Whatever You Do, Start Measuring

Aug. 14, 2007
Every supply chain needs regular maintenance, but it's hard to change the oil when the engine's running. The complex maze of mechanical systems, management

Every supply chain needs regular maintenance, but it's hard to change the oil when the engine's running. The complex maze of mechanical systems, management systems and information on what's happening inside a supply chain is hard to navigate. A fault of the design may also be the way each component of a supply chain exists within or actually creates its own silo.

It's truly an understatement to say that measuring supply chain performance is both critical and complex, but every supply chain needs benchmarks that are visible at all levels along the supply chain to ensure top performance. From a truly big-picture perspective, Karen Butner, global supply chain management lead for IBM's Institute for Business Value, says we need to measure from product innovation and design through the logistics and supply chain functions and on into service and product life cycle. There's very little good measurement in this last part, she says.

People get hungry for the benchmarks, says Butner, but you must simultaneously look at management practices along with those benchmarks. We tend to look at the companies we want to benchmark to compare lead times, logistics costs as a percentage of sales and other measures without looking at those other companies' management practices—what they're doing differently. And without doing both, it's an apples-and-oranges comparison.

Working with a number of organizations and models over the years, Butner says the IBM think tank has come up with three levels of measures. They've identified 20 key measures at the executive level, about 50 management-level measurements, and 120 at the operations level.

At the executive level, she says, they're looking at things like cash-to-cash cycle, return on assets, and other measures across the entire supply chain. Drilling down to the management level, the measurements become more specific. And, in operations, they are focused at the functional level.

Typically, companies have a strategy or executive level goals such as reducing logistics costs by a percentage, but those goals, and especially the thresholds they want to reach, don't trickle down, says Butner. It isn't just a matter of the goals not being communicated; sometimes it is an issue of making those goals meaningful to other levels. What does a customer service goal mean to a lift truck driver in the warehouse? It's meaningful if it's communicated and accompanied by some education, explains Butner. Clearly, when a careless lift truck driver runs a fork through a carton of product that is outbound to a major customer, that has a direct impact on customer service. Posting damage numbers might be one way to help translate the larger customer service goal into a meaningful benchmark for warehouse workers.

It doesn't matter what the number of benchmarks is, says Butner, what's important is to get started with some measurements and make them visible. Whether it's five or 10 or 15 top benchmarks, establish the measures, communicate them, update them, refine and improve, expand and add benchmarks, but get people thinking about them and seeing them.

One of the important benefits of communicating goals and benchmarks is that different functions can see how changes affect other areas. "A cost reduction here could quickly escalate a cost somewhere else if you don't map performance across the different areas and measure simultaneously almost daily to look at trade offs and what effect they will have," says Butner.

Another facet of the visibility that comes with communicating benchmarks and performance is a better understanding of how the benchmarks translate at different levels. A high-level customer service goal and measurement at the executive level may carry an element of on-time delivery for a management level goal. As that works down through the organization, at the operations level, important factors might be meeting delivery appointments at customer distribution centers. This carries through to driver wait time and down time, as well as the customer's labor scheduling in receiving—all meaningful benchmarks.

Butner calls it "dashboarding," but she says that doesn't mean it has to be a piece of technology, just a tool that provides some view of critical measures that everyone can see. As they expand and improve, the benchmarks become even better tools.

Continuing on the subject of trade offs, she notes that the focus has been on inventories, but Butner expands that to say you have to make sure you have the right inventory in the right place. She urges segmenting—the more you segment, the better handle you have on what is going on. Look at your A, B, and C items she suggests. Don't just view them as your high-margin, fast moving A items or your low-margin fast moving C's; consider the impact they have on other critical measures like customer service.

Going further, she says, look at your customers the same way and segment them. Who are the A customers and the B customers? "We all know there's a difference between overnight delivery and three-day delivery and we accept that," she offers by way of example. Trade offs help to set thresholds in inventory, customer service and other areas. Start measuring profitability along with segmenting, she says. You may start to find that what you gave away here, you more than gained back there. The shipping costs you waived for a customer may have translated into higher purchasing volumes from that customer. Without the "dashboard" communication, those trade offs might not be as visible—or the conversation that reveals what is really happening may not occur. This is a different attitude from just drilling down in a cost area, she notes. In that case, an across-the-board cut in transportation cost could have a negative effect on some of your best customers.

