This assignment is based on the Managing with Analytics at Procter & Gamble? case

9-613-045

A P R I L 3 , 2 0 1 3

________________________________________________________________________________________________________________

Professors Thomas H. Davenport and Marco Iansiti and Research Associate Alain Serels prepared this case. Some names and facts in this case have been disguised. HBS cases are developed solely as the basis for class discussion. Cases are not intended to serve as endorsements, sources of primary data, or illustrations of effective or ineffective management. Copyright © 2013 President and Fellows of Harvard College. To order copies or request permission to reproduce materials, call 1-800-545-7685, write Harvard Business School Publishing, Boston, MA 02163, or go to www.hbsp.harvard.edu/educators. This publication may not be digitized, photocopied, or otherwise reproduced, posted, or transmitted, without the permission of Harvard Business School.

T H O M A S H . D A V E N P O R T

M A R C O I A N S I T I

A L A I N S E R E L S

Managing with Analytics at Procter & Gamble

Alan Torres, vice-president of North America Fabric Care at Procter & Gamble (P&G), watched and listened as his embedded IDS analyst, Julia Wright, graphically displayed the initial results. Torres leaned back to take in the complete set of 12 graphs and charts on the massive 8-foot by 32- foot screens in a specialized meeting room called a Business Sphere at P&G’s headquarters in Cincinnati, Ohio. The company had introduced a more concentrated, also called compacted, powder laundry detergent in Target® stores at the end of February 2011 and results from the first two months were better than anyone had expected.

Over the previous 8 years, Filippo Passerini, Group President of P&G’s Global Business Services (GBS) organization and Chief Information Officer (CIO), had led the development of a series of systems and processes that enabled and emphasized the use of up-to-date data and advanced analytics to drive decision making throughout the company. Through these efforts, Torres and his team were able to see the results of this first wave of compaction as they became available. Data collected by P&G was seamlessly integrated so that Torres and Wright could simply click on higher level data to drill down and view performance by brand, initiative, retailer, and, in some cases, individual store. In addition, statistical forecasts for the next twelve months were automatically updated and made available.

Equipped with these tools, Wright led Torres through a series of charts and graphs that showed a three percent increase in the overall detergent category in the stores where the compacted detergents were introduced. Torres had anticipated that the compacted format would provide a slight boost to the mature $8.5 billion North America laundry detergent market, where growth rarely exceeded 2%.1 But the market had contracted by 2% in 2010 after a gain of just 1% in 2009, so growth of over 2% was beyond all expectations. Even more surprising, P&G’s powder detergent sales jumped 12% as the company’s market leading brands, including Tide®, Gain®, and Cheer®, gained market share. P&G already held more than a 50% share of the North America laundry detergent market that generated over $4 billion in revenue for the company in 2010.2

Compacted detergents were set to be introduced throughout the rest of North America in the coming months and Torres was cautiously optimistic about the impact they would have on the overall market and on P&G sales. The statistical models used to produce market size and P&G sales forecasts did not fully account for the market disruption caused by the introduction of compacted detergents and continued to produce forecasts that showed growth of 1% for the 2011 calendar year.

For the exclusive use of H. Dharod, 2021.

This document is authorized for use only by Harshmi Rajesh Dharod in IPM652 Managing with Analytics taught by Mareike Moehlmann, Bentley University from Sep 2021 to Dec 2021.

613-045 Managing with Analytics at Procter & Gamble

2

As Torres stared at the array of charts on the colossal screens of the Business Sphere, he wondered whether he should override the statistical models and increase the forecasts.

Procter & Gamble Overview and Organization

Pressured to join their businesses by their mutual father-in-law, William Procter and James Gamble partnered to form The Procter & Gamble Company in 1837. The Cincinnati, Ohio based company quickly became a leading supplier of candles and soap, achieving sales of $1 million by 1859. Three years after the introduction of Ivory® brand soap, the company decided to launch the first ever national advertising campaign that directly targeted consumers.3 Since then, P&G has focused on innovation and branding to become the world’s largest consumer packaged goods company.

