The short version
- Deliverability moved upstream. Domains can get flagged at registration, before you send a single email, so how you buy and structure infrastructure now matters as much as how you send.
- Keep it small and spread out: about three mailboxes per domain, roughly 30 sends per mailbox per day, and a wide fleet that looks like many real businesses instead of one bulk batch.
- Warm up thoroughly and never stop. Warm-up prevents burn but cannot reverse it, so run it continuously alongside campaigns.
- Retire mailboxes at the first sign of trouble and run a rotation of domains, provisioning roughly three times your peak so a warmed batch is always ready.
- Isolate cold outbound on its own subdomains and dedicated infrastructure so a bad week never touches your root domain, transactional email, or product email.
- GTM Engine runs this playbook for you, with domain and mailbox health, warmups, sending limits, and rotation built into the platform.
3
Mailboxes per domain, max
30
Sends per mailbox per day
3x
Infrastructure vs. peak volume
25%
Reply-rate drop that triggers rotation
What changed
Domains get flagged before the first send
For years the mental model was simple: send badly and your domain gets blacklisted. That is no longer the whole picture. Blocklists and reputation systems have pushed detection upstream, toward how a domain was registered rather than only how it sends. A domain can now pick up a listing during setup, before a single campaign runs.
The reason is uncomfortable but useful to understand: naive outbound looks identical to an attack fleet. Attackers register domains in bulk, same-day batches, from one cheap registrar, using lookalike names. When your fleet clusters the same way by accident, the model has no way to tell you apart. The fix is to stop looking uniform.
The old failure mode
Warm up too fast, send spammy copy, hit spam traps, then land on a blocklist. Reactive, and mostly about sending behavior.
The 2026 failure mode
Buy a batch of cheap lookalike domains from one registrar on one afternoon and get pattern-matched to attacker infrastructure before you send anything at all.
The takeaway: how you buy, warm, and rotate infrastructure is now a deliverability decision, not a procurement detail. The rest of this guide is the operating discipline that keeps a fleet healthy, and how GTM Engine runs it for you.
The operating rules
Five rules that keep a fleet healthy
These are the disciplines that separate teams who reach the inbox from teams who guess. None of them are complicated, but all of them are unforgiving. Here they are in order, with why each one matters.
Cap mailboxes at three per domain
Domain reputation is shared across every mailbox on it. One mailbox in spam means the whole domain is compromised. Keeping three mailboxes per domain limits the blast radius, keeps per-mailbox volume low, and gives you a wide, diversified fleet instead of a few high-risk sending hubs.
- Reputation is scored at the domain level, not the mailbox level
- Fewer mailboxes per domain means a smaller blast radius when one burns
- A wider fleet spreads sending pressure and looks like real businesses
Keep sends to about 30 a day per mailbox
Volume is the fastest way to burn a healthy mailbox. Low, human daily limits keep each mailbox under the thresholds that trigger filtering, and they let you scale horizontally by adding more mailboxes rather than pushing any single one harder. Ramp new mailboxes from a trickle up to their steady daily cap.
- Low daily volume keeps mailboxes under provider filtering thresholds
- Scale by adding mailboxes, never by pushing one mailbox harder
- New mailboxes ramp from 10 to 20 per day before reaching their cap
Warm up thoroughly, then never stop
Warm-up is preventative, not curative. It cannot rescue a burned mailbox, but it builds the reputation new mailboxes need and keeps healthy ones alive. Run pure warm-up for the first two weeks, add light volume, then maintain a warm-up signal alongside campaigns for the entire life of the mailbox.
- Days 1 to 14: pure warm-up, zero campaign volume
- Days 15 to 21: light sending alongside continued warm-up
- Day 22 and on: full volume with warm-up running continuously
Kill mailboxes the moment health drops
A burned mailbox is a zombie, and zombies infect the healthy mailboxes sending next to them. When health falls or a mailbox lands in spam on a placement test, retire it and every mailbox on its domain. Do not try to nurse it back. Cancel the subscription first, then remove it from sending, so you stop paying for dead infrastructure.
- If one mailbox on a domain spams, treat the whole domain as burned
- Retire at the first health drop, not after reply rates collapse
- Cancel billing before deleting so you stop paying for zombie mailboxes
Run a rotation of domains
Provision roughly three times the infrastructure you need at peak, and keep it moving through active, warming, and cooling pools. Rotate domains out while they are still healthy, not after they burn. This preserves each asset for a future cycle and means you always have a warmed batch ready to take over.
