User Motivation: Refining Customer Segments with Location
summary
Combines deterministic SDK location logs with demographic and offline interest data to build industry-specific segments using k-means clustering and elbow-method selection, operating on roughly 40M monthly active users, 200M mobile IDs and ~700M visits/month. Produces hyper-granular personas (e.g., “fast-foodies” or Whole Foods micro-segments), reports 23% lift in store visits for a targeted protein brand and 24–65% higher CTR versus traditional segments, and handles sampling bias by long time windows and large panels.
AI-written summary · about
chapters (15)
- 0:00Introduction and speaker intro
- 1:09Mobile advertising and location intelligence
- 3:02What is location data
- 5:14Visit data and customer journeys
- 6:40Hyper-focused customer segments
- 8:04Combining location and traditional data
- 9:50Why machine learning for segmentation
- 11:18Clustering and post-cluster analysis
- 12:43Industry-specific segments
- 13:46Challenges in visit data
- 15:13Results and example segments
- 17:22Using segments for insights and targeting
- 18:27Use case: Protein shake brand
- 19:35Use case: Holiday electronics shoppers
- 20:19Q&A session
related talks
50mMulti dimensional audience segmentations in products
45mLeveraging Power of Analytics for Martech
40mMachine Learning over Small Data: generating new variables with predictive power
06mMachine Learning for Phone-Based Relationship Estimation: The Need to Consider Population ...
26mThe Need for People Based Measurement
21mData Analytics for Better Product Decision Making
18mLeverage Customer Data to Build Personas
13m"Decoding the Urban DNA and Harnes