The context from which measurements are taken determines their value

Scorecarding applications summarize and highlight data for quick analysis

If you owned a racing car, you would make sure your pit crew kept it in top running order. Every dial and gauge would tick and glow with the optimum readings. Every mechanical part would move smoothly and quietly, gliding without friction. Every wheel would be aligned to within a fraction of an inch and every spark plug would be perfectly gapped.

Of course, you couldn’t even get this kind of car to run unless you lavished attention on its components. Similarly, a high-performance logistics facility requires high maintenance. You must establish a program to measure all the complex parts that work together to produce an efficient operation. But few companies have elevated their measurement programs above the tactical or facility level. Most measurement schemes are designed to rate, reward, or punish managers for performance on discrete operational factors. Although assessment at this level is a necessary part of operations management, today’s customer-centered business environment requires much broader, enterprise-wide standards that align strategic performance indicators with customer expectations and corporate goals. Your measurement program must provide the information needed to improve performance against key corporate success indicators, such as customer satisfaction and retention, market share, and profitability.

To be effective at the enterprise level, measurement programs must:

  • gauge results based on customers’ expectations;
  • align objectives with organizational goals;
  • provide real-time feedback to facilitate timely decision making;
  • Be easily accessed, measured and communicated; and
  • be understood and accepted within the framework of the corporate culture.

Implementing a scheme as complex as this is difficult, but competitive business pressures demand this level of corporate awareness for continued success.

Dents and scratches

Measurements themselves have little meaning. It is the context from which the measurements are taken that determines their value. For example, if your revenue increased 15% over the previous year, this may on the surface appear to be good performance. However, if your goal was to increase revenue by 30% and your competition increased its revenue by an average of 40%, then your performance would be considered unsatisfactory.

Context also affects corporate objectives. Take the case of a plant manager at a breakfast food production facility. Because he had responsibility for overall operations, production tonnage as a function of cost was used as one measure of his performance. For a long time his numbers looked great, but no one in the company could figure out why there was always such an excess of raisin bran flake cereal in inventory at year-end. It was not until someone realized that this product had the highest weight per package of anything in the line that the true picture emerged. Whenever there was slack time in the schedule, producing the raisin bran cereal was always the first choice because that made production numbers look best. However, when product spoilage and write-off costs were included, the plant was not meeting corporate production efficiency goals.

The above examples were intentionally simple for illustrative purposes. However, determining contextual value is often a much more complex process that depends on situational factors and fulfillment strategies.

Consider the order fill rate example on page 43. As shown in the table, this company has only one product, shipped to four customers. For the first order there were two hundred units in inventory, so all four orders were filled 100%. But now the stock level has dropped to 192 units, and the company faces a second round of the same quantity of orders. It can either short-ship its least important customer (customer stratification), or spread the shortage across all four customers (order allocation), both of which are perfectly valid strategies. But how do you account for the order fill rate? Under the customer stratification strategy, you could be rated as low as 75%, with three out of four orders being shipped correctly, or on a line-item basis as high as 96% (192 of the 200 units being sent out correctly). Under the order allocation method, the difference is even more marked. You can still claim the 96% success rate at the line item level, but you could also be given a 0% fulfillment measurement in that none of the shipments were completed as ordered.

After inventory accuracy, order fill rate is one of the most common measurements in use in the U.S. today (see chart, page 40). But even with its popularity, differences in the way that it is calculated go far beyond these simple issues of mathematics. The problem is that these measurements use the traditional operations-based point of view. What they do not provide is any evaluation of the impact of these fulfillment strategies on customer satisfaction and retention.

Reading the dials

Given the confusion over contextual measurement valuation, how does a company begin to develop effective rating standards? Fortunately, there are a few elements that are common to all successful measurement programs.

Customer expectations. Begin with an analysis of customer requirements and expectations. Order fill rates and on-time delivery percentages have very different success thresholds for just-in-time manufacturing customers than they do for customers replenishing basic inventory. Query customers routinely about what measurements are important to them and what level of performance constitutes success. Store this information in electronic customer profiles for comparison to actual results to determine true operational success. Design measurements so that successful results at lower levels of your organization contribute to success at the highest, customer-centric levels. Without this focus on the customer, company goals will not be met.

Corporate objectives. Measurements must align with performance expectations. Take two opposing strategies, one based on lowest cost and the other on maximizing customer satisfaction. Then look at six commonly used logistics measurements: order fill rate, on-time delivery, freight cost, inventory turns, over/short/damaged, and vendor compliance. The priority given to each of these should shift according to the goals of the company. If the company is pursuing a strategy of driving costs to an absolute minimum, then judging performance based on order fill rate would not only be frustrating to the shipping department, but actually counterproductive for the company’s stated policy. Conversely, if the goal is to maximize customer satisfaction, managing operations based on the number of inventory turns and total freight cost would be equally counterproductive.

You also have to avoid the trap of optimizing local behavior at the expense of broader goals. Consider the example of a nationwide retail company that sold two types of jewelry: high-value gold, silver, and precious gems, and several lines of low-cost imitations. The company built a new centralized warehouse to meet the needs of a large region, with the significant capital investment justified on the basis of being able to service the needs of both divisions. Understandably, the security and accountability requirements for the storage of the precious materials drove the cost per square foot to a premium level. Soon, the logistics directors responsible for the low-cost lines realized that they could cut operating costs by moving out of the facility and into local public warehouses. By doing this, they were able to make their numbers look good, but the warehousing costs for the company as a whole increased considerably. So, in order for an enterprise-wide measurement program to succeed, management must acknowledge that to optimize the company’s operations, individual components may need to operate at sub-optimal levels. The managers of those operations should be evaluated on factors that support corporate objectives rather than local optimization.

