3 Measuring customers

If you are operating a product or service with repeated interactions with users or customers, then you should be collecting data about those interactions in a data warehouse. Interactions in this context means interactions between the user and the product, service, or platform. (It can also include interactions with other users, mediated by the platform.) It is common to refer to such interactions as events for short, because interactions tracked in a data warehouse invariably have a timestamp telling you when they happened.

I am not going to teach you how to collect that data, but I am going to teach you how to put that data to good use. The first step in using raw data to fight churn is to turn the event data into a set of measurements that summarize the events and collectively produce a profile of the users’ behaviors. These measurements are often called behavioral metrics or just metrics, for short.

3.1   From events to metrics

3.2   Event data warehouse schema

3.3   Counting events in one time period

3.4   Details of metric period definitions

3.4.1   Weekly behavioral cycles

3.4.2   Timestamps for metric measurements

3.5   Making measurements at different points in time

3.5.1   Overlapping measurement windows

3.5.2   Timing metric measurements

3.5.3   Saving metric measurements

3.5.4   Saving metrics for the simulation examples

3.6   Measuring totals and averages of event properties

3.7   Metric quality assurance

3.7.1   Testing how metrics change over time

3.7.2   Metric quality assurance (QA) case studies

3.7.3   Checking how many accounts receive metrics

3.8   Event QA

3.8.1   Checking how events change over time

3.8.2   Checking events per account

3.9   Selecting the measurement period for behavioral measurements

3.10   Measuring account tenure

3.10.1   Account tenure definition

3.10.2   Recursive table expressions for account tenure

3.10.3   Account tenure SQL program