When to open a second sending domain for one client vs keep warming the first
Agencies open a second sending domain for two opposite reasons. One is right: the first domain is healthy and you need more capacity without raising per-mailbox volume into a danger zone. The other is panic: the first domain is struggling, so someone buys three more domains and sprays the same list harder.
Warm the first domain before you multiply. Then open a second domain when capacity, redundancy, or isolation needs it, not because the first week felt slow.
What "warm enough" looks like before you add a domain
Common agency practice for a new secondary domain:
- Auth first: SPF, DKIM, and DMARC passing before any ramp
- Age and warm window: often 2 to 4 weeks of gradual volume before steady cold sends
- Ramp shape: start very low (often mid-single digits per day), then step up across weeks rather than jumping to full volume
- Steady-state per mailbox: many cold-outbound teams keep roughly 25 to 50 cold sends per mailbox per day once healthy, and add mailboxes/domains instead of blowing past that
Exact numbers should match the client's ESP, list quality, and history. The decision rule matters more than copying someone else's ceiling: do not add domain two until domain one shows stable bounce, complaint, and placement signals under a real campaign load.
Google's sender guidelines still apply either way. Spam rate pressure (Postmaster Tools guidance commonly cited around keeping spam well under 0.3%, ideally under 0.1%) is a domain-health problem. More domains do not erase a complaint problem if you copy the same bad list onto each one.
Open a second domain when
- Capacity: Domain one is at safe per-mailbox caps and the client's ICP volume still needs more daily send capacity.
- Redundancy: You want a standby domain already warmed so a blacklist or reputation dip does not zero the client's outbound for two weeks.
- Risk isolation: You are splitting a riskier segment (new list source, new geo, new offer) away from the domain that carries the primary campaign.
- Mailbox layout: You already run a sane 2 to 3 mailboxes on domain one and adding more mailboxes there would concentrate too much volume on one reputation bucket.
Each new domain needs its own auth, its own warm plan, and its own monitoring. A parked domain with no warm history is not a spare tire.
Keep warming the first domain when
- Auth or DNS is incomplete or recently changed
- Bounce or complaint rates are elevated and not diagnosed
- Warm-up is still climbing and you have not held a steady cold volume for long enough to trust the trend
- The "need" for more volume is really a content or list problem (low reply quality, high soft bounces, filter hits)
- Client IT has not finished redirects, tracking, or seed-inbox access for domain one
Adding domain two in those states multiplies the mess. You now have two half-warmed reputations and one root cause you still have not fixed.
Decision table for agency ops
| Signal on domain one | Default move |
|---|---|
| Auth green, warm complete, metrics healthy, volume capped | Plan domain two for capacity or standby |
| Auth green, warm incomplete | Keep warming; do not split focus |
| Complaint or policy-block spike | Pause and diagnose; do not open domain two for the same content/list |
| Hard-bounce spike on one segment | Pause that segment; keep domain one if other campaigns are clean |
| Blacklist or domain reputation drop | Contain, remediate; spin standby only if it was already warmed |
How to bring domain two online without contaminating domain one
- Register early enough that age is not zero on launch week.
- Complete SPF, DKIM, DMARC, and a redirect to the client site.
- Warm on its own ramp. Do not "borrow" trust by blasting from domain one and domain two on day one of domain two.
- Split campaigns intentionally. Document which sequences ride which domain.
- Keep Postmaster, seed, and bounce boards per domain, labeled with the client workspace id.
- Retire or pause a burned domain without touching the healthy one.
Staggered warming (one domain per week) is easier to watch. Parallel warming is faster if you have the monitoring bandwidth. Pick based on ops capacity, not optimism.
Client primary domain stays out of the rotation
Whatever you decide about domain two, keep the client's primary business domain off cold outbound. Burn a secondary domain and you buy another. Burn the domain that carries invoices and support and you have turned outbound into an outage.
GTM keeps send windows, domain limits, and client-isolated workspaces so one client's domain choices do not leak into another client's reputation. Your playbook still has to decide, per client, when domain two is capacity and when it is denial.
Warm the first domain until it is boring. Open the second when you need capacity or a real standby, and only after you can explain why domain one is not the problem.
Sources: Reachly email deliverability guide, InboxWarm cold email setup for agencies, Salespipe domain warmup playbook, Octave domain warming guide, Google email sender guidelines, GTM product knowledge.