Timeliness. The third measurement success factor is timeliness. Information received too late to correct the problem before it materially degrades performance has little value. Unfortunately, most measurement programs are historical in nature. This is primarily due to the monumental task of collecting measurement data and rolling it up for corporate review. Usually this includes time-consuming processes such as manual input, batch reports, and mail transmission, all of which are fine for detecting chronic problems and for performance trend analysis. But customer satisfaction can be ruined before this approach ever discloses problems.

The combination of the pervasiveness of Internet access and new measurement scorecarding applications are alleviating the information capture and dissemination problem. Now measurement data can be captured simultaneously from sites around the globe and be transmitted in near-real-time to management at regional and corporate centers. Scorecarding applications summarize and highlight the data for quick analysis and provide drill-down capabilities for specific detailed review at any level of granularity. Thus, if the order fill rate at your warehouse in Atlanta falls below an acceptable level for an important client, quick online analysis may determine that the cause is a shortage of a key component that is readily available at your distribution center in Raleigh and can be trucked there overnight. The ability to detect and analyze the problem immediately, locate alternatives, and take corrective action turns this potential customer satisfaction problem into a non-issue.

Fix the undercarriage

You can’t implement a measurement program whose underpinnings are not valid. First, the proposed measurements must be quantitative. Statements like “increasing the level of customer support,” a qualitative standard, need to be translated into something that can be tracked on a numerical basis. For example, one e-commerce fulfillment company measured the number of rings the customer heard before being connected to a service rep. The company also measured how many times that agent was representing the geographic region appropriate for that call. These measurements took the qualitative concept of improved customer support and evaluated results with quantifiable criteria.

Second, the performance category must be easy to measure. If the measurement is arcane and buried in the minutiae of the operation, retrieving the data can be more trouble than it is worth. In the example above, all of the information was found in the automated call record of the telephone system. It was an easy matter of manipulating data in existing tables to build the required reports. Automation, such as radio frequency ID systems, enables rapid data capture.

Third, the measurement program must take the company culture into account. If your firm has historically used measurements as a stick to prod managers to perform better, it is highly doubtful that any new program will succeed without significant reorientation and communication of objectives. Using a measurement program to further organizational goals through problem detection or correction and crisis avoidance requires a much different management mindset than traditional rate, reward, or punish programs do.

Fill ‘er Up

Order Fill Rate
Customer Available
A B C D Inventory
First order 50 50 50 50 200
Second order 50 50 50 50 192
Customer stratification 42 50 50 50 192
Order allocation 48 48 48 48 192

Another cultural issue that can defeat the program is when peer pressure keeps productivity at what is perceived as a “comfortable” level. In this environment, high achievers quickly become visible and the group will work to keep their productivity down to the range of the group’s norm. Therefore, incentive/reward programs must align with corporate objectives rather than activity levels.

Road maps

The continued emergence of the customer-centric business model will require that measurement programs move beyond simply reporting results. They have to analyze the results, determine where problems or incipient problems exist based on predefined thresholds, notify all interested parties, and initiate corrective action. This convergence of automated electronic response with traditional logistics fulfillment is termed “digital logistics.”

In this context the measurement program becomes part of a broader logistics command and control framework that expands measurements from a tactical, facility-based reporting tool to an enterprise-wide strategic function for furthering corporate objectives.

Digital logistics will make it possible for logistics managers to track operations globally and provide management information on a timely and easily digested basis. This will require new and more meaningful measurements that will reflect the efficiency of the entire network. Nowhere is this better illustrated than in the rise of the “perfect order” measurement.

What is a perfect order? It is generally defined as “an order that is shipped complete without unauthorized substitutions to the right location at the right time at an acceptable cost, all on the first shipment.” Obviously, to track this kind of performance, a great deal of data needs to be obtained from a number of disparate systems. Emerging measurement systems incorporate data on orders from procurement systems and fulfillment detail from warehouse management systems, as well as delivery costs and times from carriers.

The important point is that instant access to logistical information from the entire supply chain will have a profound effect on the way that performance measurements will be used. Rather than simply reporting what has happened, they will be the basis for solving problems before the customer is adversely affected. For example, if the carrier reports a delay in delivery because of equipment problems and the procurement system specifies that the order is required by nine o’clock the following morning, the digital logistics system can automatically notify the carrier to send another truck to complete the shipment on time. The perfect order is preserved because the measurements are used for something more than reporting, that is, they are part of the larger digital logistics framework.

Accurate and up-to-date logistics measurements, combined with automated response systems, will also provide a foundation for powerful simulation and optimization capabilities that will enable marketers to continuously fine-tune their logistics processes. This will result in improved customer service and provide significant competitive differentiation.

Bill Petersen is director of the Logistics Operations Analysis program for McHugh Software International in Waukesha, WI. He has over twenty years of experience in addressing critical issues for supply chain management. He can be contacted by phone at (262) 317-2566 and by e-mail at [email protected].