Fix it now
These two codes appear before and during the socket rather than at the logon, so neither is telling you the password is wrong. Something in the path is refusing the connection or tearing it down part way through. Prove the path first, then look at what is sitting in it.
Test-NetConnection outlook.office365.com -Port 993
Test-NetConnection smtp.office365.com -Port 587
ipconfig /flushdns
- If the port test fails, Outlook is not the problem and no setting inside it will help. Test from a phone hotspot to separate the machine from the network.
- If the port test succeeds and Outlook still fails, compare the account’s server, port and encryption against what the provider documents. For Microsoft 365 that is IMAP on 993 with SSL/TLS, POP on 995 with SSL/TLS, and SMTP submission on 587 with STARTTLS.
- Turn off your security product’s mail scanning component – specifically that component, not the whole product – and run one send and receive cycle. If the error stops, you have found it.
- Start Outlook with add-ins disabled and try again, and open the same mailbox in webmail to confirm the server and credentials are healthy.
A VPN or proxy changes both the route the mail takes and what inspects it on the way. Disconnect it and retry before you change anything in the account settings.
If a full send and receive completes cleanly, you are done. If it does not, the next section explains what these codes can and cannot tell you.
Why it happens
Start with something the rest of the internet will not tell you: Microsoft does not publish a meaning for 0x800CCC0E, 0x800CCC0F, 0x800CCC0D or 0x800CCC15. Searching for them returns community questions and answer sites, not documentation. Every table that gives you a crisp one-line definition for each is repeating something somebody inferred. That does not make the codes useless – it means you diagnose from the shape of the failure rather than from the digits.
The shape is worth understanding. A send and receive cycle is an ordinary client and server conversation: resolve the host name, open a TCP connection, negotiate encryption, authenticate, transfer messages, close. A failure that happens instantly and reproduces with a plain port test never got past the second step. A failure part way through a transfer got much further, and points at something with an opinion about long-lived connections: an idle timeout on a firewall, a scanner that dropped the socket instead of releasing it, or a server closing the connection because too many sessions arrived from one client. None of those is visible in the account settings, and all of them show up in the difference between a small message and a large one.
Local mail scanning deserves particular attention because of how it works. Many endpoint security suites insert themselves between Outlook and the network for POP3, IMAP and SMTP by redirecting the connection through a proxy on the loopback address. When it works you never know it is there. When it does not – because the licence lapsed, because it cannot present a certificate the client accepts, or because it is waiting on something – Outlook sees a connection that will not open or will not stay open, and reports it as a network fault.
The port and the encryption type do not match each other
You have this one if The failure is instant, and either the account has never worked or it stopped when the provider changed its documented settings.
- Take the provider’s documented values and enter them exactly rather than adjusting whatever Outlook guessed during setup.
- For Microsoft 365: IMAP outlook.office365.com on 993 with SSL/TLS, POP outlook.office365.com on 995 with SSL/TLS, SMTP submission on 587 with STARTTLS.
- Apply, close every dialog, restart Outlook, and test the incoming and outgoing servers separately so you know which one changed.
Selecting implicit SSL on a port that expects STARTTLS produces a connection that opens and then goes nowhere, which is a reliable way to see a dropped session rather than a refused one.
A mail scanner is proxying the connection and failing at it
You have this one if Everything works with the security product’s mail protection paused, or after an expired trial suite is removed.
- Turn off the POP3, IMAP and SMTP scanning component specifically, leaving file-level protection running.
- Add the vendor’s documented exclusions for Outlook and its data files.
- If the licence has lapsed, either renew it or remove the product with the vendor’s own removal tool. A lapsed suite that still hooks the socket but no longer updates is the worst of both worlds.
- Retest with a message carrying a large attachment, not a one-line test message.
A firewall or the network is blocking the mail ports
You have this one if Test-NetConnection returns False for the mail ports, and the same account works on a phone using mobile data.
- Test from a phone hotspot first, to separate the machine from the network it is on.
- Check outbound rules in Windows Defender Firewall with Advanced Security for anything blocking Outlook or those ports.
- Ask the network team to confirm the ports your provider documents are permitted outbound from your subnet, rather than assuming they are open because they were last year.
The mail host is not resolving, or resolves to the wrong address
You have this one if The failure looks like a connection error but name resolution for the mail host fails or returns an address that does not match the provider’s.
- Clear the resolver cache with
ipconfig /flushdns, then test resolution again. - Inspect the hosts file at
C:\Windows\System32\drivers\etc\hostsfor a stale manual entry somebody added during a migration. - Run
ipconfig /alland confirm the machine is not still using DNS servers left behind by a VPN client that is no longer connected.
The Outlook profile itself is damaged
You have this one if A brand-new profile configured with identical settings connects without complaint.
- Open Control Panel, then Mail, then Show Profiles, and add a new profile.
- Configure the account manually with the provider’s documented settings, and set Outlook to prompt for a profile while you test.
- If the new profile works, attach any local data files to it and retire the old profile rather than deleting it.
Creating a profile does not delete mail. Server-side mail resynchronises and local data files are reattached rather than recreated – but note where they live before you start.
