Why “last activity” sounds simple
Last touch before purchase is a classic question. Marketing wants attribution. Product wants behavior insights. I’ve seen teams answer it confidently — and incorrectly.
Where it usually goes wrong?
- Events after purchase
- Events from another session
- Events triggered by the purchase itself
Suddenly “last interaction” happens after the conversion. What BigQuery forces you to respect?
You must clearly separate:
- Before purchase
- After purchase
Anything else is storytelling, not analysis.
How I approach it now
For each user, I lock the purchase timestamp. Then I only look backwards.
- Strict time filter
- Ordered events
- One last valid action
That’s when last-touch analysis starts making sense.
Want to get all my top Linkedin content? I regularly upload it to one Notion doc.
Go here to download it for FREE
