A spam filter is a program or set of algorithms that evaluates incoming email and decides whether to deliver it to the inbox, route it to spam, or block it entirely.
Types of spam filters:
- Gateway filters: Applied before email reaches the inbox (e.g., Proofpoint, Mimecast, Barracuda)
- ISP-level filters: Google, Microsoft, and Yahoo's proprietary filtering systems
- Client-side filters: Rules-based filters within the email client itself
How spam filters evaluate email:
- Sender IP and domain reputation
- Authentication status (SPF, DKIM, DMARC pass/fail)
- Content analysis (keywords, link analysis, HTML quality)
- Engagement history (past behavior of the sender with this ISP's users)
- Blacklist status
- Behavioral signals (sending patterns, volume consistency)
Warm-up's effect on spam filters: Building reputation through warm-up essentially teaches the ISP's ML-based spam filter that your domain produces wanted, engaging email — shifting your messages from "unknown/suspicious" to "trusted sender" in the filter's model.