By 2011, P&G marketed personal care and cleaning products in 180 countries that were ultimately used by 4.2 billion people around the world.4 It was estimated that 98% of all North American households contained at least one P&G product.5 The company marketed approximately 300 brands, 25 of which generated over $1 billion in annual sales, including Always®, Bounty®, Charmin®, Crest®, Duracell®, Gillette®, Head & Shoulders®, Olay®, Pampers® and Pantene®.6 In total, P&G was on track to amass sales of over $80 billion in fiscal year 2011 (ended June 30, 2011). (See Exhibit 1

for company revenue and profitability)

Organizational Structure

In an effort to spur profitable growth, P&G embarked on an aggressive two year-long corporate restructuring in 1999. The initiative, called Organization 2005, aimed to reorganize the many small, relatively independent businesses that had been loosely organized by region. By configuring as global product line-focused business units, the company hoped to more fully take advantage of P&G’s established brands, tremendous scale, and worldwide presence. Critically, the new structure would allow P&G to more effectively develop and implement innovations on a global scale. The 1999 annual report described Organization 2005 as a “realignment of the organization structure, work processes and culture designed to accelerate growth by streamlining management decision-making, manufacturing and other work processes to increase the Company’s ability to innovate and bring initiatives to global markets more quickly.”7

Following the restructuring, P&G’s 129,000 employees were organized into four independent global organizations: (Exhibit 2)

Global Business Units (GBUs) The primary driver of commercial activity at P&G was the GBU. The product line-based GBUs had primary responsibility for developing strategy, building brands, manufacturing products, developing and bringing to market innovations, contending with competitors, and delivering profits. By centralizing at the global level, GBUs were better able to standardize manufacturing processes, coordinate marketing activities, build global brands, identify significant innovations and bring new products to market quickly. As of 2011, P&G consisted of GBUs covering five product-based segments: Beauty, Grooming, Health Care, Fabric Care and Home Care, and Baby Care and Family Care.

Market Development Organizations (MDOs) By maintaining a local presence, the MDOs provided P&G with the knowledge and relationships required to compete locally in each geographic region. MDOs were responsible for understanding the preferences of local consumers and retailers so that the global initiatives and brands of the GBUs could be tailored for local markets. In addition,

For the exclusive use of H. Dharod, 2021.

This document is authorized for use only by Harshmi Rajesh Dharod in IPM652 Managing with Analytics taught by Mareike Moehlmann, Bentley University from Sep 2021 to Dec 2021.

Managing with Analytics at Procter & Gamble 613-045

3

having leaders in each market enabled the company to identify trends in local markets faster. The MDOs were organized into five geographies: North America, Western Europe, Asia, Latin America, and Central & Eastern Europe, Middle East and Africa.

Global Business Services (GBS) Business support services including accounting, payroll, employee services, information technology, order management, and logistics were moved into the newly formed GBS. Prior to the reorganization, these functions were duplicated in each unit. In centralizing these activities, P&G looked to gain efficiencies through scale, standardization of shared services and the transfer of best practices across the organization.

Corporate Functions Corporate accounting, legal, communications, governance, corporate strategy, and other corporate support functions were provided by a lean Corporate Functions group.

GBS

GBS functioned as a service to the GBUs and MDOs across P&G. As the company’s back office, GBS looked to provide the high quality services required by its internal customers “at lowest cost and with minimal capital investment.”8 GBS had to break even each year, and, therefore, could only supply services that were valued by the GBUs and MDOs. In total, GBS provided more than 170 shared services and solutions to the various groups in the organization.9 (Exhibit 3)

In 2003, Passerini was put in charge of GBS. A native of Italy, Passerini earned a doctorate in statistics and operations research from the University of Rome. He had served many roles across several geographies and product lines since joining P&G in 1981 and he understood both the technical demands of GBS and the needs of a diverse set of P&G business units.

