How do mailbox providers filter and classify emails?

Mailbox providers use layered filtering systems that evaluate messages at multiple checkpoints. The first layer checks authentication: does the sender have valid SPF, DKIM, and DMARC records? Messages failing authentication face immediate suspicion or rejection.

The second layer examines sender reputation, including IP address history, domain reputation, and past engagement patterns from this sender. Third comes content analysis, where machine learning models scan for spam indicators, phishing attempts, and malware signatures.

Finally, user engagement data influences placement. If recipients consistently ignore, delete, or mark similar messages as spam, the MBP learns to filter accordingly. Each provider weights these factors differently, which is why the same email can hit inbox at Gmail but spam at Outlook.