What are provider-specific algorithms?

Every mailbox provider runs its own filter stack. The inputs overlap (authentication, sender reputation, content, user behavior) but the weights, thresholds, and proprietary signals are different. That is why the same campaign can hit the inbox at Gmail and land in junk at Outlook on the same day.

Here is what each of the big four actually leans on.

Gmail. Domain reputation and per-user engagement do most of the work. Google looks at how recipients interact with your mail over time: opens, replies, archives without reading, moves to spam, moves out of spam, and stars. The 2024 bulk sender rules made the floor explicit. If you send more than 5,000 messages a day to Gmail accounts, you need SPF, DKIM, DMARC, one-click unsubscribe (RFC 8058), and a spam complaint rate under 0.3% measured in Postmaster Tools. Cross 0.3% and placement drops fast. Stay under 0.1% and you have headroom. See Gmail's complaint thresholds and the bulk sender guidelines for 2024+ for the specifics.

Microsoft (Outlook.com, Hotmail, Live, Office 365). SmartScreen was retired as a standalone product in 2024, but the filtering philosophy carried into the current Exchange Online Protection and Defender stack. Microsoft weights IP reputation, content patterns, and complaint rates from the Junk Email Reporting Program heavily. You get visibility through SNDS (Smart Network Data Services) and JMRP. Microsoft is also stricter than Gmail about new IPs and sudden volume jumps. A clean Gmail sender can still get yellow or red in SNDS the week they switch ESPs.

Yahoo (also AOL, Verizon). Yahoo runs a complaint-driven model with hard authentication requirements that mirror Gmail's 2024 rules. They publish their stance at postmaster.yahooinc.com. Same 0.3% complaint ceiling, same DMARC requirement at the bulk sender level. Yahoo tends to penalize affiliate-heavy content and shared IP neighborhoods harder than Gmail does.

Apple Mail (iCloud, me.com, mac.com). Apple does not publish a postmaster tool and does not give senders a reputation dashboard. The big variable here is Mail Privacy Protection, which pre-fetches images for most Apple Mail users and inflates open rates. That breaks open-based engagement signals for anyone using opens as a quality input. See the impact of open tracking under Apple MPP for what that does downstream at Gmail.

Why this matters in practice

A fix that helps at one provider can hurt at another. Some examples we see weekly when cleaning client lists:

• Aggressive list pruning based on Gmail engagement removes Apple subscribers who actually opened (MPP inflated their opens, you can't tell who is real).
• Switching to a fresh dedicated IP improves Gmail placement within a week but tanks Microsoft for 4 to 6 weeks because there is no IP history.
• A subject line full of emojis can keep you in the Gmail Promotions tab but trigger content scores at Microsoft.

The practical move is to segment your deliverability reporting by provider. Look at inbox placement, complaint rate, and bounce rate per MBP, not as a single blended number. If Gmail is at 98% inbox and Outlook is at 40%, those are two different problems with two different fixes. Read more on why different MBPs produce different inbox results for the same campaign.

One last thing. None of these algorithms are static. Gmail rolled out relevance filtering in late 2024, Microsoft tightened tenant-level filtering in 2025, and Yahoo aligned with Gmail's 2024 rules within months. Any tactic you build today should be reviewed every quarter against current provider documentation, not a blog post from 2022.