Shortly after joining GBS, Passerini embarked on the second phase of the reorganization of shared services at P&G. The company looked to achieve even greater scale efficiencies by leveraging external partners to deliver many of the services initially centralized into GBS. In 2003, P&G entered into contracts worth $4.2 billion for several activities where Passerini felt that the company could gain efficiencies without sacrificing strategic capabilities.10

Initially GBS planned to utilize third party providers for all of P&G’s IT capabilities. But managers realized that it would be difficult for an outside entity to work closely with P&G executives on important decisions. Also, these external providers would not understand the company’s business well enough to create innovations. So, they decided to keep in-house the IT capabilities for innovation, system design and architecture, and those that supported decision-making.

Passerini, whose title was augmented to include CIO, renamed these essential IT activities “Information and Decision Solutions,” (IDS) befitting the group’s support for decision processes around P&G. IDS built up its human and technical capabilities for decision support and analytics in both the supply chain and the sales and marketing domains. IDS also worked throughout the decade of the 2000s to consolidate systems and integrate them into a much smaller set of global platforms.

In 2011, GBS was comprised of more than 6,000 employees in one centralized service organization. The group operated 6 hubs around the world that provided consistent services to all P&G units globally.11 In addition, a portion of GBS employees were embedded in the GBUs and MDOs to whom they provided service. GBS worked with a core of 15 external partners that supplied a wide range of support services to P&G’s operating units.12 Over the first 8 years since Passerini had taken over, GBS delivered $900 million in cost reductions for P&G.13 Equally impressive, GBS did so while expanding

For the exclusive use of H. Dharod, 2021.

This document is authorized for use only by Harshmi Rajesh Dharod in IPM652 Managing with Analytics taught by Mareike Moehlmann, Bentley University from Sep 2021 to Dec 2021.

613-045 Managing with Analytics at Procter & Gamble

4

the number of services offered and improving service levels, which demonstrated the increased value that the GBUs and MDOs placed on the services provided by GBS.

Analytics at P&G

We see business intelligence as a key way to drive innovation, fueled by productivity, in everything we do. To do this, we must move business intelligence from the periphery of operations to the center of how business gets done.

— Bob McDonald, President, CEO, and Chairman, P&G

Initially, P&G’s management felt that IT could support data analysis throughout the company by acting as a central repository for data and by providing software tools to aid analysts. Passerini, however, envisioned a much larger role. IDS, he proposed, would support business leaders by performing much of the analytics and by making data as accessible as possible. The P&G website described the division’s broad role, “IDS stretches far beyond the traditional aspects of ‘hardware and software.’ Instead, we are a key business enabler that advances technology tools, strategic development, collaboration, and decision making.”14

By the end of the decade, Passerini felt that P&G had the systems, data, and people within IDS to change the process of managerial decision making at the company. Fortunately for Passerini and IDS, in July 2009, Bob McDonald became the President and CEO, and later Chairman of the Board, of P&G. Having studied operations research, computer science, and engineering at West Point, McDonald shared Passerini’s enthusiasm for data and analytics.

McDonald, who had previously served as Chief Operations Officer and Vice-Chairman, Global Operations, had long understood that the analytics capabilities of IDS were critical to P&G’s continued success. The Global Operations section of the 2005 annual report, for which McDonald was responsible, stated that “GBS provides P&G with an often-unseen competitive advantage. GBS combines information management and analysis capabilities to take P&G’s use of information to a far more strategic level. These efforts help P&G’s Market Development Organizations and Global Business Units make better, smarter, real-time business decisions.”15

