise the Bar: These are additional areas where improvements can be realized. Current performance in these areas may be adequate, but the area is important enough to a company's specific strategy and business goals that improving performance is worthwhile. For example, a company's new product development time may be on par, but because expansion into new products is a major business goal, performance on this metric should be brought up a notch.

• Monitor and Revisit: These areas will likely be impacted by improvements in the root cause problem areas identified in “Immediate Focus,” and therefore do not require immediate action themselves. This may include, for example, high raw material inventory which is a result of supplier performance issues.

Clearly, the primary lever—and therefore focus—for Company A is to improve its demand forecast. While many companies face supply chain complexities that they cannot control, Company A did have some opportunity to reduce and better manage some of its complexity through item rationalization and customer segmentation practices, would help with the demand forecast. Improvements in its existing S&OP process and introduction of forecasting technologies and disciplines to support the process are the additional action items that would advance this company's progress in the pursuit of excellence.

On the supplier side, the first order of business for Company A is to investigate further the primary reasons for its poor supplier performance, and determine how much of it, if any, is due to the long payment terms. Right on the heels of that should be implementation of a consistent supplier performance measurement program with scorecards to make performance readily visible and easy to address. This should be combined with a supplier segmentation program to rationalize the supplier base. Supplier constraints should be taken into account in the S&OP process, and forecasts and plans should be shared with suppliers as much as possible.

As noted earlier, some of the problem areas showing up in Company A's hierarchy could be related to issues other than demand and supply variability. For example, the combination of high warehousing costs, long order entry-to-ship time and late shipments might indicate an issue in the warehouse processes themselves. Improving demand visibility will “drain the swamp” and allow Company A to see more clearly whether there are additional issues that need to be addressed.

For Company B, it's not the quality of the demand forecast itself that's the problem—its demand forecast accuracy is already good. Company B's focus should be on improving the translation of that demand forecast all around its network, utilizing the foundation of strong connections it already has with its suppliers. First order of business, then, should be to introduce a strong S&OP process to take the great demand signal it already has and help translate it back and around its supply chain, both with external trading partners and through its own internal processes.

In addition to providing its suppliers with better forecasts, Company B will need to identify any additional reasons for the poor supplier performance and revisit its agreements with suppliers to ensure that it gets the service it is paying for.

Of course, these examples are simplified for the sake of brevity. Supply chains are complex, and there are more things happening than we can address in this short space. The point, however, is that you can use these steps to approach the results of your measurement exercise with an objective eye and clearly identify the connections between metrics, the possible root causes, and the follow-up actions.

Measurement is often viewed as a means by which to set performance targets for the organization. That's true, but it goes further than that. Analysis of the data and what it's telling you—understanding the interdependencies, identifying the tradeoffs the business is making, and revealing the possible root causes—is what allows an organization to achieve its targets. Knowing how to use the data and turn it into action is the difference between wasted effort and success.