Email Migration Checklist: 8 Steps for Deliverability

Email migration checklist for ESP and domain moves: inventory senders, stage authentication, carry over suppressions, and monitor inbox placement.

Published on

•

Email Migration Checklist: 8 Steps for Deliverability
Do not index
Do not index
A team moves email to a new platform on Friday. DKIM validates, the cutover finishes on time, and Monday's dashboard shows a delivery rate above 99%. By Wednesday, customers are reporting that password reset emails never arrived, although the application logged them as sent.
The receiving servers accepted the messages, but acceptance does not confirm inbox placement. A message may go to spam, a promotions tab, or a quarantine folder instead. Mailwarm's network study reported that 26.5% of emails delivered to the inboxes it monitors during the seven days ending October 14, 2026, were filtered to spam. That figure is a snapshot of inbox behavior, not a verdict on any single sender.
An email migration checklist should account for these separate failure points, from transactional streams to tracking domains and suppression data. The eight steps below cover the work to complete before, during, and after a move. Teams seeking expert support can work with a deliverability team; smaller senders can also use this email deliverability checklist for SMBs alongside the migration-specific steps here.
Table of Contents

1. Inventory every sending stream and domain before cutover

Most migration failures begin with a stream nobody wrote down. Marketing gets mapped, lifecycle flows get rebuilt, and then an invoicing tool or a support desk keeps sending from the old subdomain long after the DNS was cleaned up. Those messages fail authentication, and the first person to notice is a customer waiting on a receipt.
The planning gap is not unusual at the project level either. In the Cloud Migration 2025 study published by IDG's CIO and Computerwoche with T-Systems and Lufthansa Industry Solutions, 38% of surveyed companies had launched or completed at least one cloud migration project, while complexity delayed 38% of projects (Cloud Magazine).

Build a stream register, not a mailbox list

A stream register lists every source that sends under the brand. For each one, record the sending domain and subdomain, the From address, the platform or application that sends it, the IP or shared pool, the tracking domain, and one named owner.
  • Pull 90 days of sending logs from each ESP and from the application itself. Senders you forgot show up there, not in the kickoff deck.
  • Assign one owner per stream, not per team. Shared ownership is why a stream goes unmigrated.
  • Check each domain's authentication status before you change anything, so you know the starting point rather than guessing at it.
  • Flag every stream that shares the root domain with marketing mail. Those are the ones that will take damage together.
notion image
Say a SaaS company moves from one ESP to another and maps campaigns, lifecycle, and sales sequences. Its billing platform sends receipts from a subdomain that never appears in the plan. When the old sending setup is retired, receipts begin failing alignment and land in spam for a week before anyone connects the two events.

2. Plan the SPF, DKIM, and DMARC cutover as a staged change

A missing record is easy to spot. A record changed out of order can break one stream while another keeps passing. When marketing moves to a new ESP but transactional mail stays on the old platform, each may depend on different keys and SPF entries.
Google's sender requirements provide a baseline. Since February 1, 2024, senders delivering to personal Gmail accounts must use at least SPF or DKIM. Senders delivering more than 5,000 messages per day to Gmail must use SPF, DKIM, and DMARC, with the visible From domain aligned to the SPF or DKIM organizational domain. Google also expects user-reported spam rates below 0.1% as a working target and below 0.3% as a critical threshold (Google's sender guidelines).

Change one record, then verify it

Make DNS changes in sequence, with a check after each one. Publish and validate the new DKIM key before disabling the old key. Add each sending service to SPF rather than replacing the existing include list wholesale. Keep DMARC on a monitoring policy until reports show clean alignment across every stream. A focused authentication audit can help identify which records and streams need attention before changes begin.
Send test messages to Gmail and Microsoft 365, then inspect their headers to confirm SPF, DKIM, and DMARC results. A checker can help catch record errors between sends. Use Microsoft 365 DNS checks for senders as a reference when reviewing that environment.
notion image
For example, marketing moves to a new ESP while transactional mail remains on the old platform. If both streams use different keys, deleting an SPF include that appears redundant can cause transactional mail to fail authentication. Confirm which stream uses each record before removing it.

3. Migrate suppression lists, unsubscribes, and consent status

The most damaging migration mistake is also the easiest to avoid, and it happens in the data import. Contacts get moved without their suppression status, so the new platform sends to people who opted out years ago or whose addresses hard bounced repeatedly. Complaint and bounce rates then climb on the new domain from the very first campaign.
Tool selection gets discussed as a speed question, but data quality is what practitioners rank first. In a survey of more than 100 managed service partners across North America, Australia/New Zealand, and Europe run for the AMI program, 60% identified overall data quality as the key driver of tool selection, and more than half ranked user experience as the first or second most important feature (AMI survey report).