When McDonald became CEO, the effective use of information and technology at P&G became one of his primary emphases. In his inaugural letter to shareholders, published just one month after becoming CEO, McDonald wrote that P&G was “making the investments to be networked globally and digitized from end to end. And we’re building capability to operate on a demand-driven, real- time, future-focused basis.”16 In the 2010 annual report, he strengthened the corporate focus on data and analytics, “With digitization, our goal is to standardize, automate and integrate systems and data so we can create a real-time operating and decision-making environment. We want P&G to be the most technology-enabled company in the world.”17

McDonald emphasized the holistic nature of this transformation. “Moving business intelligence to the center of how business gets done is absolutely critical, as business analytics is a new way to win in the marketplace. And, it is going to be the tools, the talent, and the culture, not just the data, which are the true competitive advantages to driving business value.”

Analytics as a Centralized Service

Operating much of the data and analytics activities as a centralized service presented some noteworthy opportunities for improvement across P&G. Prior to Organization 2005, each part of the

For the exclusive use of H. Dharod, 2021.

This document is authorized for use only by Harshmi Rajesh Dharod in IPM652 Managing with Analytics taught by Mareike Moehlmann, Bentley University from Sep 2021 to Dec 2021.

Managing with Analytics at Procter & Gamble 613-045

5

business collected and stored data in independent systems. By bringing together the data collection activities, IDS was able to establish standards for data type and quality. Additionally, aggregating and comparing data across product lines and regions became significantly less complicated with the data stored in standard formats in the central data warehouse.

Over time, managers found that the single, company-wide database played a much more strategic role in aiding decision making by serving as the “one truth” for the entire corporation. As Passerini explained, “Instead of spending our time debating the data, we eliminate the need for those discussions. The data is the data, and leaders can spend their time concentrating on the business and the decisions that need to be made to move the business–instead of which set of data points is correct. This speeds decision-making and ultimately improves time to market.”18

IDS automated the generation of many reports that were used across multiple business units. In addition to simplifying the task of retrieving data and performing some basic analyses, the reports served to standardize the way data was visualized across the company. The same types of charts and graphics used to track sales and market share of Old Spice® deodorant in the US were also used to monitor Iams® dog food in the Philippines and Oral-B® toothbrushes in Brazil. With consistent visuals, analysts and managers from one unit could step into a role, or even a meeting, with a different product or region and quickly understand the situation.

The centralized IDS team also acted as a hub for sharing analytic techniques and ideas. McDonald explained that prior to the integration of analytics as a centralized service there were “5,000 people in the company who used to do demand planning. So, each one would have their own spreadsheet. And they would try to plan demand based upon the inputs that they saw. Of course, they all had their own technique.”19 IDS served as a platform for sharing best practices, which improved the overall quality of analytics at P&G.

Data and Enabling Technologies

As a service, IDS had to respond to the needs of the GBUs and MDOs. Any unit within P&G was able to request that IDS provide a new service, but IDS focused on building out scaled solutions that could be performed centrally and leveraged in multiple markets.

One of the first company wide solutions developed by IDS was the Decision Cockpit, a web-based customizable dashboard that tracked the most relevant data and news for each individual employee. (Exhibit 4) Each user was able to specify the data and reports displayed on their personal dashboard and could set alerts if data were outside of a specified range or if important information, like a competitor product introduction, became available. Linked to the centralized data warehouse, the dashboards were immediately updated as soon as new data or information became available, allowing managers to be notified and react in real-time. In 2011, over 50,000 employees had access to the Decision Cockpits.

This ambitious project had the goal of addressing two of Passerini and McDonald’s key passions: accelerating decision-making with real-time information, and changing the entire focus of information-based decisions from “what” to “why” and “how.” On the Decision Cockpit, users were able to drill-down into the data to reveal not only “what” was happening, but also “why” it was happening and better understand the options available for “how” P&G could react. Previously, once an exceptional event was noticed, managers would have to decide what data was available, how they could get it, and how it should be presented, spending a great deal of time producing a report to better understand the “what” and maybe the “why.” By simply double-clicking on a chart or

For the exclusive use of H. Dharod, 2021.

