Assessing Whether Friend-Authored Content Is More "Social" Than Unconnected Content

Background

You have impression- and engagement-level logs for a social feed. Posts are created by authors and viewed by audiences who may be either:

• Connected to the author (e.g., friends/follows), or
• Unconnected viewers (no direct connection).

The product team wants to know if posts from friends lead to more social interactions, and what additional value unconnected audiences provide.

Task

Outline how you would validate, using available data, that content from friends is more social than content from unconnected viewers. Include:

1. Clear definitions of "socialness" and the primary metrics.
2. Hypotheses to test and appropriate statistical methods.
3. Concrete, step-by-step measurement for connected vs. unconnected audiences, including controls for confounders.
4. If engagement from unconnected viewers is lower, propose alternative value they provide (e.g., reach, discovery, network growth) and how to quantify it.

Hints

• Define socialness (e.g., reactions per view, reshare rate, comments per view, thread depth).
• Use A/B-style comparisons where possible; otherwise, apply robust observational methods controlling for confounders.
• Consider reach, content discovery, and network growth as alternative value.