- One pool sending, one warming up, one cooling or replacing
- Rotate on a cadence while mailboxes are still healthy
- A fresh warmed batch is always ready to take over
None of this requires a heroic effort, but it does require consistency. GTM Engine enforces these limits and cadences across your whole fleet, so you connect domains and mailboxes and the platform handles the daily discipline: warmup ramps, send caps, health checks, and rotation.
Enterprise architecture
Isolate cold email from the domain your business runs on
Most teams stop sending cold from their root domain the first time a campaign tanks their transactional deliverability. The usual reaction, buying lookalike domains, solves the wrong problem: it dilutes your brand, reads as spoofing, and starts you from zero reputation with Microsoft.
The cleaner architecture is subdomains of your root, on dedicated infrastructure. They inherit your domain's age and trust, keep cold reputation isolated from transactional and product email, and scale horizontally without brand dilution. If a subdomain gets hot, the blast radius stops at the subdomain wall, and your CFO's wire confirmations still land.
Root domain
company.com
Years of age and authority. The trust signal Microsoft cares about.
Transactional
billing.company.com
Receipts, password resets, and alerts stay on your existing ESP. Untouched by cold volume.
Cold Outbound
send.company.com
Dedicated, pre-warmed sending on isolated subdomains. If a lane gets hot, the blast radius stops here.
Product Email
app.company.com
In-app notifications keep their own reputation and their own rails.
Three separate reputations, all under your brand, none able to take the others down.
This matters most for enterprise targets. Microsoft and gateways like Proofpoint weigh domain age and IP reputation above almost everything else, and treat anything brand new as guilty until proven otherwise. An aged subdomain on a dedicated, reputable IP, warmed for 30 or more days, is what consistently reaches those inboxes.
The rest of the playbook
Buy clean, and write to survive filters
Two more things decide whether a healthy fleet stays healthy: the domains you buy, and the copy you send. Diversity has to be real rather than cosmetic, and copy fatigue quietly kills mailboxes that no infrastructure can save.
Buying domains
Spread across registrars
Attacker fleets cluster on one cheap registrar. Split purchases across several reputable registrars so your fleet does not pattern-match to abuse. Cap around 250 domains per registrar account to survive a single-account ban.
Space out purchases
A 100-domain same-day batch is the exact signature reputation systems hunt for. Buy in small amounts, staggered over days and weeks, so your domains never look like one bulk order.
Name like a real business
No hyphens, no numbers, no lookalikes or brand tokens. Use plain, pronounceable, real-word names on .com, .co, .net, or .info. Avoid cheap gTLDs and fingerprinted extensions.
Give every domain a reason to exist
Do not 301-redirect a sending domain to your site. Use domain masking or a simple landing page so each domain reads as a legitimate business, not a throwaway bounce-through.
Sending copy
Vary the shape, not just the words
Providers fingerprint the structure of an email, not its wording. Spintax on greetings is not enough. Run several structurally different variants and rotate them before any one gets stale.
One sender name per mailbox
Reusing a handful of names across hundreds of mailboxes groups them instantly. A unique sender identity per mailbox buys weeks of extra lifespan.
Strip the signature
Images get blocked and the signature is the one constant across personalized sends, so it becomes your fingerprint. Plain text only: name, title, company as a word, no linked .com.
Drop open and click tracking
Tracking pixels and click redirects are detected and downgraded, and the open data has been broken since Mail Privacy Protection. Optimize on replies, the only signal that means the inbox.
Copy fatigue is the number one killer. Providers fingerprint the structure of an email, so the same template sent thousands of times gets filtered no matter how healthy the mailbox. When you rotate infrastructure, ship structurally new copy with it, or the new mailboxes inherit the old fingerprint within days.
Where GTM Engine fits
Managed infrastructure, not a spreadsheet and a prayer
You can run all of this by hand. Most teams do, at least until a spreadsheet of domains, mailboxes, warmup dates, and health scores becomes its own full-time job. GTM Engine turns the discipline above into a managed part of the platform, so your team focuses on the message while the infrastructure stays healthy.
One fleet, centrally managed.
Against a spreadsheet and five registrar logins
A diversified fleet by hand means juggling logins across registrars and tracking domains, mailboxes, and health in a spreadsheet that is out of date the moment you save it.
