Skip to content

Book a call

Email bounce codes in a campaign: 5.x.x, 4.x.x and when to pause sending.

An email bounce is the return of a message by the recipient's server together with a code that says whether the problem is permanent, as with 5.x.x codes, or temporary, as with 4.x.x codes. An address with a permanent error is removed from the campaign, a temporary error is left to the server to retry, and a rejection due to the recipient's policies requires checking the sender. When bounces build up quickly, it is better to pause sending before they harm the domain.

Published
Updated
Reading time
8 min
Category
Deliverability
Author
DYNTRO team

What are email bounces and where do they come from?

A bounce is a notice from a mail system that a message did not reach the addressee. In its Exchange Online documentation, Microsoft calls it a nondelivery report (NDR). In a B2B campaign, the cause may be a mailbox that no longer exists, a company domain after a name change, or a filter that did not accept the message.

The recipient's server can refuse a message straight away, during the connection with the sender's server. It can also accept it and send the nondelivery report back to the sender's return address later, sometimes after several hours or days. M3AAWG notes in its sender best practices that such late bounces tend to be more common in B2B mail. That is why bounces are checked not only on the day of sending but throughout the campaign.

How do you read email bounce codes?

A bounce usually contains a three-digit SMTP reply code from the RFC 5321 standard and an enhanced status code from RFC 3463, written with dots. An example line from a bounce:

550 5.1.1 Bad destination mailbox address

The first digit tells you the most:

  • 2 means success: the message was accepted;
  • 4 is a temporary error: the message is valid, but a temporary condition is in the way, so a later attempt may succeed;
  • 5 is a permanent error: resending the message in the same form is unlikely to achieve anything.

The second number indicates the area of the problem: 1 is the address, 2 the mailbox, 3 the mail system, 4 network and routing, 5 the protocol, 6 the content, and 7 security and policy, e.g. the recipient's filters. The third specifies the cause. Microsoft describes it the same way: a 5.x.x code is a permanent error and 4.x.x a temporary one.

The code alone is not yet a diagnosis. M3AAWG points out that codes usually follow the standard, but each receiver writes the text description in its own way. Only the description shows whether the issue is the address or, for example, the sender's IP address being on a blocklist.

What do the most common codes mean?

CodeWhat it meansWhat to do
5.1.1no mailbox exists at this addressremove the address, do not retry
5.1.2the destination domain or server does not exist or does not accept mailremove the address, check the company's domain in the contact list
5.1.10Microsoft 365 could not find the recipientremove the address
4.2.2mailbox fullleave the retry to the server
4.4.7the message expired in the queue before it could be deliveredcheck the address and the recipient's server
4.7.xthe recipient is temporarily not accepting mail because of its policies, e.g. after detecting suspicious activityslow down sending and look for the cause
5.7.1the sender is not permitted to send to this recipientdo not retry, read the description
5.7.23, 5.7.509rejection after an SPF or DMARC failurecheck the sender's domain records

Codes 5.1.1, 5.1.2, 4.2.2, 4.4.7 and 5.7.1 come from RFC 3463. Codes 5.1.10, 5.7.23 and 5.7.509 are Exchange Online codes from Microsoft's documentation, and other mailbox providers may describe the same situations with different codes.

What do you do with each type of bounce?

Permanent address error.

5.1.x codes say that the address does not exist or the domain does not accept mail. RFC 5321 recommends not repeating the same request after a reply with a 5 code. M3AAWG adds that a “no such user” address must be excluded from further sends, and that a large number of permanent errors lowers the reputation of the sender's IP address and often indicates an old or poorly collected list.

So the address is removed from the campaign and flagged in the contact list so that it does not come back in the next one. It is also worth taking a wider view. If all the addresses at one company written in the same format bounce, the problem is the format or the domain, not an individual person.

Temporary error.

4.x.x codes indicate a temporary obstacle, e.g. a full mailbox or an overloaded server. The sender's server then retries on its own, without manual resending. RFC 5321 recommends an interval of at least 30 minutes between attempts and giving up only after at least 4-5 days. M3AAWG points out that some receivers deliberately reject the first attempt temporarily (greylisting) to check whether the sender behaves according to the standard. So a 4.x.x error is not treated as permanent. However, if the same address bounces on successive steps of the sequence, it is better to exclude it.

Policy rejection.

x.7.x codes relate to security and policy, e.g. filters on the recipient's side. When a rejection concerns a single address, e.g. a group mailbox that only accepts mail from selected senders (Microsoft gives this example for code 5.7.1), the address is skipped. When similar rejections come from many receivers, the cause lies on the sending side: in the SPF, DKIM or DMARC records, in the reputation of the IP address or domain, sometimes in the content. M3AAWG advises reading the description of such bounces, because it may point, for example, to a blocklist from which the sender's IP address has to be removed before mail reaches recipients at that domain again.

