How Many Domains and Inboxes Do You Need for Cold Email? (The Real Math)
"How much infrastructure do I need?" has an actual formula, not a vibe. You work backwards from daily volume — and you always build headroom, because domains get burned and rotated. This is the same math I use to spec fleets sending 100k+ per day.
The formula
Domains needed = inboxes ÷ 2–3 per domain — then add 20–30% headroom for rotation and warmup pipeline.
Why per-inbox volume stays low
A Google Workspace account can technically relay thousands of messages a day — but cold email isn't judged on technical limits, it's judged on engagement signals from strangers. At 15–30/day per inbox, each address looks like a normal professional. At 80/day it looks like a machine, and machines get filtered. The provider-by-provider details are in cold email sending limits.
Why 2–3 inboxes per domain
The domain is the reputation unit. Stack 10 inboxes on one domain and a single reputation dip takes out 10 inboxes of capacity at once. Spread the same 10 inboxes over 4 domains and a hit costs you a quarter of that — and you can rotate the burned domain out without touching the rest.
Build in rotation headroom
Fleets are living systems. At any moment some domains are warming, some are prime, some are cooling down. If your math says 100 domains, buy and warm ~125 — the spare capacity is what lets you swap instantly instead of pausing campaigns for three weeks while replacements warm. Naming and DNS for large fleets: domain setup at scale.
Mixed ESP fleets
At larger scale, split infrastructure across Google Workspace and Microsoft 365 — partly for provider-diversity resilience, and partly because matching sender ESP to recipient ESP measurably helps placement. A common split I run: Google-heavy for Gmail-dominated TAMs (startups, agencies), Microsoft pools for enterprise-heavy lists.
Three worked builds, with cost
The formula is only useful once you have run it at your own scale, so here are three fleets I would spec today, at 25 cold sends per inbox per day, 3 inboxes per domain, 25% headroom, and a 3-step sequence — which means daily sends include follow-ups, and new conversations started per day are roughly a third of the send number.
The last row is the one that breaks plans. A 2,000-a-day fleet needs roughly 15,000 fresh, verified prospects a month, every month. If the list side cannot supply that, the fleet is oversized and the money should go to data instead. Seat cost at each size, on Google, Microsoft or mixed, is in the inbox cost calculator; the list decay calculator shows how fast that monthly list need grows once bounces from stale data are counted.
Building the pool in waves
Do not build the whole fleet on day one. Domains registered together, mailboxes created together and campaigns started together are a fleet-shaped signal, and one bad decision — a wrong SPF include, an unverified list — hits every domain at once. I build in waves of 10–15 domains, two to three weeks apart. Wave one is sending at cruise while wave two warms; wave three registers while wave two starts campaigns. The fleet reaches full capacity in six to eight weeks, but every wave is a test of the previous one, and a problem in wave one is fixed before wave two inherits it.
Waves also give you the rotation pipeline for free. If a domain in wave one burns in month four, wave four is already warming, and its replacement slots in without a gap. This is what the 20–30% headroom actually buys: not idle inboxes, but a warmup pipeline that is always one wave ahead of demand. The domain and inbox planner lays the waves out on a calendar against a target date.
The three numbers that change the answer
Every fleet size above assumes 25 sends per inbox. Move that to 15 and the fleet is two-thirds larger; move it to 40 and the fleet is smaller and, on most lists, shorter-lived. The second number is inboxes per domain — 2 is the conservative choice for high-value lists where a burned domain costs a lot of warmed reputation, 3 is the default, and 4 or more concentrates risk for a saving in domains that is usually not worth it. The third is the sequence length, because a 5-step cadence at the same daily send count starts 40% fewer new conversations than a 3-step one.
Run the formula with your own three numbers rather than mine. The capacity calculator takes all three as inputs; the sequence planner shows what a given cadence does to new-prospect throughput; and the ROI calculator is where the fleet cost, list cost and expected reply rate finally meet, which is the number that decides whether the fleet is the right size or the right idea.
Take the domain count from the formula above and buy the batch together so the fleet ages together. Namecheap does bulk registration with free WHOIS privacy, which keeps the WHOIS data consistent across all of them — one fewer fingerprint shared by fifty names.
Bulk-register on Namecheap →Affiliate link — it costs you nothing and helps keep this site free.Do the maths with your own numbers
The domain & inbox planner turns a daily target into inboxes, domains, cost and a purchase-to-launch calendar at the ratios above. The sequence planner then shows what a multi-step cadence does to that daily volume once follow-ups stack.
Tools for this
Inbox cost comparisonCompare Workspace, Microsoft 365 and SMTP.List decay calculatorHow fast a list dies, and what that costs.Domain & inbox plannerVolume target to domains, inboxes and cost.Capacity calculatorSize a sending fleet against a target.Frequently asked questions
How many inboxes should I put on one cold email domain?
Two to three inboxes per domain is the standard I run at fleet scale. More inboxes on one domain concentrates risk: if the domain's reputation dips, every inbox on it dips together, and recovery is slower.
How many cold emails can one inbox safely send per day?
After full warmup, 15–30 cold emails per day per inbox is the sustainable range I use. Providers technically allow far more, but cold email is judged by recipient engagement, and pushing 50+ per inbox reliably degrades placement within weeks.
Is it better to add more domains or more inboxes to scale?
More domains. Domains are the reputation unit — spreading volume across more domains with 2–3 inboxes each keeps any single reputation hit small and lets you rotate burned domains out without denting total capacity.
How many domains does an agency need per client?
Enough that a client's domains are separate from every other client's — never share sending domains across clients, because one client's list quality becomes everyone's deliverability. Beyond that, the formula is per client: their daily target, divided by the safe per-inbox rate, divided by inboxes per domain, plus headroom. A client sending 500 a day is 9 domains; most agencies run that as a dedicated pool under the client's name.
Can I run more than 3 inboxes on a domain if they send less each?
The domain's reputation is the sum of its inboxes' behaviour, so five inboxes at 15 sends carry the same domain-level volume as three at 25 — but they concentrate more accounts on one domain's fate, and receivers do see the number of active mailboxes on a domain. Below 3 there is little benefit; above 3 the saving in domains is small and the blast radius when the domain dips is bigger. Three is the default for a reason.
How much should I budget per month for a cold email fleet?
Roughly: inboxes × seat price, plus domains × annual price ÷ 12, plus the sending tool, plus data. For a 2,000-a-day fleet that is about 100 seats, 34 domains and 15,000 verified prospects a month — the seats and domains are the predictable part, the data is usually the larger one. The inbox cost calculator prices the infrastructure side by provider; the data side depends on where you source.