How P&G uses real-time data for supply chain resiliency

Meet supply chain disruption with fast data-driven insight

Real-time data is essential for supply chain resiliency, per research by economic advisory firm, Oxford Economics, because it gives organizations “superior visibility up and down the supply chain”. While 75% of leaders see the importance of increased visibility for more informed decisions, few actually have it.

Procter & Gamble, the multinational consumer goods corporation, is actively working to make this a reality.

When a hurricane eliminates supply from Florida or the Suez Canal blocks with 6 months of orders stuck on a boat, or a pandemic shuts down borders, the supply chain repercussions are huge.

Identifying exactly which products, supplier networks, and plant equipment this will affect is a tall order when you have one of the largest supply chains in the world. P&G’s Personal Health Care sub-sector alone counts more than 5,000 unique products. Those 5,000 products are made up of over 22,000 individual components, with the supply and demand of each affecting the one that follows it.

P&G was already using data from multiple systems to answer these supply questions, but integrating the data was often time-consuming, manual, and required up to 10 different types of expertise. So when the pandemic upended retail, the need for real-time, data-driven insight became even more acute.

A major step towards superior visibility through real-time forecasting

Brennon Sessions, Sr. Scientist in P&G’s R&D group, and a chemist by trade, is one of the functional experts at P&G who has been identifying which data reporting practices can be improved by automation to improve speed, accuracy, and practice. In a collaboration with analytics provider, phData, they developed a solution that gathers and merges all the relevant technical and supply data to tell a single, holistic story – getting closer to superior visibility up and down the supply chain.

Data activation and automation is the focus of Sessions’ team. They faced the challenge of integrating complex data from multiple sources: from manufacturing, to lab information, to supply chain, marketing, and quality assurance.

From individual pieces of information to a single, holistic data story

Supply chain data is one of the largest sources with records easily in the 10s of millions. Every lot of material and batch of a product has associated quality testing records, with at least 10 quality tests performed on each batch.

“It’s common for final dataset sizes to exceed a million rows after appropriate filtering & cleaning; a single Excel sheet maxes out at 1,048,576 rows,” says Sessions.

Previously data collection was manual. It had to be performed by people with expertise for that data type. The manufacturing-, lab-information-, supply chain-, marketing-, quality-assurance-data experts spent hundreds of hours gathering and integrating data. In daily meetings, the integrity of the data was checked and involved going through multiple, differing Excel files, by hand.

Automatic data integration provides immediate insight

Combining the technical expertise in KNIME from phData with subject matter expertise from Brennon’s team, P&G has upskilled to the next level. They developed a KNIME solution to fully automate data collection to produce a single, accessible pool of information.

Managers, scientists, and researchers are now enabled to explore and gain insight from this accessible pool. What took hours now only takes minutes and they have answers to supply and demand questions instantaneously.

Instead of people spending valuable time every day collecting data they can put their expertise to best use, exploring the data for additional insight.

The KNIME solution automatically combines multiple data sources together to assess the potential impacts of unexpected changes to supply & demand forecasts, and provide guidance on options that could help mitigate risks.

Live data delivers precise supply and demand forecasting

To provide precise real-time supply and demand forecasting, the KNIME solution brings together:

• Bill of materials data to identify the 22,000 components making up each of the 5000 products
• Supply chain data on all the permitted vendors for each of the 22,000 components, plus the involved manufacturing plants, warehouses, and distribution center
• Current inventory data to understand the potential risks to supply and demand changes

Using all this data the solution is able to produce a real-time summary of expected demand together with a real-time supplier projection of the quantities and timings for component deliveries.

With the new solution, leadership has a more efficient view of the data and can make more specific decisions. Daily, regional decision meetings have now been replaced with one global decision meeting.

User-friendly dashboards enable multi-level views of live data

The dataset generated by the KNIME workflows were used to produce a summary specification view, designed to provide live data to as many different people as possible in a user-friendly interface. 

P&G leveraged data and analytics services provided by phData, a KNIME partner, to create the initial iterations of the dashboard. With this foundation, P&G’s technical experts can now modify and expand without having to become dashboard experts themselves.

Different views exist for different levels in the business, with the global summary providing insight for leaders and close-up views allowing more hands-on colleagues to dig deeper to look up all the parts associated with specific products. Previously, digging through all the systems to find all the product parts could take hours. It now only takes seconds.

Benefits at a glance: before and after KNIME

• From manual data collection to automated data integration
• From 10+data navigation experts to 0 people
• From 2+ hours/day to get answers to questions to instantaneous insight
• From daily regional review meetings to a single global decision meeting

Organization upskilled to the next level

P&G researchers and scientists had already started to use KNIME as a way to layer data science over their work. In fact, KNIME is used throughout P&G in multiple sectors ranging across both technical and commercial functions, including marketing, finance, sales and manufacturing operations. The intuitiveness of the low-code environment makes it easier for “citizen developers” with little or no programming experience to get started and learn quickly.

“KNIME gives us access to advanced data science techniques and we can still focus on our work. There are now around 8,000 KNIME users across P&G globally,” said Brennon Sessions.

With so many employees using KNIME it’s a great opportunity to share knowledge, collaborate, and learn from each other.