Applying Data Science and Analytics at P&G Custom Case Solution & Analysis

1. Evidence Brief: Business Case Data Research

Financial Metrics

• Annual Net Sales: 83.5 billion dollars for fiscal year 2012.
• Net Earnings: 10.8 billion dollars.
• IT Investment: Approximately 2 billion dollars annually, representing 2.5 percent of sales.
• Marketing Spend: 9.3 billion dollars, much of it targeted for digital optimization.
• Cost Savings Target: 800 million dollars through Global Business Services (GBS) initiatives.

Operational Facts

• Scale: 127,000 employees operating in 180 countries with 300 brands.
• Infrastructure: 50 Business Sphere locations globally for real-time data visualization.
• Digital Access: 50,000 employees equipped with Decision Cockpits on personal desktops.
• Data Volume: Processing billions of data points across the supply chain and consumer touchpoints.
• Organizational Structure: GBS provides shared services to four Global Business Units (GBUs).

Stakeholder Positions

• Bob McDonald (CEO): Views digitization as the primary path to becoming the most digital company on earth.
• Filippo Passerini (CIO/President of GBS): Architect of the One Version of the Truth strategy. Focuses on eliminating time spent debating data accuracy.
• Guy Peri (Director of IT): Tasked with moving the organization from descriptive to predictive analytics.
• Guy-Charles Bonnet (Associate Director): Leads the execution of data science applications within specific business units.
• Business Unit Managers: Historically relied on intuition and retrospective monthly reports; now required to use real-time predictive tools.

Information Gaps

• Specific retention rates for data science talent compared to Silicon Valley competitors.
• Granular ROI for individual Business Sphere installations.
• The exact percentage of mid-level managers who actively use Decision Cockpits daily versus those who rely on traditional spreadsheets.
• Capital expenditure requirements for transitioning legacy ERP systems to support automated predictive modeling.

2. Strategic Analysis: Market Strategy Consultant

Core Strategic Question

• How can P&G transition from data visualization to automated predictive modeling while maintaining organizational agility across 300 diverse brands?

Structural Analysis

Applying the Value Chain lens, P&G has digitized the support activities (IT and Procurement) but faces a bottleneck in the primary activity of Marketing and Sales. The current data infrastructure provides high visibility but low agency. While the Business Sphere identifies a sales drop in real-time, it does not automatically trigger a corrective supply chain response. The structural problem is the gap between data insight and operational action.

Using the Jobs-to-be-Done framework, the manager job is not to view data, but to mitigate risk and capture growth. Current tools provide the view, but the predictive modeling required to mitigate risk is still in its infancy and lacks standardized deployment.

Strategic Options

Preliminary Recommendation

P&G must adopt the Hybrid Platform Model. This approach ensures One Version of the Truth through a centralized data lake while allowing individual brands to develop custom predictive algorithms. This balances the need for global scale with the requirement for local market sensitivity.

3. Implementation Roadmap: Operations and Execution

Critical Path

• Month 1-3: Standardize the predictive modeling API. All 300 brands must use the same underlying data definitions to prevent fragmented insights.
• Month 4-6: Launch a pilot predictive program in one high-volume category, such as Fabric Care, focusing on automated inventory replenishment.
• Month 7-12: Scale the pilot findings to all Global Business Units, replacing retrospective monthly reviews with weekly predictive sprints.

Key Constraints

• Talent Scarcity: The primary constraint is the availability of data scientists who understand CPG dynamics. P&G cannot outbid tech firms on salary alone; it must sell the scale of its data.
• Legacy Resistance: Mid-level managers who have spent 20 years relying on gut feeling represent a significant cultural barrier to automated decision-making.

Risk-Adjusted Implementation Strategy

The strategy assumes an 18-month rollout but includes a 20 percent time buffer for data cleaning in emerging markets. We will utilize a shadow-running approach where predictive models run alongside human decision-makers for one quarter to build trust in the algorithm accuracy before granting the system autonomy over purchase orders.

4. Executive Review and BLUF

BLUF

P&G has achieved world-class data visibility but lacks data-driven autonomy. The Business Sphere and Decision Cockpits are successful communication tools that have reduced internal debate. However, they remain retrospective. To maintain its competitive edge, P&G must pivot from visualization to predictive automation. The recommendation is to implement a hybrid architecture that centralizes data governance while decentralizing model creation. This shift will move P&G from describing what happened to dictating what should happen. Success depends on shrinking the time between insight and execution from weeks to hours. Failure to automate these decisions will result in a talent drain and a loss of market share to more agile, digitally-native competitors.

Dangerous Assumption

The analysis assumes that data visibility automatically leads to better decision-making. In reality, providing more data to a manager who lacks the statistical literacy to interpret predictive intervals can lead to paralysis or incorrect interventions.

Unaddressed Risks

• Model Drift: Predictive models built on historical data may fail during black swan events or rapid shifts in consumer behavior, leading to massive over-stocking or stock-outs.
• Data Privacy Regulation: Increasingly stringent global privacy laws may limit the ability to integrate consumer-level data into the predictive models, weakening their accuracy.

Unconsidered Alternative

The team did not consider an Outsourced Analytics Model. Instead of building internal data science capabilities, P&G could partner with specialized AI firms to manage the predictive layer. This would solve the talent acquisition problem but would sacrifice long-term strategic control over proprietary consumer insights.

Verdict

APPROVED FOR LEADERSHIP REVIEW