Optimizing Purchasing at Procter & Gamble

The Problem

Procter & Gamble (P&G), the consumer products giant, seeks to operate a “consumer-driven supply network.” But the complexity inherent in sourcing materials for a global business with nearly 300 brands produced by over 100,000 people and thousands of suppliers in nearly 80 countries for millions of consumers was creating costly inefficiencies. P&G’s original sourcing strategy was to simplify the process by limiting the scope of auction events. P&G’s sourcing managers created groups of products for suppliers to bid on, limiting the amount and type of input suppliers could provide in a sourcing event. The effect was forcing suppliers to squeeze their margins instead of enabling them to bid competitively with their own cost efficiencies. In addition, the system made it difficult for P&G to balance price considerations with such “side constraints” as product quality, supplier reliability and delivery schedules.

The Analytics Solution

P&G tapped an optimization-based, two-sided purchasing system known as “expressive competition.” Sellers bid on self-constructed packages of items instead of being limited to bidding on predetermined lots. Suppliers can make creative proposals including bundled offers, conditional volume discounts and innovative alternative item specifications to exploit cost efficiencies. They also can incorporate a wide variety of constraints, such as delivery time options, colors and other variables.

Analyzing such complex bids is performed by the “expressive allocation evaluation” side of P&G’s “expressive competition” sourcing system. It enables P&G to express preferences with rich, precise and compact language. P&G can, for example, specify that a particular percentage of purchases be allocated to minority contractors, or give an existing supplier a specified degree of advantage over new, untested suppliers, or limit the number of suppliers from a particular geographic region. The system is designed for users without any expertise in optimization techniques. Users can express allocation preferences using language like “I don’t want any one supplier to win more than 15%” (to maintain a diverse supply chain). In addition, the system allows P&G’s purchasers to quantify the cost of such preferences.

The Value

This analytics-based system was used by P&G over a two-year period to manage purchases of $2.13 billion worth of goods yielding savings of $305 million, or 14%. The application also enabled P&G to achieve other benefits, including improved supplier relationships, reduced exposure to the risk of non-performance by suppliers, and expanding its supplier network. In addition, the optimization system has enabled P&G to understand the quantitative impact of various trade-offs used in its purchasing decisions, thereby helping to align all the stakeholders in P&G’s purchasing decisions.