Export by reason, import by reason

Export suppression data with its reason code, date, and source, then reconcile counts per category rather than as a single total.
  • Global unsubscribes must apply to every list, not only the list where the contact originally opted out. Test this with one address before the first live send.
  • Hard bounces, complaints, and manual suppressions are separate categories. If the new platform collapses them into one flag, you lose the ability to suppress complaints separately from invalid addresses.
  • Re-verify stale contacts before import. Anything that hasn't engaged across two full sending cycles is a candidate for email list building work rather than a straight transfer.
  • Compare import counts against export counts, category by category, and sign off before campaigns are scheduled.
notion image
Picture a marketing team that exports contacts but leaves the unsubscribe column out of the file, then imports into a new ESP and sends a win-back campaign. Every complaint that follows is recorded against the new domain, which has no reputation history to absorb it.

4. Set up custom tracking and link domains on the new platform

Open and click tracking, plus image and link hosting, is the part of a migration that gets skipped most often. Leave it on the platform default and every link in your email points to a domain you don't control, shared with thousands of other senders on the same infrastructure. If one of those senders triggers a block, your links break with them.
A custom tracking domain aligned with your sending domain gives mailbox providers a consistent signal and keeps link reputation under your control. Pick a subdomain you can explain, such as links.example.com, confirm its certificate is issued and renewed automatically, and verify that it resolves before the first live send rather than after.

Keep the old domain resolving

Previously sent emails stay clickable for months. Old bookmarks, forwarded messages, and archived campaigns all point at the previous tracking domain, so retiring it on cutover day turns working links into dead ends for customers.
The exact configuration steps differ by platform and change often, so follow the platform's official help center for setting up a custom tracking or branded link domain rather than a tutorial written two years ago.
A brand launches a welcome series from a new platform with default tracking. Links route through a platform-owned redirect domain shared with thousands of senders, one of which gets blocked, and the brand's new subscribers hit warnings on the first click.

5. Run a parallel send and move traffic in controlled stages

Switching every stream on the same day guarantees you can't tell which change caused a problem. Move one stream or one audience segment at a time, keep the old setup live for everything not yet moved, and start with the most engaged contacts, because they generate the positive engagement signals that shape how a mailbox provider judges a new sender.
The case for a measured pilot is well documented on the infrastructure side. Microsoft's migration guidance notes that mailbox extraction is CPU and memory intensive and can affect end users when the source system is already loaded, and it recommends monitoring performance during a pilot instead of opening with an aggressive schedule. Default migration concurrency is 20 mailboxes, configurable upward through PowerShell, but pushing concurrency without validating source capacity, network throughput, and throttling can slow the whole job down (Microsoft 365 migration best practices).

Define the baseline and the thresholds first

Record open, click, complaint, and bounce rates per receiving provider for at least four weeks before the move. Write the go and no-go numbers down before the first stage runs, not during it. A staged rollout for agency teams benefits from the same discipline described in this platform migration strategy, where stages and owners are agreed in advance.
  • Stage by engagement, not by list size. Your most engaged 10% gives you clean signal on a small volume.
  • Keep the old platform active until the new one has run a full sending cycle cleanly, including a monthly newsletter and a trigger-based flow.
  • Compare per provider. Gmail and Microsoft 365 often diverge during the same migration, and a blended rate hides both patterns.
A team moves its most engaged segment first and sees complaint rates rise only among Microsoft 365 recipients within two sends. It pauses that stage, checks DMARC alignment for the stream in question, and continues once the authentication gap is closed.

6. Protect transactional email during the switch

Password resets, order confirmations, shipping notices, and account alerts usually belong to the product or application team, not marketing. They're business critical, time sensitive, and noticed within minutes when they stop. During a migration they need their own plan, covering sending domain, authentication, bounce handling, and the webhook endpoints that record delivery events.
The authentication bar is worth checking stream by stream. Microsoft's requirements for high-volume senders to Outlook.com, Hotmail.com, and Live.com consumer mailboxes require SPF to pass and accurately identify authorized hosts, DKIM to pass so message integrity can be validated, and DMARC to publish at least a p=none policy aligned with SPF or DKIM, preferably both (Microsoft's high-volume sender requirements). A record showing SPF is present proves nothing on its own.

Verify credentials, not just templates

The most common transactional failure in a migration isn't a rendering bug. It's an SMTP credential or API key still pointed at the old platform after the account was closed on schedule.
  • Separate subdomains. Transactional mail on its own subdomain cannot be damaged by a marketing complaint spike, and reputation stays attributable.
  • Test every template in the new environment, including edge cases like multi-language resets and attachment-heavy invoices.
  • Confirm webhook delivery so bounces and complaints from transactional sends reach your suppression logic.
The sequence that keeps resets working through a cutover is short and worth following in order.
notion image
A SaaS product moves marketing email to a new platform and leaves password resets on the old SMTP account. The old account closes on the planned date, and resets stop arriving for a subset of users, in some cases for hours, until the application config is updated.

7. Configure bounce, complaint, and deferral handling on the new setup

After a cutover, delivery can fail because the new platform lacks the old system's suppression rules, alert routing, or retry behavior. For example, an address suppressed on the old platform may receive another message if its hard bounce history was not migrated. Set up and test each handling rule before shifting traffic.
Rate limits can trigger deferrals, especially when a notification burst is retried aggressively. Exchange Online sets a recipient rate limit of 10,000 recipients per mailbox per day and a message rate limit of 30 messages per minute. The recipient count for a single message can be configured up to 1,000 (Exchange Online limits). Throttling can resemble a reputation problem in a dashboard, so check the response text before changing sending volume or retry settings.