This document is authorized for use only by Harshmi Rajesh Dharod in IPM652 Managing with Analytics taught by Mareike Moehlmann, Bentley University from Sep 2021 to Dec 2021.

613-045 Managing with Analytics at Procter & Gamble

6

highlighting a portion of it, the Decision Cockpit displayed a more granular version of the same data or other related data that enabled users to rapidly identify the “what” and understand the “why.”

For example, if a manager noticed a drop in sales of Gillette® Fusion™ razors in Australia, the “what,” she could quickly understand the “why” by looking at sales by city or retailer, shipments, prices, promotions and competitor actions. With a few clicks of a mouse, or touches on her tablet computer, she could check if the overall market dropped or just the P&G brand. She could also check if a particular retailer decreased orders, or if customers just purchased lower quantities. Just as easily, she could investigate if there were any changes in pricing, if a competitor came out with a new product or if a competitor launched a new marketing campaign. With much less time needed to identify the “what” and understand the “why,” leaders were able to quickly move to the crucial decision making of “how” to fix the problem or maintain improved performance.

IDS worked with P&G business units, suppliers, customers and external vendors, such as Nielsen and IRI, to collect data about the “what,” “why” and “how.” Passerini and his team continuously looked for ways of improving the data available both in type and frequency. To collect data from social media, IDS developed Consumer Pulse, which provided real-time sentiment analysis for web buzz of each P&G brand. It leveraged best-in-class technology, including approaches used by the intelligence community to monitor terrorism.

For management teams, P&G developed a patent-pending physical environment for information- based decision-making called the Business Sphere. (Exhibit 5) The meeting room, located adjacent to the executive offices at P&G’s headquarters in Cincinnati, was spherical in shape with a conference table in the center. What made the Business Sphere unique was the computing and network capabilities to bring data, images, voice, and videoconferences into the room. Most noticeable was the surrounding set of two oversize 8-foot by 32-foot screens to display visualizations of data and analyses. Drawing on the 200 terabytes of data in P&G’s data warehouse, leaders could drill down into the data “on-the-spot” using color coded charts and displays to visualize all the relevant information at once.20 Passerini explained that the Business Sphere provided an environment that was “physically surrounding business leaders with the data they need to make actionable decisions. The visualization of the data makes it easy to focus on the exceptions and realize business opportunities and where interventions are necessary.”21

Leadership teams would typically gather in the Business Sphere at regular intervals, generally weekly or monthly, to review product and market performance for the previous periods and make decisions about how to proceed going forward. McDonald, for example, held his Monday morning global meetings of direct reports in the Business Sphere. They could also be used for ad hoc meetings to address a particular issue.

Forecasting and Business Sufficiency Models

In 2008, IDS began to assemble a centralized forecasting team that would use statistical methods to provide the GBUs, MDOs and company leadership with a clearer view of future market developments. Staffed with analysts specialized in statistics and econometrics, the team took under two years to develop models that provided 12-month forecasts for many P&G product-market combinations worldwide. The statistical forecasts were stored in the centralized database and made available through the Decision Cockpits. By 2011, the centralized forecasting team produced statistical forecasts for 6,000 product-market combinations covering market size, P&G sales and market share.

For the exclusive use of H. Dharod, 2021.

This document is authorized for use only by Harshmi Rajesh Dharod in IPM652 Managing with Analytics taught by Mareike Moehlmann, Bentley University from Sep 2021 to Dec 2021.

Managing with Analytics at Procter & Gamble 613-045

7

The initial forecasting models only provided point forecasts, single numbers for each time period that represented the most likely outcome. To give managers a more complete picture of the potential result, the team decided to provide probabilistic ranges that showed the spread of prospective outcomes. The forecasting team ran thousands of virtual simulations of how the future might unfold and tracked each to determine the likelihood of being, for example, 10% above the forecast. The ranges alerted managers to markets with higher uncertainty and variability, shifting the focus from large but stable markets to those that warranted increased attention.

