BayLou Featured Project: Pepsico Business Intelligence Analysis

As a featured project, BayLou’s business intelligence analysis engagement with Pepsico illustrates a range of services and core competencies to deliver enterprise results…

PepsiCo Foodservice Division [PFS] actively pursues innovations that yield greater value for their channel partners. In 2013, PFS engaged BayLou to solve a complex problem – develop a process and logistics analysis toolkit that optimizes product portfolios to increase operators vending machine level profits. BayLou delivered and executed its Product Optimization Model [POM] that addressed four core aspects of the problem – data management, logistical algorithms, data mining, and reporting. ii-PepsicoFS-POM

End-to-end ETL (extract-transform-load) and management of a robust data repository required for business intelligence, logistical, and other analyses — both standard and custom. Data warehousing requirements spanned acquisition, cleaning, codifying, structuring, storing, and delivery. The scope further included workflow, scheduling, policy development, backup and recovery procedures, security, retention, capacity planning and management, and other general functions that enable seamless and reliable fulfillment of customer demand for data. A number of transformations to the data were required in order to support the intended analytics.

The focus will be to transition Operators from low-to-high profitability SKUs based on machine-level portfolio analytics and management. The process follows each machine through recommendation, conversion (or Swap). Note, this process requires a number of addition to POM in order to move from POM I (demonstration) to POM II (PFS ROI orientation). This includes developing/applying an Operator Selection assessment and pre-analysis, which means possible 2X-to-4X Operators analyzed than the number of Operators ultimately taken through the POM II process.

Analytics that yielded insights into tier-I questions for PFS, other Pepsico divisions, and Operators. Data mining leveraged the large data repository that contains multiple Operators, Segments, Categories, and other key business entities to perform uni-variate and multivariate, descriptive (e.g., Analysis of Variance) and prescriptive (e.g., regression) analytics. The scope of these analysis (i.e., nature of the questions) include: impact of strategic products, machine product SKU portfolios, category contribution to profits, etc.

Multiple organizations and various departments with them had direct interest during the frontend and backend of the POM project. Likewise, departmental performance measures gin, logistical efficiency) created various prisms through which reports were viewed. BayLou delivered a battery of reports at points along the project that enabled PFS Category Management effectively translate project insights to diverse stakeholders.

We are pleased to have delivered a solution that added value to one of the worst most recognized brands. The core competencies brought to the PFS project can be transportable to and scaled for other clients faced with complex questions..