A 20,000-recipient campaign sent from a new provider can change your sending volume, IP path, and audience response in one move. If the same cutover also introduces new DNS records, a drop in inbox placement can be hard to trace. This guide to sender reputation during migration gives you a staged cutover plan, example DNS records, and measurable pause points.

Sender reputation is not a portable score that moves between providers. Keep the established sending domain, authentication, recipient consent, and suppression practices consistent where possible; shift traffic in stages; and measure the new provider’s results against a baseline. A new IP and provider-specific sending history must be established through sending behavior and recipient response.

How to preserve sender reputation during migration

Sender reputation comes from several signals, not one score you can transfer to a new provider. Your domain identity and sending practices can remain familiar, while a new provider’s infrastructure and sending history establish themselves through actual mail and recipient response. Plan the move around those different kinds of continuity instead of expecting a new route to inherit the old route’s results.

It helps to distinguish two signals:

  • Domain reputation is associated with the domains used to identify and authenticate messages. Keeping the established visible From domain and authentication consistent gives receiving systems a familiar identity to evaluate. Recipient response and message quality still affect how that mail is treated.
  • IP reputation is associated with the sending IP address. When a provider changes the IP, the new address has its own sending history. With shared infrastructure, other senders may also use the pool, and you do not control an individual IP’s full history.

Authentication helps identify authorized mail; it is not a reputation score. DKIM supports multiple public keys for a signing domain through selectors. RFC 6376 describes publishing old and new keys concurrently during a key transition, switching the outbound server to the new private key, then removing the old public key after the transition. That helps receivers validate signatures across a key change; it does not transfer IP history. RFC 6376

SPF authorizes specified sending systems for a domain identity, while DMARC compares SPF and DKIM domain identifiers with the visible From domain. They help receivers evaluate authentication, but they do not create a portable measure of inbox placement. Google Postmaster Tools is one provider-specific view: it reports on messages sent to personal Gmail accounts, not every mailbox provider. Its data is typically updated within 24 hours after a sending change, but can take longer. Google Postmaster Tools

In an email service migration, preserve the parts you control—domain identity, authentication, consent records, suppressions, and sending practices—and treat the new provider’s IP and sending history as separate. Volanea sends through Amazon SES, so moving to Volanea changes the sending infrastructure even if your From domain remains the same. The practical objective is to avoid changing several major signals at once and to catch a negative change before it spreads to more traffic.

Audit your current sending setup before the move

Build a baseline before changing providers so you can compare like with like afterward. Record each sending domain and subdomain, the visible From domain, the envelope MAIL FROM domain, and the sending IP addresses used by each system. Separate transactional and campaign traffic; a combined daily total can hide a sharp change in one stream behind steady volume in another.

For each domain, save the complete SPF TXT value, DKIM signing domain and selector, and DMARC record and policy. Record which provider or application sends each message type. A DKIM selector identifies the public key used for a signature: RFC 6376 gives foo.bar._domainkey.example.com as the DNS lookup name for selector foo.bar on example.com (RFC 6376).

Track daily sending totals and peaks by message type, recipient domain, sending IP, and sending domain where your logs support it. Include hard bounces, soft bounces, complaints, deferrals, delivery timing, and any inbox-placement results. Record the calculation method and reporting window too: comparing a one-day campaign spike with a monthly average can conceal a real change.

A compact baseline might look like this. These figures are an illustrative example for a fictional SaaS company, not recommended targets or industry benchmarks.

StreamCurrent route and volumeBaseline observationsMigration comparison
Password resetsProvider A; 4,000/dayHard bounce 0.3%; complaints 0.01%; delivery usually under 2 minutesCompare reset mail separately by recipient provider and delivery time
Receipts and alertsProvider A; 6,000/dayHard bounce 0.4%; complaints 0.01%; transactional templatesCompare each event type and watch for duplicate sends
Weekly campaignProvider B; 20,000/sendHard bounce 0.8%; complaints 0.04%; 30-day engaged audienceCompare the same audience segment and campaign type

Save dashboard screenshots or exports with their date ranges. Google Postmaster Tools data applies to personal Gmail accounts, including @gmail.com and @googlemail.com; some dashboard data can be omitted when sending volume is low. Its available dashboard ranges are the last 7, 30, 60, 90, or 120 days, so note the chosen range when comparing results (Google Postmaster Tools dashboard details). If a dashboard has a gap, mark it as missing data rather than recording zero complaints or zero spam.

The comparison should use matching slices: Gmail campaign complaints against Gmail campaign complaints, and transactional bounces against transactional bounces. Preserve raw message logs as well as summary metrics. Logs help determine whether a change came from DNS, a message template, a recipient segment, or the new delivery route.

