Data-driven Company: From Data to Decisions

What Does It Mean to Make Data-Based Decisions

Data-driven is a culture and practice of making decisions based on data analysis, rather than intuition or experience. Amazon: Bezos introduced a “culture of metrics”—every decision that can be measured must be measured.

Stages of data analytics:

Descriptive: what happened? (Reports, dashboards, KPI)
Diagnostic: why did it happen? (Root cause analysis, drill-down)
Predictive: what will happen? (Forecast models, ML)
Prescriptive: what should we do? (Optimization, AI recommendations)

Key Metrics and KPI

North Star Metric: one key metric reflecting value for customers. Airbnb — “number of nights booked.” Facebook — “daily active users.” All other metrics are subordinate to it.

Funnel: for e-commerce: Impressions → Clicks → Add to Cart → Purchase → Repeat. Conversion at each step is the key to optimization.

Cohort Analysis: grouping users by the date of the first action. Allows you to see how the behavior of cohorts changes over time (retention, LTV).

A/B Testing: scientific method for making product decisions. You create two versions (control and experiment), measure the difference, and make decisions based on statistics. Booking.com conducts thousands of A/B tests simultaneously.

Data Infrastructure

Data Warehouse: structured storage of processed data for analytics. Snowflake, BigQuery, Redshift. Optimized for analytical queries (OLAP).

Data Lake: storage of raw data of all formats. AWS S3, Azure Data Lake. Flexibility due to order.

Data Lakehouse: hybrid—structure of warehouse + flexibility of lake. Databricks Delta Lake, Apache Iceberg.

Practical Assignment

An online store wants to reduce customer churn. Data: transaction history, browsing history, support data, NPS surveys. (1) Which metrics will help understand the reasons for churn? (2) What type of analytics is needed (descriptive/diagnostic/predictive)? (3) How to build a predictive churn model? (4) What to do with clients from the “risk group”?