Companies with high-performing supply chains differentiate themselves from ordinary performers through the superior application of six management practices, McKinsey research finds. As a result, these companies enjoy lower distribution and logistics costs, better service outcomes, and better inventory performance than others do. Against a backdrop of economic uncertainty and rising supply chain risk, our research has implications for high-tech, manufacturing and assembly, packaged-goods, pharmaceutical, and retailing companies.

To study the link between supply chain performance and the underlying practices driving it, we conducted in-depth interviews with operations executives at more than 60 companies across Europe and North America. The research, conducted together with the Georgia Institute of Technology’s College of Management, assessed the performance of the respondents’ companies in more than 50 aspects of supply chain management, including business processes, corporate culture, network configurations, organizational structures, strategy, supporting infrastructure, and the capabilities of personnel. After interviewers plotted the executives’ responses on a scale of one to five (five was the highest), the results were organized into tertiles and compared with supply chain performance metrics provided by the respondents on cost, inventory, and service levels.

When we examined the relationship between the scores and performance, six broad practices emerged as significant (Exhibit 1). While no company we studied had mastered all six, organizations displaying strength across them appear to outperform competitors in service (the timeliness and completeness of customer deliveries), inventory (the ratio of inventory levels to cost of goods sold), and distribution and logistics costs (Exhibit 2). That finding belies the notion that trade-offs are inevitable in these areas.¹

To manage complexity, for instance, 50 percent of our top service performers use segmentation techniques to create distinct supply chains (in their broader supply networks) to serve different customer groups with unique products. Only 29 percent of average companies do. Notably, top companies take this approach without increasing their supply chain costs relative to average companies. Likewise, two-thirds of top companies consider supply chain design processes and product- and portfolio-development processes concurrently, versus half of average companies. Leaders also regularly examine product portfolios to eliminate complexity that doesn’t add value.

The case for developing best-practice segmentation skills is compelling. Organizations scoring among the top one-third of survey participants on this dimension are 2.5 times as likely to be leaders in inventory performance and twice as likely to be leaders in service as companies in the bottom third.

Top companies also excel in optimizing end-to-end efficiency. Ordinary companies apply cost reduction techniques within individual functions (such as transportation or manufacturing), but top ones apply lean-management tools throughout the supply chain. The companies with the lowest supply chain costs are also more likely to entrust managers with end-to-end control over them, thus increasing the likelihood that management decisions will improve the whole business and not just certain functions. These companies are also nearly three times more likely than the others to share both information (say, about customer demand) and improvement metrics across functions.

Top performers take a methodical approach to demand and production planning, as well. Using cross-functional teams empowered to make decisions ranging from revising inventory safety stocks to reallocating products to manage shortages, these companies tightly integrate forecasting, supply planning, and production-scheduling processes. They rigorously forecast performance across multiple dimensions—for instance, by customer group, product, or distribution channel. They are twice as likely as ordinary companies to use planning and performance information to adjust inventory levels and storage locations dynamically in order to minimize inventory holding costs without compromising quality of service, among other purposes. One executive commented, “We turned the corner when we started rewarding disciplined decision making rather than heroic acts made in the name of customer service.” The results? Companies scoring among the top one-third of survey respondents in this area are 4.0 times more likely than low performers to be inventory leaders, 2.6 times more likely to be cost leaders, and 1.3 times more likely to perform strongly on service.

Top companies achieve the discipline required to excel in these areas partly by improving the skills of their employees. Talent leaders are more likely than others to use functional rotation and to groom supply chain leaders through formal capability-building programs (50 percent versus 33 percent). Such programs help companies to increase the competency and awareness of leaders across the business on supply chain issues, while also ensuring that consistent approaches are used—and constantly improved—over time.

What practices don’t appear to be important? Notably, our survey suggests that investments in formal IT systems (beyond basic enterprise-resource-planning ones) don’t improve supply chain performance as much as some managers expect. Companies relying more on spreadsheets and other informal solutions were nearly twice as likely to be cost leaders—and nearly three times as likely to be inventory leaders—as companies using formal IT systems extensively. That finding supports our view that they cannot replace strong processes, capabilities, and decision making. As one executive put it, “Our systems cannot be smarter than our colleagues, or we will have problems.”

Our research also identified important interdependencies between practices: in particular, two of the six we studied require strong performance in other practices (Exhibit 3). Companies seeking a roadmap for improving the efficiency and performance of their supply chains may therefore well have different natural starting points.