Prepare the new provider without changing everything at once

Treat provider setup as a controlled DNS and message-format change. Keep the visible From domain where possible, configure the new provider’s authentication before production traffic moves, and test the message paths. That makes it easier to distinguish a DNS or rendering fault from an issue with volume or audience quality.

Work through SPF, DKIM, and DMARC

SPF is published as a single TXT record for a domain, not as a separate record for every provider. Add the new provider’s authorization to the existing policy rather than creating a second SPF record: RFC 7208 says multiple SPF records can cause a permanent error. Its example syntax is v=spf1 +mx a:colo.example.com/28 -all. This is an RFC example, not a ready-to-publish migration record; the correct mechanisms depend on your authorized sending systems. RFC 7208

During a provider transition, keep the old SPF policy valid until legitimate mail from the old system can reasonably be expected to have been checked, as RFC 7208 advises. Count DNS-query-causing terms before publishing: SPF implementations must limit them to 10 during evaluation, including include, a, and mx. An include can authorize a provider’s hosts, but adding one can bring the included policy’s lookups into the total.

DKIM can use concurrent selectors. A key record is a tag-list: p= is required and contains base64 public-key data; an empty p= value means the key has been revoked. For RSA, that value contains the base64-encoded ASN.1 DER-encoded RSAPublicKey. A schematic record might look like v=DKIM1; k=rsa; p=BASE64_PUBLIC_KEY; replace the placeholder with the actual public key supplied for your domain and selector. Publish the new selector separately, and do not create multiple TXT records for the same selector name. RFC 6376

DMARC alignment compares SPF and DKIM domain identifiers with the RFC5322.From domain. DMARC.org’s deployment guidance describes starting with a p=none policy and reports, analyzing the data, and then moving policy toward quarantine or reject as the mail streams are understood. The updated DMARC protocol deprecated the pct=, rf=, and ri= tags, added np= for non-existent subdomain policy, and brought in psd= from RFC 9091 (updated DMARC specification). Don’t copy obsolete tags into a new record just because they appear in an older template.

Test DNS and message routing

After publishing records, confirm that they resolve publicly and send a test message. Inspect its SPF result, DKIM signature (d= and s=), and DMARC result. A passing SPF result alone does not show that SPF aligns with the visible From domain. Test real message types such as password resets, receipts, and alerts for the correct From and Reply-To headers, working links, template rendering, and expected authentication results.

Choose deliberately whether the new system will use a custom return-path domain or branded tracking domain, if supported. Document the choices because they affect the domains used in authentication or visible links. Send test messages to controlled mailboxes at Gmail and other major providers before moving production traffic. Google Postmaster Tools requires domain verification before it displays information about that domain; adding the domain and verifying it are separate steps (Google’s setup documentation).

Volanea provides guided setup for SPF, DKIM, and DMARC and checks domain authentication automatically. For staging, Volanea’s sk_test_ keys render and log messages but do not deliver them. Use that mode to validate application payloads, templates, and event routing before switching a production key into the send path. For production integration, Volanea offers a REST API and SMTP relay; its AWS Lambda sending guide is useful when a function owns the application’s outbound mail route.

Clean and segment the audience before sending

Recipient data is part of the migration, not just a contact export. A new provider will not automatically have the previous provider’s suppression history. Transfer exclusions and consent records before sending so an address previously suppressed for a hard bounce or complaint does not re-enter the sendable audience simply because it is new to the platform.

Export suppression data separately from active contacts, preserving the reason for each suppression where available. Carry forward at least these exclusions:

  • Hard bounces: exclude addresses recorded as permanently undeliverable. Importing one into a clean contact table does not make it safe to send again.
  • Complaints: retain spam-complaint suppressions across providers. Do not reintroduce a recipient who reported your mail as spam.
  • Unsubscribes: transfer opt-outs and apply them to the appropriate campaign audiences. Keep consent and unsubscribe records alongside contact data.

Keep the suppression file distinct from the sendable audience. Test the migration process with a small sample, including a known suppressed address, and confirm that it is excluded from a real send. If the old provider distinguishes temporary failures from permanent bounces, use the recorded status and your sending policy rather than treating every soft bounce as a permanent suppression.

When using Volanea, suppression records are part of the API surface, so the migration can include them alongside contacts and sending logic rather than leaving exclusions in a separate, undocumented application rule. Import and test those exclusions before sending to the migrated audience. Volanea’s sending endpoint also processes suppression checks, contact upserts, template rendering, tracking, and delivery in one pipeline; keep your own eligibility logic clear so application code does not accidentally bypass the intended audience rules.

