Journal

Outlook.com refused mail without DMARC: what to fix now

In early November Outlook.com refused mail from domains without DMARC. Microsoft blames a faulty update. Why the gap matters and what to check now.

From 5 November, mail to Outlook.com, Hotmail and Live addresses started to come back. The bounce read 550 5.7.1 Message rejected: Message lacks SPF/DKIM/DMARC Auth Records (AS9031519). It hit small senders too, below the volume for which Microsoft requires DMARC. Microsoft has since put it down to a faulty update of its spam rules and rolled it back. The fault is fixed. The gap it showed may still be open in your own domain, and the first step to close it takes a few minutes.

What happened

Senders and mail providers began logging the refusals on 5 November. On 6 November a user asked on Microsoft's own support forum, Microsoft Q&A, why mail to their outlook.com address was being blocked. A moderator listed as Microsoft external staff answered the same day that a personal address needs no records of its own and that the senders' domains were the likely issue. That was accurate as far as it went, and it did not help: the cause turned out to be on Microsoft's side. Other people in the thread reported the same error code from about 4 and 5 November.

The pattern senders reported was clear, though not every message failed. Mail from a domain without a DMARC record was refused, even when SPF and DKIM passed. EasyDMARC tested it with a sending domain that had both aligned and no DMARC record, and the message bounced at live.com. According to EasyDMARC, business mailboxes on Microsoft 365 with their own domains did not appear to be affected as recipients. The refusals came from the consumer service.

Microsoft did not explain it at first. When CSO Online reported on 9 November, Microsoft had announced no policy change and published no postmaster notice or blog post. It tracked the problem as incident EX1545193 in the Microsoft 365 admin centre, titled "Some users may be intermittently unable to send email messages to Outlook.com, Hotmail.com, or Live.com email addresses". According to the incident text, as quoted by MailRoute and by a user in the Q&A thread, "A recent update to the spam enforcement rules caused email messages to be incorrectly marked and blocked". Microsoft rolled the update back. It gives the impact as 6 November, 18:00 UTC, to 9 November, 11:02 UTC, and closed the incident on 10 November at 05:36 UTC. On 10 November a Microsoft spokesperson said, as quoted by CSO Online: "We've resolved an issue that may have prevented some email messages from reaching Outlook.com, Hotmail and Live addresses".

The incident notice does not mention DMARC. That the faulty rule turned on the missing record is what senders observed, not what Microsoft wrote.

Two things are worth keeping apart. The rule that applies today has not changed. Since 5 May 2025, domains that send more than 5,000 messages a day to Outlook.com must pass SPF and DKIM and publish a DMARC record with at least p=none. Smaller senders are not formally covered. What changed, for less than three days by Microsoft's count, was that the smaller senders were treated as if they were.

Why the mail did not come back by itself

A code starting with 5 is a permanent refusal, so the sending server does not try again once the fault is fixed. MailRoute points out that mail refused during the incident was never delivered and will not be retried.

So the damage is not limited to the days of the fault. An offer, an invoice, a confirmation or a reply to a citizen sent in that window is lost unless someone read the bounce and sent the message again.

What DMARC is, in one paragraph

DMARC is the record in which a domain says how its mail can be recognised as genuine and what a receiver should do with mail that fails. In May 2026 it moved to the IETF Standards Track: RFC 9989, with RFC 9990 and RFC 9991, replaced RFC 7489 of 2015, which had only been informational. The record is a TXT record at _dmarc in front of the domain, so _dmarc.example.ch. SPF says which servers may send for a domain, DKIM signs each message, and DMARC ties both to the address the reader sees in the From line. Without DMARC, a receiver has passing SPF and DKIM results but no statement from the domain owner about what they should mean.

What it means for a company or an administration

If some of your customers, members or citizens use an Outlook.com, Hotmail or Live address, part of your mail may not have reached them between 5 and 9 November. That holds for a company sending offers and for a municipality answering a request.

The larger point is that a domain without DMARC depends on the goodwill of every receiver's filter. In November one filter update at one provider was enough to refuse it for days. Microsoft treats this one as an error. A rule that today covers only senders above 5,000 messages a day can be extended to everyone by a decision instead of a fault. The rule of 2025 was announced in April and enforced from May.

There is a second reason that has nothing to do with Microsoft. Without a DMARC policy that enforces something, anyone can try to send mail with your domain in the From line, and receivers have no instruction from you to refuse it.

The cost of fixing it is small. The cost of fixing it carelessly is not. A record that jumps straight to p=reject before every sender is known can get your own newsletter, your invoicing software or the form on your website refused. The order matters.

Where Elchi Studios stands

EMX, our business mail, checks incoming mail against SPF, DKIM and DMARC, signs outgoing mail with DKIM and sends it from our own servers. Messages are stored encrypted in Geneva. EMX is in beta, and we wrote down what that means.

A DKIM signature is not a DMARC record. Whoever hosts your mailboxes, us included, the DMARC record is published in your domain's DNS by whoever manages that zone. A mail provider can sign every message correctly and your domain can still have no policy at all. That is exactly the case that bounced in November: SPF and DKIM passed, and the record at _dmarc was missing.

So the answer from our side is plain. Mail that leaves EMX is signed. Whether your domain tells receivers what to do with mail that fails these checks is a question for your DNS, and it is worth answering this week whoever your provider is. The same applies to every system that sends in your name, not only the mailboxes. We made a similar point in our post on the end of Exchange Web Services: what matters is knowing what depends on what before a vendor changes it.

What to check this week

  1. Look up your record. Run dig TXT _dmarc.example.ch +short with your own domain, or use any DMARC lookup. No answer means no policy.
  2. Search for the bounce. Look in the sent mail and the system mailboxes for AS9031519 or the words SPF/DKIM/DMARC Auth Records between 4 and 10 November. Send again what was refused. Nothing else will deliver it.
  3. Publish a policy that watches first. If there is no record, start with v=DMARC1; p=none; rua=mailto:dmarc@example.ch. p=none states no preference for how failing mail is handled; the rua address receives the reports.
  4. List everything that sends in your name. Mailboxes, the newsletter tool, the invoicing software, the CRM, the website's contact form, the booking system. Each one has to pass SPF or DKIM for your own domain, not only for its provider's.
  5. Read the reports for two to four weeks. They show who sends with your domain and whether it passes. Fix the senders that fail.
  6. Then enforce. Move to p=quarantine, and later to p=reject, once the reports show only senders you know.
  7. Do not forget the quiet domains. A domain that sends no mail at all, such as an old name or a second spelling, can publish p=reject at once. There is nothing legitimate to break, and receivers are told to refuse mail that uses its name.