A Microsoft 365 tenant applies Microsoft's filtering and then its own. Anti-spam policy thresholds, transport rules written by somebody's IT contractor in 2019, connection filters, allow and block lists, and quarantine — all configured per tenant, all invisible to you as a sender. If you sell to businesses, this is the filter that decides whether your mail is read, and consumer Outlook seeds do not sample it.
The three outcomes a tenant can produce
| Outcome | What the recipient sees | What you see |
|---|---|---|
| Inbox | Normal delivery | 250 accepted |
| Junk folder | Message in Junk Email | 250 accepted |
| Quarantine | Nothing, unless they check a digest | 250 accepted |
All three look identical from your side. That is the whole argument for seeding inside a tenant: without it, "delivered" covers a message nobody will ever see.
Quarantine is the outcome senders underestimate. Depending on policy, a quarantined message may surface in an end-user digest, or may be held for an admin and never surfaced at all. A campaign can be fully quarantined across a tenant with a 100% delivery rate in your ESP.
Building the seed set
- At least one Microsoft 365 mailbox on a tenant with default anti-spam policy. Default matters: a tenant where your domain is allow-listed produces a guaranteed pass that measures nothing.
- Consumer Outlook seeds as well, if any of your audience is on outlook.com, hotmail.com, live.com or msn.com. They are a separate reading, not a substitute.
- Several mailboxes per context. Per-recipient variation is real; a single seed is a coin flip.
- Seeds excluded from warm-up. A seed your warm-up has been conversing with carries interaction history and reports the best case rather than the typical one.
- The production sending path. Same credentials, IP pool, signing domain, template and tracking setup as the real campaign.
- Folder checks that distinguish Junk from quarantine, and per-seed results kept rather than averaged.
Run it on a fixed cadence so you have a baseline to compare against, and re-run after any deliberate change: new subdomain, new ESP, higher volume, new template family.
What the tenant can be doing to you
When seeds inbox on consumer Outlook and land in Junk or quarantine on a tenant, the cause is usually one of:
- A stricter spam-confidence threshold than Microsoft's default, so borderline mail is junked.
- A transport rule matching a keyword, an attachment type, a sender pattern or an external-sender condition.
- Connection filtering on the sending IP, independent of Microsoft's own reputation view.
- Bulk-mail thresholds that treat campaign-structured mail as bulk and quarantine it by policy.
- An anti-spoofing rule firing because your DMARC alignment is imperfect.
You cannot change any of them. What you can do is fix your side so you clear a default policy comfortably, and — for high-value accounts — ask your contact to have their admin allow-list you. That request is far more credible when you can say which folder your last message landed in.
The authentication floor comes first
Do not interpret tenant seed results until authentication is clean. Microsoft tightened its requirements for high-volume senders to consumer Outlook domains during 2025: SPF, DKIM and a published DMARC policy, with the visible From domain aligned to SPF or DKIM. Anti-spoof rules inside tenants are stricter than the consumer side, so imperfect alignment shows up in tenants first. The explicit rejection is 550 5.7.515 Access denied, sending domain does not meet the required authentication level.
What Microsoft tells you, and does not
SNDS reports per-IP volume, complaint rate, spam-trap hits and a colour grade for IPs you can prove you own. JMRP forwards you a copy of every message an Outlook user marks as junk. Those are the Outlook-side equivalents of Google's Postmaster Tools, and Microsoft publishes no checkable blocklist at all. Neither tool reports tenant placement, and neither covers shared ESP IPs you do not own.
For contrast on the Google side: Postmaster Tools v2 reports spam rate, authentication, delivery errors and compliance status, plus a Deliverability analysis report added in June 2026 — and the domain and IP reputation dashboards are permanently gone, retired with v1, which stopped serving senders on 30 September 2025 and was gone by the end of 2025.
What not to do
- Do not treat outlook.com placement as Microsoft 365 placement. They are two filters in series, and only one of them is yours to influence.
- Do not seed a tenant that has allow-listed you. You have measured your own configuration.
- Do not equate a `250` with delivery. Quarantine accepts the message and shows nobody.
- Do not reuse warm-up mailboxes as seeds.
- Do not ask an admin to allow-list you instead of fixing alignment. A tenant will not make an exception for a sender that fails anti-spoofing, and the next tenant will not either.
- Do not expect seeding to change the result. It reports; the warm-up, the list and the authentication change it.
How MailPilot helps
MailPilot warms Microsoft 365 and Outlook mailboxes over OAuth, or any SMTP/IMAP sender, at a ramp that avoids the volume spikes Microsoft reacts to, and measures Microsoft placement as its own number through a seed-inbox network kept separate from the warm-up traffic. SPF, DKIM and DMARC alignment is checked on the sending domain, which is the condition tenant anti-spoofing rules enforce most strictly.
Ready to reach the inbox every time?
Automated email warmup across real peer mailboxes. Live inbox placement monitoring. Free 7-day trial - no credit card required.
Start free trial