Context - Facebook in 2004-2006

Untitled

Phase 1 — The Question to Understand the Problem

• The Wrong Question
• The Correction

This question reframing was everything for them. It shifted the analysis from who users are (demographics, college, device) to what actions they took early. It became a behavioral cohort analysis, rather than a segmentation study.

Phase 2 — Analyzing the Data & The Insights

• Step 1: Split users into cohorts
• Step 2: Replay their first 10 days
• Step 3: Correlate behaviors with retention
• Step 4: Find the signal

The findings: Users who connected with 7 or more friends within their first 10 days were dramatically more likely to become long-term active users. Below that threshold, most churned.

<aside>

The “magic number”

It wasn’t a product requirement, a guaranteed cause of retention, a universal truth for all social apps or discovered in a single afternoon.

But rather, it was:

• A proxy for the "aha moment" — feeling the social graph come alive
• A leading indicator to optimize onboarding around
• A signal that the network effect had clicked for that user
• A north star metric for the growth team's experiments </aside>

The friends you had on Facebook determined whether the product was useful or useless. 7 friends was the threshold where your feed became genuinely interesting every time you opened the app.

Phase 3 — Project Sense & Features

The insight wasn't just intellectually interesting but it gave the team a specific, measurable target to engineer toward. The question became: how do we get every new user to 7 friends in 10 days?

<aside>

Feature 1

"People you may know"

• Surfaced in feed and on profile pages
• Used mutual friends + network graph to rank suggestions
• Reduced friction to zero, one click to send request
• The most direct lever to hit 7 friends faster </aside>

<aside>

Feature 2

Email contact importer

• Import Gmail, Yahoo, Hotmail contacts
• Show which contacts are already on Facebook
• Batch friend request to known people instantly
• Collapsed the friend-finding funnel dramatically </aside>

<aside>

Feature 3

Onboarding friend step

• New users shown friend suggestions before seeing feed
• Couldn't easily skip past the friend-adding step
• Profile completion tied to adding friends
• Engineered the first 10 days deliberately </aside>

<aside>

PM lesson — How a problem turned into a solid Feature

None of these features were complex to build. The hard part was the analysis that revealed what to build. Once the team knew the metric (friends in first 10 days), every product decision had a clear test: "does this move more users past 7 friends faster?" That's the power of a well-defined activation metric , it turns subjective product debates into measurable experiments.

</aside>

PM frameworks this story teaches

<aside>

Cohort retention analysis

• Segment users by behavior, not demographics
• Compare retained vs churned cohorts on early actions
• Look for leading indicators, not lagging ones
• Tool: retention curves split by behavior buckets </aside>

<aside>

Magic number / activation milestone

• Every product has one — find yours via data
• Twitter: follow 30 people. Slack: 2,000 messages sent. Dropbox: 1 file synced
• It's a proxy for value, not value itself
• Optimize onboarding to reach it faster </aside>