Kapil Malhotra: Head – Analytics at PepsiCo speaks at Cypher 2017
summary
New analytics function starts with two analysts, three org-structure options evaluated and a category-aligned model chosen; forecasting moves from pure time series to causal/econometric models and a version two improves accuracy. Reporting overhaul cut reports by half and instituted simple data dumps, which reduced delivery time 33%. Pilots include Watson for digital-sentiment-to-volume proof of concept; market-mix models for 2015 are delivered but limited by data quality and refresh resources.
AI-written summary · about
chapters (19)
- 0:00New analytics role at PepsiCo
- 1:44Business challenges at PepsiCo
- 2:26Building analytics from scratch
- 4:36Gap assessment and team structure
- 5:39Hiring for culture and skills
- 7:06Processes for analytics success
- 10:22Partnering with functions
- 11:47Business-driven analytics
- 13:11Data harmonization and dashboards
- 14:38Service provider mindset
- 15:20Solution design leadership
- 16:48Reporting overhaul
- 17:53Democratizing data
- 19:18Forecasting with causal models
- 20:42Forecasting and business cycles
- 21:43Econometric models for CPG
- 23:09Sales and marketing spend optimization
- 24:56Market mix modeling limits
- 26:03Hyper-targeting and innovation
related talks
53mBusiness transformation through data and ML
31mTransforming Data Science at PepsiCo: The Kubeflow Revolution
29mDeciphering Sales Drivers at PepsiCo
12mInterview
29mBecoming a data-driven enterprise
11mInterview with Wes Nichols
43mConsumer Intelligence and the Age of Predictive Marketing Analytics
32mMLDS 2019: Pradeep Gulipalli of Tiger Analytics at Machine Learning Developers Summit 2019