Why Inbox Placement Differs by Mailbox Provider

Your campaign report is one row. The inboxes are not. Gmail filed it in Promotions, Yahoo filed it in spam, and Outlook sent a 550 before the customer could file anything. Averaging those into “deliverability dropped” sends you into the wrong fix.

One send can produce three different outcomes at three providers, and your ESP report shows all of them as one delivery rate. A split is not always a provider problem, though: if all three move in the same hour, the cause is usually something you share across them, like the list or the sending domain.

Last verified: 26 September 2026, against Google’s sender guidelines, Yahoo Sender Hub, and Microsoft’s 550 5.7.515 article.

Read inbox placement by provider, not in aggregate

InboxEagle seed mailboxes are Gmail, Yahoo, AOL and Outlook. A seed test can show the first two rows, and Outlook folder placement for accepted mail. It cannot show an Outlook rejection, because that happens at SMTP time. AOL follows Yahoo’s sender rules closely enough that Yahoo’s page is the policy reference, and the placement result is still worth reading on its own.

Why the same message gets different answers

Each provider measures complaints on its own users. Gmail’s spam rate is user-reported spam on mail that reached an engaged Gmail inbox. Google’s guidance is to stay under 0.10% and avoid 0.30%. Yahoo’s bulk complaint rate is under 0.30% on mail that reached the Yahoo inbox. Those denominators are different populations. Your win-back that annoyed Gmail users and not Yahoo users moves one chart.

Authentication is also evaluated per receiver. Microsoft’s high-volume rule, enforced since May 5, 2025, requires domains sending more than 5,000 messages a day to Outlook.com, Hotmail.com, and Live.com to pass SPF, DKIM, and DMARC (at least p=none, aligned). Non-compliant mail goes to Junk first, and rejection with 550 5.7.515 follows later. The email deliverability checker shows all three on your domain. A message can satisfy a receiver that accepts p=none and still fail a check you broke only on the Outlook path, or the reverse. Do not “fix Yahoo” by changing a Microsoft-only failure.

Content is the least likely explanation for a one-provider split. The same HTML was just accepted next door. Studies of trigger words, images, and link counts are about patterns across many sends, not about one ISP flipping overnight. Content versus reputation keeps those studies in their lane.

A temporary provider event versus a structural placement problem

A deferral or a short filtering wave at one provider, with clean authentication and a normal complaint rate, can pass while you hold volume. Deferrals are the 4xx case. You still watch the next send.

A structural problem repeats after you retry: spam rate over Gmail’s published line, Yahoo complaints over Yahoo’s line, a standing 5.7.515, or a blocklist the receiver is honoring. A 4xx that clears on a smaller retry is not that. SMTP 4xx versus 5xx is the code. Those structural cases do not clear because you waited overnight. The recovery time page is explicit that providers do not give you a date. The postmortem is where you write which provider it was, and monitoring across providers is how you catch the next split early, so the next incident does not get averaged into “email is down.”

Sources

• Email sender guidelines, Google. Gmail spam-rate guidance.
• Sender best practices, Yahoo. Yahoo complaint-rate guidance.
• Fix NDR 550 5.7.515, Microsoft. Outlook rejection conditions.