Everyone has to see this every day, no matter where you sit, she says. Put the numbers on a dashboard or in whatever incentive or performance objectives the company has. Directors and managers should all be looking at the same numbers. "Changes happen; communication happens, and it's not mandated [as a straight cost cut may be]." This more corporate or holistic view Butner describes takes time. But when it starts to take hold, there is a shift in attitude and in practice.

Take this approach down through the organization then to your partners, says Butner. If you're using a third-party logistics provider (3PL), they have to be onboard with what you're trying to achieve. By implication, they have to be party to the benchmarks—they need to see where things stand and their impact on the results.

To get things started, begin with the executive level. Get everyone on board and then take it to the management level, says Butner. Just as she said to start with a few measurements and get them out to everyone, it might be necessary to begin by highlighting a particular area. You could start with one business unit or begin geographically so you get your biggest bang. Then roll it out to a wider area.

The biggest problem is not with getting the executive level onboard. It's not even with pushing the benchmarking down from executive goals to the managerial level. The real problem is getting down to the operations level—the detail level.

Things are happening in your supply chain real time, and the dashboard approach Butner describes has to keep expanding to reach every level of the supply chain, it has to keep getting better, and it has to be more real-time itself. The benchmarks have to be meaningful and tied to goals. Don't wait to build it, says Butner, get it out there, even if it's a weekly e-mail with three or four items on it.

John Perkins, senior architect for Forte Industries, is on the front line when it comes to monitoring and measuring results in operations. His perspective on the automated distribution center offers some parallels to management structure and highlights ways the mechanized and nonmechanized benchmarks operate.

Perkins describes three levels of systems in the warehouse: the warehouse management system (WMS), the pick system, and the warehouse control systems (WCS).

The WMS takes orders, processes inventory, handles replenishment, and does wave planning. The pick system directs and measures personnel on items picked, error rates, etc. And the WCS manages the automated systems, mapping and moving goods through the warehouse.

One of the first challenges is that very few companies operate systems from a single supplier on all levels. Add to that the impact of acquisitions, expansions, and other changes in network make up and the problem multiplies with each warehouse or distribution center in the network. Selecting a best-of-breed and pulling the plug on all the rest is very unlikely. The expense and risk to such a transition would keep many companies from even considering such a bold move.

When the systems are tied together, says Perkins, the WMS will be directing metrics such as stock keeping units (SKUs), orders, orders shipped per day or shift by type and quantity. The pick system will measure quantity picked per zone, quantity picked per shift, individual worker performance, etc. A problem that surfaces almost immediately is that these systems all use different units of measure.

The WCS is tracking cartons moved by each mechanical system. The pick system is reporting item-level activity. And the WMS is looking at orders and SKUs. When it all communicates with the transportation management system (TMS), the TMS wants pallets or shipments—eventually containers and truckloads.

Finding a common unit of measure to establish key performance indicators is a challenge. One company, says Perkins, chose to measure product shipped per employee working that day or that shift.

Each of the systems has incredible metrics and data on what is coming in, what is going out and at what cost. The problem is in trying to take advantage of what each system knows.

This points to a larger issue, that people are often working towards goals that are in contention with one another. On the receiving dock, the goal may be to get the dock cleared and have everything put away as quickly as possible, using the least amount of labor hours. But at the same time receiving is storing items in warehouse locations, some of those receipts may be needed in the pick module so the shipping side can see them and use them to fill orders. If the pick module isn't immediately available for put-away, that would mean the goods would have to be held in the receiving area. Clearly, the priorities of the two distribution operations are in conflict.

Such efforts to optimize individual operations or functions can easily miss the point. If you don't know where the bottlenecks are, says Perkins, you don't know what to optimize. There is very little bidirectional communication between the systems or the functions, Perkins points out. Communication between the systems tends to be linear—usually flowing down from the WMS.

Either those management people in charge of the functions will have to take the metrics out of those subsystems and coordinate the metrics to identify the bottlenecks, says Perkins, or those subsystems will have to communicate bi-directionally so information can be shared up to someone to make a more intelligent, informed decision. Basically, there has to be communication in both directions on what is happening in every system and subsystem whether it is done by people or automated.

Summing up, Perkins points to three critical needs: Common units of measure, Bi-directional communication, And a common data exchange.