Inventory Optimization at Proctor & Gamble

Inventory Management in a company of P&G’s size, scale and complexity requires leveraging the right people, organization and tools. By marrying the appropriate operations research (OR) techniques with a unique planning-organization structure, P&G has achieved two step-change improvements in inventory levels. The first improvement came from the broad application of spreadsheet-based inventory models; this work produced four tools that locally optimize different portions of the supply chain. The second improvement integrated multi-echelon inventory software in P&G’s more-complex supply chains. In 2009, these tools, implemented through a well-coordinated planning community, were instrumental in driving $1.5 billion in cash savings, while maintaining or increasing service levels. This work produced 4 tools that locally optimize different portions of the supply chain.

The first step of inventory optimization led to the development of the following 4 spread-sheet based tools:

  • FIM: Finished goods Inventory Model for the optimization of the finished goods inventory held at P&G plant locations

  • XIM: Extended Inventory Model for the optimization of inventory held at distribution centers with several remote stocking locations

  • RIM: Retail Inventory Models developed for retail level inventory optimization by incorporating customer variations

  • MIM: Materials Inventory Model meant for optimizing raw material inventory at plant locations


These Single-Stage Inventory models provide benefits to the business as statistical inventory models avoid supply chain planner bias towards excessively high levels of stock. Safety stock at too low a level would quickly result in customer service incidents and the planner would choose higher safety stock targets. However, when a planner manually sets a high safety stock target, no automatic correction mechanism is available to adjust the safety stock levels downward.




Formula 1: Saving Lives

Formula 1 for some is a mindless sport of screaming cars, for some it is epitome of automobile engineering, while a few consider it modern horse race of developed nations. But of many possible things that Formula 1 racing can be, it is the breeding ground of technological innovations which have impacted our lives for decades that passed and many more to come. 

In one such development, technologies from Mclaren F1 team are employed to save children’s lives at Birmingham Children’s hospital in UK. At first glance, it appears that a toddler with a heart condition and a Formula 1 racing car, have little in common.

Cars on the race track are very much like patients in hospital, they need to be constantly monitored. Sensors on the vehicles send back real time data to a team of engineers who make adjustments, either trackside or after the race, to improve performance next time around. The Formula 1telemetry system is used to gather and process vast amounts of data – each car has about 130 parameters, 25000 possible failures and 15,000 health checks. Nearly 120 onboard sensors send back 2-4 Mbps of data back, so the software not only has to quickly process this data, but also “learn” the new normal for individual components.

Telemetry is highly automated communications process, which is employed to measure data from remote or inaccessible points and further transmitted for monitoring. But the telemetrics that assess a car’s performance on the track are now being used to monitor children with serious medical problems in intensive care ward of Birmingham Children’s Hospital (BCH). It is the first time such a system has been adapted for use on humans.


By contrast, most hospitals rely on a combination of paper charts – plotting the vital signs of patients, from heart rate to blood pressure – and the four-hourly observations conducted by nurses and doctors. When a patient is in intensive care, he or she is wired up to a series of monitors and these are used by the McLaren system to collect continuous data about heart rate, respiration, oxygen levels and blood pressure. Constant monitoring offers two huge benefits for doctors:

  • It warns the doctors about the deviations/fluctuations from normal data patterns, and
  • It also enables doctors to learn about individual’s normal conditions. All of us are different and have different normal settings for our bodily parameters. [We are engineered to order (ETO) and not Made to stock (MTO)]

Currently data collected in the intensive care ward has to be deleted after 96 hours in order to accommodate new readings. So having long-term data on patients for the first time could prove extremely valuable in predicting sudden occurrences. For patients with chronic conditions, it would be possible to monitor them at home because they van be connected to experts in hospital.

Data analysts from Mclaren F1 racing have been drafted in to help make sense of the information and create a simple-to-use early-warning system – with red signifying changes to patients that need immediate attention, amber suggesting slight changes and green representing normality.

The trial system at BCH is in the early stages and McLaren engineers are working closely with doctors and data analysts at the hospital to make sure it is successful.