Case Study: Pepsi Digest
Transforming a data ecosystem to uncover insights.
Changing the CIO office from data providers to insight providers.
Role
Head of Design, North America - Cognizant / Insight, Strategy & Design
The problem
The CIO office had all the data and knowledge within PepsiCo but was viewed as a reactive provider of charts and reports.
The solution
Working directly wih the CIO and senior leadership, I led a team to define a data transformation approach to empower the CIO team to proactively provide insights to the business. We chose the daily sales report as the pilot project as it was a mission-critical tool for the company.
From 40 pages to 3 key metrics.
The daily sales report was a 40 page spreadsheet that took most of the day to understand, and by that time the data was out of date. Based on our user research we learned that 80% of users relied on 3 key metrics to determine if they were on track or needed to investigate.
Co-designed with users.
We interviewed over 30 users to understand the end to end lifecycle of the data report from how it was produced, to how is was consumed.
Speed of work.
We worked on-site with the Pepsi team to co-create solutions in low-fidelity, test with users and then prototype with real data.
Defining an ecosystem.
In order to create a larger transformation the company needed a framework to holistically approach data from identifying sources, business processes and tools to understanding how users engage with data both as individuals and as collaborative teams.
Natural Language Generation.
Natural Language Generation (NLG) was explored as an alternative to charts and attempting to draw insights immediately from the data. We worked with Arria Natural Language Generation product for the initial launch.
Instant contact.
One of the biggest challenges people had with data was figuring out who they could contact with questions so we mapped each data point with direct contacts.
Deeper dives.
Based on our research, we understood the dimensions that users would diagnose business problems and provided the ability to dive into data through multiple dimensions such as region, customer or brand to identify root causes.
Results
- A hollistic framework to approach data transformation.
- A tangible demonstration of both new capabilities and sensibilities.
- A pilot to demonstrate both Natural Language Generation and AI.