Build engagement segments from actual activity. Keep recent clickers or purchasers separate from contacts who have not opened or clicked recently, and distinguish brand-new signups from older imported records. Begin campaign sends with recipients likely to recognize your messages. Hold inactive contacts back until you have a reason to include them; a provider cutover is a poor time to reactivate a stale list and change infrastructure at once.

Plan the cutover and warm up traffic safely

Treat cutover as a transfer of individual message streams, not a single switch for the whole business. Assign each stream—password resets, receipts, alerts, and campaigns—to one active provider at a time. That prevents an application retry or scheduled campaign from sending duplicates through both systems. Keep the old route available for rollback, but disable its sending path for each stream once the new route owns it.

A staged sequence might work like this:

  1. Validate application behavior without delivery. Use test credentials to inspect rendered content, headers, template variables, and event routing. For Volanea, sk_test_ keys render and log but never deliver, making them suitable for this staging pass.
  2. Move a small transactional slice. Route a limited share of password resets or account verification messages to the new provider. Verify delivery, authentication, latency, and application retry behavior before expanding.
  3. Move other transactional traffic. Transfer receipts and alerts as separate cohorts. Keep campaign traffic on its existing route while you check that each transactional stream is stable.
  4. Move campaigns by engagement segment. Start with recently engaged, eligible recipients; expand to additional segments only when the current cohort is performing within the thresholds you set. Pause the old provider’s scheduled send to that same segment before launching the new one.
  5. Retire the old route after checking every sender. Inspect scheduled jobs, workers, retries, and fallback paths. Remove old DNS authorization only after legitimate mail from the old route can reasonably be expected to have been checked.

Here is an illustrative volume schedule for a service that previously sent 10,000 transactional messages per day and a 20,000-recipient weekly campaign. The stages are based on observed results, not a fixed calendar: start with 1,000 transactional messages, then 2,500, 5,000, and the full 10,000; for the campaign, test 2,000 recently engaged recipients, then 5,000, 10,000, and the remaining eligible audience. Hold each stage until its results meet the team’s pre-set ramp criteria. These are planning examples, not universal rates or a prescription for every provider.

Warm-up depends on the IP arrangement. With a shared IP, you generally are not warming an IP dedicated to your account. Focus on the volume and quality of your own traffic, and keep an eye on recipient response. With a dedicated IP, the new address has no established sending history for your mail, so move traffic onto it progressively rather than transferring the full daily volume immediately.

Volanea sends through shared Amazon SES infrastructure and does not offer dedicated IPs. A Volanea cutover therefore centers on staging your own traffic and monitoring the domain’s results, not following a dedicated-IP warm-up schedule. Volanea’s reputation guard slows sending when a list may harm the sender domain, providing another control when a questionable audience or sudden volume change is detected. The guard does not replace segmentation, suppression transfer, or your own stop rules.

Monitor deliverability signals and decide when to ramp

Set ramp and stop rules before the first production send. Compare each new cohort with the baseline using the same recipient providers, message types, and reporting windows. Assess transactional and campaign traffic separately: a password-reset stream and a re-engagement campaign have different recipient expectations and should not share one pass/fail decision.

The following thresholds are an example of an internal operating policy for the fictional baseline above, not universal provider rules. If a transactional cohort’s hard-bounce rate reaches 2%, or a campaign cohort’s complaint rate reaches 0.1%, pause that cohort and investigate before increasing volume. Also pause if either rate doubles its comparable baseline in a small sample, even if the example threshold has not been reached. A team should choose thresholds that fit its own history and recipient mix; the point is to define an action before launch.

SignalExample ramp conditionExample stop action
Hard bouncesRemain near or below the comparable baselinePause a cohort at 2% in this example; inspect address quality and source data
ComplaintsStay near the matching message-type baselinePause a campaign cohort at 0.1% in this example; review targeting and message recognition
Deferrals and delivery timeNo sustained increase versus the prior stageHold volume if the queue grows or delivery slows; inspect provider responses
AuthenticationTest messages show the expected SPF, DKIM, and DMARC resultsPause the affected route if authentication fails
Inbox placement and engagementRemain consistent with comparable cohortsHold the next increase if placement or engagement falls materially

Track these indicators by mailbox provider where possible:

  • Hard and soft bounces: Separate permanent address failures from temporary delivery failures. A hard-bounce increase can point to address quality; a soft-bounce or deferral increase can indicate temporary delivery issues.
  • Complaints: Compare rates as well as counts and keep the cohort size visible. A small count can matter in a small audience.
  • Engagement: Compare opens and clicks by message type and recipient cohort, but treat them as supporting indicators rather than proof of inbox placement.
  • Inbox placement: Use controlled seed tests or placement monitoring to check whether messages reach the inbox or spam folder. SMTP acceptance does not establish inbox placement.
  • Delivery timing: Review provider responses and time to delivery. Growing deferrals or delayed messages can show trouble before complaints appear.