P&G employed two types of statistical models for forecasting. The first type, generally time series- based forecasting models, produced the most accurate forecasts. These models constructed forecasts using previous results as well as macroeconomic conditions and other basic inputs. Although the outputs were precise, this set of models provided little insights as to why the outcome would occur. Propensity models, on the other hand, attempted to better understand the causes, or drivers, of the outcomes. By incorporating information on drivers such as new product introductions, marketing campaigns, and competitor actions, propensity models served as useful tools to measure the potential impact of management actions.

The forecasts and ranges were extended to a set of seven “Business Sufficiency Models” developed by IDS in conjunction with senior executives. These models captured the most recent data and conducted a large number of simulations to provide leaders with a measure of the likelihood that the business would be able to deliver on previously set goals and targets. Areas that seemed unlikely to meet objectives were highlighted so that leaders could quickly move to explore actions of “how” to rectify the situation and alert other parts of the organization that may be impacted.

Andy Walter, who headed the Business Intelligence area within P&G, explained the concept of Business Sufficiency Models:

The concept of business sufficiency is that we commit to a performance goal, and the key question is whether I have sufficiency to deliver that goal. It may be a supply chain goal, a production goal, or a sales volume goal. We want to make sure that we have all the factors lined up to achieve it, and the models provide data on those factors. They are mostly backward-looking reports now, but in some areas we’re producing forecasts and projections. With the seven models we cover about 90% of the information the company needs to run itself. But our executives are always asking for more and better information, so we are always refining and adding to the models.

The business sufficiency models addressed key business domains for P&G, including product, customer, supply chain, distribution channels, human resources, and the “Top 50” global product- market combinations. While most Sphere meetings used only one of the models, it was also possible to work across them in a particular meeting. If a team was focused, for example, on whether P&G had the inventory and the people to introduce a Swiffer® product in Korea, it could use both the supply chain and human resources models.

The Role of Embedded IDS Analysts

The Business Sphere meetings and the daily use of the business sufficiency models were facilitated by a group of 300 IDS analysts who were embedded in the business units and functions. IDS had centralized the quantitative analyst function in the mid-2000s, and over time had developed stronger relationships between analysts and the business leaders they supported. By 2010, when IDS was ready to roll out the business sufficiency models and Business Spheres, it became obvious that the use of all this information by business leaders required facilitation, and the IDS analysts were the

For the exclusive use of H. Dharod, 2021.

This document is authorized for use only by Harshmi Rajesh Dharod in IPM652 Managing with Analytics taught by Mareike Moehlmann, Bentley University from Sep 2021 to Dec 2021.

613-045 Managing with Analytics at Procter & Gamble

8

obvious candidates. IDS embedded analysts within the business units, each dedicated to working alongside leaders and managers of a particular team on a daily basis. In some cases, the embedded IDS analyst was a member of the business unit’s leadership team.

The embedded IDS analyst role had multiple components. The analyst was responsible for bringing technology, visualization, and automated alerting functions to the business unit. They were responsible for integration across diverse data domains and models (as in the Korea Swiffer® example above). They were required to develop analytical models driving forward-looking insights and simulations within and across the business domains they supported. And, they led market

Collepals.com Plagiarism Free Papers

Are you looking for custom essay writing service or even dissertation writing services? Just request for our write my paper service, and we'll match you with the best essay writer in your subject! With an exceptional team of professional academic experts in a wide range of subjects, we can guarantee you an unrivaled quality of custom-written papers.

Get ZERO PLAGIARISM, HUMAN WRITTEN ESSAYS

Why Hire Collepals.com writers to do your paper?

Quality- We are experienced and have access to ample research materials.

We write plagiarism Free Content

Confidential- We never share or sell your personal information to third parties.

Support-Chat with us today! We are always waiting to answer all your questions.