Full reference
Narrowing it down by symptom
| Symptom | Where the fault is |
|---|---|
| Fails instantly and the port test also fails | The network path, a firewall, or the wrong port |
| Port test succeeds but Outlook still fails | Local interception, the wrong encryption setting, or the profile |
| Works with the security product’s mail scanning turned off | The scanner’s mail proxy |
| Only fails on large messages | The session is dropping mid-transfer, usually a timeout or a scanner buffering |
| Fails on one network and not another | Something on that network is inspecting or blocking mail traffic |
| Fails for every account in the profile at once | The machine or its network, not the mailbox |
Settings you can actually check against a source
| Service | Microsoft 365 value |
|---|---|
| IMAP | outlook.office365.com, port 993, SSL/TLS |
| POP | outlook.office365.com, port 995, SSL/TLS |
| SMTP (client) | smtp-mail.outlook.com, port 587, STARTTLS |
| SMTP AUTH client submission | smtp.office365.com, port 587 recommended, TLS enabled |
| Authentication | OAuth 2.0 modern authentication |
If your mailbox is somewhere else, get the equivalent list from that provider rather than from a general article. The point of writing them down is that you then have something to compare the account against field by field, instead of adjusting settings until something works.
Separating the machine from the network from the mailbox
- Open the mailbox in webmail. If that fails too, the problem is the account or the service and nothing on the machine will fix it.
- Open the same account on a phone over mobile data. If that works, the mailbox is fine and the fault is local to the machine or its network.
- Connect the machine to a phone hotspot and retry. If it now works, the fault is the network the machine was on.
- Disable the security product’s mail component and retry. If it now works, that component is the fault.
- Start Outlook with add-ins disabled and retry. If it now works, an add-in is the fault.
- Create a fresh profile. If that works, the old profile is the fault.
Why a code table is not the diagnostic here
It is tempting to read 0x800CCC0D as a name resolution failure and 0x800CCC15 as a socket failure, because that ordering is plausible and it is what every article says. It may well be right. What it is not is documented, and the practical consequence is that you should not spend an afternoon acting on a code alone. If you see 0x800CCC0D, test name resolution – and if resolution is fine, do not conclude the code is lying to you and start rebuilding the network stack. Test the next thing instead.
When it only breaks for one person
- Compare their account settings field by field against a machine that works, rather than against your memory of what they should be.
- Check whether they are on a different network segment, a different VPN profile or a guest wireless network.
- Check whether their machine has a security product the others do not, or a different version of one that everybody has.
- Check the size of what they are trying to send. A single oversized message in the Outbox retries on every cycle and makes an intermittent fault look constant.
- Check whether their mailbox is the one with a protocol switched off, since POP and IMAP are settings per mailbox rather than per tenant.
When a licence is the actual fix
Most of this costs nothing: ports, DNS entries and scanner settings are all free to change, and on a single machine the honest answer is usually to turn off one component of the product you already have. A licence becomes a reasonable answer in one situation – an estate running an expired or consumer suite whose mail component cannot be turned off centrally and keeps being re-enabled locally on machines you cannot visit. Arco supplies business antivirus licences from the main vendors and can tell you which of them proxy mail at the socket and which rely on file-level scanning, so you can pick one that stays out of every send and receive. For one machine, do not buy anything.
Every code this article covers
| Code | What it points at | Source |
|---|---|---|
0x800CCC0E |
Reported when the connection to the mail server does not open. No published meaning; diagnose with a port test from the same machine | not published by the vendor |
0x800CCC0F |
Reported when a session that had started ends before the exchange finishes. No published meaning; treat it as a dropped connection and look for what is holding or timing out the socket | not published by the vendor |
0x800CCC0D |
Seen at the same point in the cycle and commonly associated with the mail host not resolving. No published meaning; test name resolution rather than assuming it | not published by the vendor |
0x800CCC15 |
Seen alongside the two connection codes. No published meaning; treat it as the same class of failure and diagnose it the same way | not published by the vendor |
Confirm the fix worked
- A full send and receive completes with no error dialog.
Test-NetConnectionreturns True for both the incoming and the outgoing port you configured.- A message with a large attachment both leaves and arrives, since a dropped session often only shows on longer transfers.
- The security product’s other components are back on, with exclusions in place, and mail still flows.
- The same test passes after a reboot, rather than only in the session where you changed something.
Questions people ask about this
Does this mean my password is wrong?
No. Both codes appear at or before the connection, which is earlier than authentication. A rejected credential produces a different failure and usually a credential prompt rather than a connection error.
Why does the same account work on my phone?
The phone takes a different network path, has no local mail scanner in the way, and often uses a different protocol entirely. That it works proves the server and the credentials are fine, which narrows the problem to this machine or its network.
Would switching to a different email program fix it?
No. The codes belong to Outlook, but the underlying failure is a socket that will not open or stay open, and any client would meet it. A different client might report it less clearly.
Is there a definitive list of what these codes mean?
Not from Microsoft. The lists you will find are inference, and while the common readings are plausible, they are not something to make an expensive decision on. Diagnose from the behaviour: does the port answer, does it fail instantly or part way through, does it fail on every network.
Do I need to buy anything to fix this?
Usually not. Correcting a port, an encryption setting or a scanner exclusion costs nothing, and the diagnostic commands ship with Windows.