With GTM Engine
GTM Engine connects domains and mailboxes into one workspace with live warmup status, daily limits, sending windows, and health scores in a single view. Diversity is the default, not a chore.
Reputation that is actually yours.
Against shared-pool mailbox resellers
Cheap resellers put you on shared IPs where one bad neighbor tanks your reply rates. The mystery drops most operators blame on themselves are usually contamination they did not cause.
With GTM Engine
GTM Engine runs cold outbound on isolated subdomains and dedicated sending infrastructure, so your reputation is built and owned by you, walled off from transactional and product email.
Built for cold, not receipts.
Against transactional ESPs pushed into cold
SendGrid, Mailgun, and SES are shared pools tuned for password resets. Push cold volume through them and you trigger abuse teams and risk the rails your product runs on.
With GTM Engine
GTM Engine keeps cold outbound on its own pre-warmed infrastructure and lane, so a bad campaign week never touches the CFO wire confirmations or the app notifications your business depends on.
FAQ
Common questions, direct answers
The rules are not complicated, but they are unforgiving. Here is how GTM Engine handles the questions operators ask before they trust their outbound to a platform. The through-line to all of them: look like hundreds of real businesses, not one spam fleet.
Getting startedHow many mailboxes and domains do I actually need?+
Work backward from volume. At roughly 30 sends per mailbox per day and three mailboxes per domain, one domain gives you about 90 sends a day. Then buy three times the infrastructure you need at peak so one pool is always sending, one is warming, and one is cooling. GTM Engine sizes and manages this for you as your target volume changes.
Size for peak, then triple it for rotation.
Warm-upCan I just warm up a burned mailbox and keep using it?+
No. Warm-up is preventative, not curative. Once a mailbox is contaminated, no warm-up brings it back, and it will infect healthy mailboxes on the same domain. The right move is to retire the mailbox and its domain, and let a pre-warmed batch from your rotation take over. GTM Engine flags the decline early and swaps in ready infrastructure.
Warm-up prevents burn. It cannot reverse it.
VolumeThirty sends a day feels too slow. Can I push more?+
You scale by adding mailboxes, not by pushing any single one harder. Volume is the fastest way to trip provider filters and burn an asset you spent weeks warming. Keeping per-mailbox sends low and human is exactly what keeps a wide fleet healthy and compounding, and GTM Engine makes running a larger fleet as simple as running a few mailboxes.
Scale out with more mailboxes, not up per mailbox.
Enterprise targetsMy prospects are behind Microsoft and Proofpoint. Does any of this land?+
For the hard inboxes, the domain and the IP decide everything. Microsoft weighs domain age and IP reputation above almost anything else, and treats brand-new lookalike domains as guilty until proven otherwise. Sending from an aged subdomain on dedicated, reputable infrastructure, warmed for 30 or more days, is what consistently reaches enterprise inboxes. Gateways like Proofpoint cannot be warmed and need a whitelist or a route around them.
Aged subdomains plus dedicated IPs win the enterprise inbox.
Brand safetyWill cold email put my main domain at risk?+
Not if you isolate it. The failure mode is sending cold volume from your root domain or from lookalikes that read as spoofing. GTM Engine keeps cold outbound on isolated subdomains and dedicated infrastructure, so transactional email, product email, and your root domain each keep their own reputation. A bad week on cold stops at the subdomain wall.
Three lanes, one brand, reputations that never touch.
MonitoringHow do I know when something is going wrong?+
Reply-rate degradation is the earliest warning. A drop of more than about 25 percent week over week with the same audience and copy means rotate now, before the mailboxes are gone. GTM Engine tracks health, warmup progress, and reply performance per mailbox and domain, and surfaces the underperformers automatically instead of leaving you to reconstruct it from a spreadsheet.
Move at the first signal, not after the collapse.
CopyIf my infrastructure is clean, why do results still dip?+
Copy fatigue is the number one killer. Providers fingerprint the structure of an email, and the same template sent thousands of times gets filtered no matter how healthy the mailbox. When you rotate infrastructure, ship structurally new copy with it, or the new mailboxes inherit the old fingerprint within days.
New mailboxes plus burned copy just makes new zombies.
Send from infrastructure built to inbox
GTM Engine runs the modern playbook so your outbound lands: capped mailboxes, disciplined warmups, health-based retirement, and a rotation of domains, all managed for you.