You sent a "Upgrade to Premium" email to 5,000 users. 200 upgraded within 7 days.
Question: Did the email cause those upgrades, or would they have upgraded anyway?
The Problem: Users who received the email aren't random. Your email system targeted:
- Active users (logged in 5+ times last week)
- Power users (used advanced features)
- Engaged users (high session duration)
These users were already more likely to upgrade. We need PSM to separate email impact from user intent.