Moving from a bundled platform to bring-your-own-provider changes two things at once: where the software lives, and where the sending happens. Doing both simultaneously and abruptly is how migrations go wrong. This guide sequences it so they do not compound.
Before you start: export everything
Do this first, while you still have full access, and store it somewhere durable.
- Contacts, with every custom field.
- Suppressions — unsubscribes, complaints, hard bounces. This is the export most likely to be awkward to obtain and the one whose absence does the most damage.
- Consent records, with dates and sources, if the platform can produce them.
- Templates, as HTML.
- Automation definitions, even if only as screenshots. You will be rebuilding, and the rebuild goes faster with a reference.
- Recent reporting, as a baseline. You need to know what normal looked like before you can tell whether the migration hurt.
Step 1: Choose and set up the provider
Pick the provider account you will own: Amazon SES, SendGrid, Mailgun, Postmark, Brevo, or an SMTP host.
The main considerations are volume pricing, whether you need dedicated IPs, regional presence, and how the provider surfaces bounce and complaint webhooks. For most operators moving from a bundled platform, SES is the cheapest at volume and Postmark is the easiest to operate.
Create the account, request any production access it requires, and get the sending limits raised before you need them. Approval can take days.
Step 2: Set up domains — and consider a subdomain
Decide the sending domain now.
If your existing platform sends from mail.yourbrand.com, you have a choice: reuse it, or start on a new subdomain such as news.yourbrand.com. Reusing it inherits whatever reputation it has, good or bad. A new subdomain starts clean but needs full warmup.
Reuse if your current deliverability is healthy. Start fresh if it is not — you are unlikely to repair a damaged domain faster than you can warm a new one.
Add and verify each domain by purpose in SendNectar: sending, tracking, return_path, and landing_page. Publish SPF, DKIM, and DMARC, and confirm alignment between your from-address and the verified sending domain. Do not proceed until every domain shows verified.
Step 3: Build the structure in SendNectar
Create the legal entity, then the brand, then the sender profile pinned to your verified domain and provider identity.
Define subscription topics before importing, because consent records reference them. Map your old platform’s groups, interests, or tags onto topics deliberately — this is the moment to simplify a category structure that grew organically, and it is much cheaper to do now than later.
Step 4: Import suppressions first
Always first. Before a single contact.
Import unsubscribes, complaints, and hard bounces with the appropriate scope — brand in most cases. If the old platform cannot export suppressions cleanly, reconstruct what you can from bounce reports and unsubscribe logs, and treat anything ambiguous as suppressed. The cost of over-suppressing is a slightly smaller list. The cost of under-suppressing is a complaint spike on cold infrastructure.
Step 5: Import contacts with an explicit consent strategy
Choose deliberately:
preserve_evidence— you have genuine consent records with dates and sources, and they carry across.pending_confirmation— you do not have real evidence, so contacts must confirm before receiving marketing. Safest.admin_attested— you are personally attesting to the consent, recorded with your name on it.none— imported with no consent claim; they exist, you may not market to them yet.
Also set the duplicate strategy — update, skip, or reject — rather than accepting a silent merge.
Map every custom field to a declared type before import. A date column landing as free text will break segments quietly.
Step 6: Rebuild content and automations
Import templates as content, then rebuild the pieces you repeat as reusable blocks. Publish each template to create the immutable version your campaigns will pin to.
Rebuild automations rather than trying to replicate them one-for-one. Migration is the rare moment when you have permission to simplify a flow that accreted steps over three years. Publish each workflow to snapshot a version before enrolling anyone.
Set frequency caps before you enable anything.
Step 7: Warm up and cut over gradually
This is where migrations succeed or fail. Do not switch all traffic on day one.
Week 1. Send only to your most engaged segment — contacts who opened or clicked in the last 30 days — at a few hundred a day. Watch bounce and complaint rates daily.
Week 2. Roughly double volume every two to three days, extending to 90-day engaged contacts. If bounce or complaint rates rise, hold at the current level rather than pushing through.
Weeks 3–4. Continue ramping toward full volume. Begin sending your less-engaged segments last, and consider a re-engagement flow rather than simply resuming.
Throughout. Keep the old platform live and read-only. It is your comparison baseline and your fallback.
Step 8: Verify before you cancel
Before shutting off the old account, confirm:
- Delivery, bounce, and complaint rates are at or better than your pre-migration baseline.
- Domain, sender, and provider reputation are all being scored and look healthy.
- Provider webhooks are flowing — delivered, opened, clicked, bounced, complained.
- Unsubscribe links resolve correctly and suppressions apply.
- The preference centre works end to end from a real email.
- Transactional mail is routed separately from marketing.
- You have exported everything you will ever want from the old platform.
Then cancel. Not before.
What usually goes wrong
Suppressions imported second. The single most damaging mistake, and the most common.
Volume too fast. Cold infrastructure at full volume gets filtered, and filtering is much harder to recover from than to avoid.
Alignment missed. SPF and DKIM pass, DMARC fails, and nobody notices because the mail is technically authenticated.
Consent laundered on import. A blanket subscribed flag converts an unknown into a false claim, and it is not reversible once the provenance is gone.
Old platform cancelled too early. The problem you did not anticipate surfaces in week three, and by then you cannot compare against the source of truth.