Is an auto-reply a bounce?

No. A bounce is sent by the mail system when it cannot deliver a message. According to RFC 5321, such a notification goes to the sender's return address and itself has an empty return address, so that it does not trigger further notifications in a loop. An auto-reply, e.g. “out of office”, is sent by a mailbox that has accepted the message. RFC 3834 describes automatic responses separately and notes that delivery status notifications (DSNs) are covered by a different standard. It also recommends marking auto-replies with the header:

Auto-Submitted: auto-replied

An auto-reply therefore tells you that the mailbox works, and it does not count as a bounce. A separate case is a message saying that the person no longer works at the company: the address works, but further steps to that person make no sense.

When should you pause sending?

It is worth checking bounces as you go, on every sending day, and comparing them with previous days of the same campaign. There is no single bounce rate at which you have to stop. What matters is a sudden change and the type of codes. It is better to pause sending and establish the cause first when you see:

  • many 5.1.1 codes from one batch of the contact list or one segment: the contact list is out of date or badly built;
  • many x.7.x rejections from different receivers: a problem with authentication, reputation or a blocklist;
  • a series of 4.7.x codes and deferrals at one mailbox provider: the receiver is limiting how much mail it accepts from the sender.

M3AAWG recommends tracking bounces separately for each mailbox provider, because a rise at one of them often points to a specific cause, e.g. an authentication error, a missing PTR record or a poor IP address reputation. In its sender guidelines, Google recommends reducing sending when messages start to bounce or are deferred, and increasing it again slowly as the errors decrease. If bounces do not go away even at low volume, individual messages need to be reviewed.

After pausing, the order of steps is simple:

  1. Group the bounces by code and mailbox provider.
  2. Remove addresses with permanent errors, and check the next batches of the contact list before sending, as described in our article on email address verification.
  3. For policy rejections, check the sender's domain records and blocklists.
  4. Resume sending with a smaller batch and watch the bounces over the following days.

Few bounces are the result of the earlier campaign stages: a verified contact list and prepared domains. We describe deliverability as part of the service on the B2B email campaigns page.

Example of a bounce summary from one campaign day

Example company, people and addresses; domains example.com and example.org.

AddressCode or textTypeDecision
a.nowak@example.com5.1.1permanent address errorremove from the campaign and flag in the contact list
p.wisniewski@example.org4.2.2temporary, mailbox fullleave the retry to the server
j.kowalski@example.com5.7.1recipient policyread the description, check whether it is a single address
m.zielinska@example.org“out of office until 6 October”auto-reply, not a bouncecounts neither as a bounce nor as a reply

Summary.

  • A 5.x.x code is a permanent error and 4.x.x a temporary one. The second number tells you what the problem concerns: the address, the mailbox, the network or security policies.
  • An address with a 5.1.x code is removed from the campaign and flagged in the contact list. 4.x.x errors are retried by the server, without manual resending.
  • x.7.x rejections from many receivers point to a problem on the sender's side.
  • An auto-reply is not a bounce.
  • When bounces suddenly increase, it is better to pause sending, find the cause and resume with a smaller batch.

We run email and LinkedIn campaigns to decision-makers, from data and domains to replies, while your salespeople focus only on conversations. See the service: B2B email campaigns.

Book a call

Questions and answers.

Is an auto-reply a bounce?

No. A bounce is sent by the mail system when it cannot deliver a message, and it includes an error code. An auto-reply, e.g. “out of office”, is sent by a mailbox that works and has accepted the message. RFC 3834 recommends marking such messages with the header Auto-Submitted: auto-replied. An auto-reply does not count as a bounce. In practice, the exception is a message saying that the person no longer works at the company: the address works, but further steps to that person make no sense.

Should you resend after a 4.x.x error?

Not manually. A 4.x.x code means that the message is valid and the obstacle is temporary, so the sender's server retries on its own. RFC 5321 recommends an interval of at least 30 minutes and giving up only after at least 4-5 days. If the same address bounces on successive steps of the sequence, it is better to exclude it. M3AAWG considers it good practice to remove an address after at least two consecutive bounces over a period of at least two weeks.

Sources.

Related articles.

All articles

Let's start talking to the right companies.

Leave your details.

Your answer helps us see which channels bring enquiries.

DYNTRO sp. z o.o. processes the data from this form to reply to your enquiry. More in our privacy policy.

What happens next.

  1. On the first call, we define your customer profile, market and exclusions.
  2. We choose the scope: email, LinkedIn or both channels, in one or several segments.
  3. If working together makes sense, you receive a proposal with terms for your company.

Prefer email? kontakt@dyntro.pl