Read the bounce text, not the status code

Enhanced status codes provide a category; the human-readable response often identifies the specific cause, such as a rate limit, policy rejection, reputation block, or full mailbox. Route that detail to someone who can take action.
  • Suppress hard bounces on the first event. Confirm that the new platform applies the rule immediately.
  • Send complaints to a review queue. Use complaint signals to identify a problematic list segment, then suppress the complainant.
  • Watch deferrals by receiving provider. Repeated deferrals at one provider can point to throttling or list-quality problems. Adjust volume and review the affected list before changing retry behavior.
  • Treat hard bounce rates above 2% as a reason to investigate the source list. Review where the addresses came from and whether the list needs cleaning.
A deferral that repeats for 24 hours warrants investigation. If it continues for four days, check whether the receiving provider is blocking the traffic.

8. Verify placement, reputation, and ownership after the cutover

The migration isn't finished when the last campaign sends. Reputation builds over weeks, and enforcement has become less forgiving. Google's escalation of enforcement from November 2025 includes temporary or permanent rejection risk for non-compliant traffic, including traffic from senders approaching bulk-sender volume to personal Gmail accounts (Google Workspace email sender guidelines).
Segment the monitoring. Google's bulk-sender rules apply to mail sent to personal Gmail accounts, while Google Workspace accounts are evaluated differently, so one blended delivery rate across all recipients tells you very little (Gmail sender guidelines). Report placement and complaints by recipient domain and by stream, and read the domain reputation data in Google Postmaster Tools alongside Microsoft's sender reputation views for the rest of your list.

One owner, one register, one weekly review

Post-migration, the risk shifts from technical to organizational. Marketing, product, IT, and outbound teams all make changes to a shared domain, and nobody is accountable for the aggregate effect.
  • Review placement daily for two weeks, then weekly. Daily only works if one person owns the number.
  • Require a register entry before any new sender goes live. That single rule prevents most of the slow degradation that follows a successful migration.
  • Keep rollback criteria written down for each stream, with the thresholds and the person who can trigger them.
  • Compare against the pre-migration baseline rather than last week, and treat a plateau in one provider as an open question.
Two weeks after cutover, a team sees engagement on one provider drop while others hold steady. Checking the email deliverability fundamentals against the actual setup, it finds an unverified link domain, fixes it, and recovers inside the next two sends. In another team, an outbound tool starts sending from the main domain without review, complaints rise, and the change log is the only reason the unauthenticated sender is found quickly.

What each migration step protects, and what to watch after cutover

Use post-cutover checks to catch failures the migration plan cannot prevent on its own. A sending stream can authenticate correctly while its links fail, transactional messages stall, or suppression rules stop working. Set a baseline before the move, then compare each stream against it.
Migration step
Watch after cutover
Inventory every sending stream and domain before cutover
Investigate any stream that is missing from reporting or shows a new authentication failure.
Plan the SPF, DKIM, and DMARC cutover as a staged change
Check authentication results for each stream. Pause its traffic if SPF, DKIM, or DMARC fails or alignment changes unexpectedly.
Migrate suppression lists, unsubscribes, and consent status
Compare complaint and bounce rates with the pre-migration baseline. Stop the affected send if either rises.
Set up custom tracking and link domains on the new platform
Test link resolution and SSL, and confirm tracking data arrives. Pause links that fail or point to the wrong destination.
Run a parallel send and move traffic in controlled stages
Compare placement, engagement, complaints, and bounces by provider. Hold the next traffic increase if a provider falls below its baseline.
Protect transactional email during the switch
Confirm delivery events arrive and check bounces and template rendering. Keep the old path available until critical messages work on the new setup.
Configure bounce, complaint, and deferral handling on the new setup
Check that invalid addresses are suppressed and provider deferrals trigger review. Pause affected traffic if complaints or bounces rise.
Verify placement, reputation, and ownership after the cutover
Review placement, reputation, engagement, and authentication trends against the pre-migration baseline. Assign an owner to every unexplained decline.

Treat the Migration as the Start of Monitoring

After cutover, assign one person to track deliverability by sending stream and recipient provider. They need clear escalation thresholds and authority to pause or adjust sends when placement, complaints, bounces, or deferrals shift.
A migration can expose ownership gaps: no one may be responsible for a transactional stream, a tracking domain, or the suppression data that shapes who receives mail. Keep those responsibilities visible after launch, and review performance as a continuing part of the migration rather than treating the DNS change as its finish.
MailAdept embeds a dedicated deliverability expert in the customer's team to audit the setup, address technical issues and reputation risks, and improve inbox placement. Once the core setup is in place, MailAdept monitors email health, sender reputation, blacklist exposure, authentication health, and inbox placement trends every day, with weekly expert reviews. Start with a free deliverability audit at MailAdept.

Fix Your Email Deliverability Before It Costs You Revenue

Get expert insights on why your emails go to spam and how to consistently reach the inbox.

Get a Free Deliverability Audit