Ecommerce going offline - parcel lockers network analytics
summary
Allegro builds a fulfillment and last mile stack around a proprietary parcel locker network called One Boxes and aims for 3000 lockers this year already at about 1,600 installed in Poland where 20,000 lockers exist. Lockers stream IoT data — air quality, temperature, GPS, motion, cameras — into Grafana and feed a predictive stack: ML models that forecast monthly orders and seasonal peaks, ETL into BigQuery, a site selection app tied to Salesforce, and KPIs for population coverage and locker utilization; installs take two weeks to a month.
AI-written summary · about
chapters (11)
- 0:00Introduction
- 1:55Allegro’s delivery network
- 3:19Parcel lockers in Poland
- 5:48Sensors and IoT in lockers
- 7:15Data sources and tech stack
- 9:01Expansion strategy stages
- 12:52Closing and resources
- 13:13Q&A on talent
- 14:59Design and branding impact
- 16:47Cannibalization and coverage
- 17:51Current network status
related talks
49mUnderstanding Customer Journeys with Deep Learning
24mSegmenting Customers with Machine Learning
20mPartial automation of the dispatching process to reduce empty kilometers
58mReal Data Science Examples at OLX's Online Marketplace
34mData Drives Business
36mML: Predictive analytics in Ad Engineering
48mUrban Data & Decisions
26mDriving towards a cleaner future through data