Propose a multi-touch attribution model and its data pipeline
This tests your ability to choose a practical data model under real-world constraints. Propose a time-decay or position-based model, then describe the data pipeline: event collection, identity resolution, and aggregation. A red flag is ignoring signal loss.
This tests your grasp of trade-offs between model complexity and data quality under signal loss. A strong answer proposes a rule-based model like time-decay, details the data pipeline (event collection, identity resolution, aggregation), and specifies a long attribution window (e.g., 90 days) for B2B. The primary red flag is ignoring that MTA coverage is down to 30-60% of 2020 levels or proposing a data-driven model without the required 2,000+ monthly conversions, showing a lack of practical awareness.
Read the original → improvado.io
- #data engineering
- #analytics
- #data modeling
- #metrics
- #system design
Get five bites like this every day.
Tezvyn delivers a daily feed of 60-second tech bites with quizzes to lock in what you learn.