Google defines a bulk sender as one sending about 5,000 messages or more to Gmail accounts in a 24-hour period. For Gmail visibility, verify your sending domain in Google Postmaster Tools. The tool covers personal Gmail accounts, not every recipient domain, and the Spam Rate dashboard reports the percentage of DKIM-authenticated messages recipients manually mark as spam. Dashboard data is typically updated within 24 hours after a sending change, but can take longer; use it as one signal rather than a real-time control panel (dashboard details).

Record each decision: the cohort, volume, DNS state, bounce and complaint results, deferrals, placement observations, and whether the next stage was approved. If a metric crosses its stop threshold, pause that stream and investigate the specific cause before increasing volume. Do not let a successful test message overrule worsening production signals.

Common migration mistakes and how to recover

The most difficult cutovers are often those where a team cannot isolate the cause of a delivery change. Avoid changing the sending domain, IP arrangement, list, and message content together unless each change is necessary. When a signal worsens, pause the affected stream, return to the last stable configuration, and investigate before trying another change.

  • Sending the full volume at once: If bounces, complaints, or deferrals rise, stop the increase and return to the last stage that met your thresholds. Inspect the recipient cohort and provider responses before resuming.
  • Moving an inactive audience with the infrastructure: Exclude hard-bounced and complained-about addresses, and keep inactive campaign contacts out of the initial cohorts. If a campaign segment performs poorly, stop sending to that segment rather than retrying it through the old provider.
  • Breaking authentication during overlap: Do not remove the old SPF authorization as soon as the new sender passes a test. RFC 7208 calls for a transition period so legitimate in-flight mail can still be checked. For DKIM, publish old and new public keys concurrently during the key transition, as described in RFC 6376. If authentication fails, pause the affected traffic and restore the last known-good DNS or signing configuration.
  • Sending from both providers by accident: Check application retries, fallback routes, and campaign schedules. If duplicate sends occur, disable one route for the affected stream before restarting it; do not leave two providers competing to send the same event.
  • Ramping through deteriorating results: Use the pre-set stop conditions. If one is crossed, pause that cohort, review authentication, audience eligibility, provider responses, and message changes, then test the correction with a small cohort.

A practical rollback changes routing, not recipient history. Point each affected message stream back to the old provider, disable the new route for that stream, and confirm a test event sends once. Keep the suppression state and consent records consistent across both systems during rollback so a retry cannot reintroduce a complained-about, unsubscribed, or hard-bounced contact.

For a Volanea cutover, keep the monthly sending allowance in the plan in view as you schedule campaign stages. At the monthly limit, sending pauses; there is no overage billing. That can affect a migration that moves more volume than expected, so account for production tests and campaign traffic in the monthly total rather than treating them as outside the cutover plan. If your application sends over SMTP, Volanea’s SMTP relay uses username volanea and the project’s secret API key as the password. Keep the test key out of that production path because sk_test_ messages do not deliver.

FAQ

Does sender reputation transfer to a new email service provider?

No portable reputation score transfers. Your sending domain, authentication setup, consent records, and sending practices can remain consistent, but a new IP has its own history and the new provider builds sending history from your mail and recipient response. Keep the familiar parts stable and use staged traffic to measure how the new route performs.

Should I warm up a new IP or a new domain during migration?

The warm-up plan depends on the infrastructure. A dedicated IP needs a progressive increase in traffic because it has no established sending history for your mail. With shared IP infrastructure, focus on a measured increase in your own traffic and response; Volanea sends through shared Amazon SES infrastructure and does not offer dedicated IPs. Keep the established domain when it remains appropriate.

Can I test a Volanea cutover without delivering messages?

Yes. Volanea sk_test_ keys render and log messages but never deliver them, so you can validate payloads, templates, and event routing before a production cutover. Once you switch to a production key, send a controlled test to confirm headers, links, and authentication, then begin with a limited stream rather than moving every message type at once.

What happens if Volanea reaches my monthly sending limit during a cutover?

Sending pauses at the monthly limit, and there is no overage billing. Plan campaign stages and production tests within the month’s sending allowance so the cutover does not rely on continued sending after the limit is reached. Check the sending plan and volume you expect to use before assigning a production stream to the new route.

We built Volanea to combine transactional email, campaigns, and automations on one API, with an MCP server for AI coding agents. If you want to stage a provider cutover on that setup, start free with 1,000 emails a month.