A common planning mistake shows up whenever someone maps out a new sending setup: they divide target daily volume by mailbox count, get a comfortable number, and call it done. 500 emails a day across 7 domains and 21 mailboxes works out to about 24 emails per mailbox — easy, well under any provider cap. Then warmup starts running on all 21 mailboxes at 30 messages a day each, follow-ups stack on top of new sends, and suddenly a few mailboxes are throwing throttling errors at 2pm and silently dropping the rest of the day's queue.
Warmup is not free volume. It's mail. Your provider counts it.
{"h2": "Yes — warmup counts, and here's why"}
Warmup works by sending real messages from your real mailbox through your real provider. That's the entire point: Gmail and Microsoft only build reputation for a sending identity when that identity sends, gets delivered, and gets engagement. There is no side channel and no exemption. A warmup message authenticates with the same SPF and DKIM records, leaves through the same SMTP connection or API, and increments the same outbound counters as a cold email. From the provider's side, nothing distinguishes it — the From address, the IP, the authentication result, and the mailbox are identical.
Any tool that claims warmup doesn't touch your quota is either routing those messages through its own shared relay — which builds reputation for someone else's IP rather than your mailbox — or it's simply not telling you the truth. If you're evaluating warmup that runs outside your own workspace, that's the first thing to test: send warmup for a day, then open the mailbox's sent folder and count. If the messages aren't there, they aren't yours.
So when you plan capacity, treat warmup as a line item in the same budget as campaign sends. It's a real cost, paid in volume, to buy reputation.
{"h2": "You're budgeting against two different ceilings"}
The hard provider cap is the one people quote, and it's rarely the one that bites. Google Workspace allows roughly 2,000 messages per day per user on paid business tiers, with lower caps on trials and some legacy plans, and it counts recipients rather than just messages. Microsoft 365 and Exchange Online allow around 10,000 recipients per day, but enforce a per-minute throttle of about 30 messages and a 500-recipient-per-message cap. Third-party SMTP providers vary wildly, often with burst limits much tighter than the headline daily number suggests.
The soft cap is the one that matters. Nobody sending cold email should be anywhere near 2,000 a day from a single mailbox. The practical safe ceiling for a cold-sending mailbox sits between 30 and 60 total outbound messages a day — and that ceiling includes warmup. A mailbox doing 30 campaign emails plus 30 warmup emails is a 60-message-a-day mailbox from the provider's point of view, from a filtering engine's point of view, and from a reputation point of view.
That's the reframe. The number you should care about is total daily outbound per mailbox, split between warmup and campaign — not campaign volume with warmup bolted on invisibly underneath it. If you've been sizing your infrastructure using only campaign numbers, your real per-mailbox load is probably 50-80% higher than you think.
{"h2": "Warmup replies count too, and people forget them"}
Warmup isn't one-directional. A decent warmup network has your mailbox replying to messages it receives, because reply behavior is one of the strongest engagement signals a provider can observe. Those replies are outbound sends from your mailbox, and they consume quota exactly like anything else.
If your warmup setting says "30 per day," check whether that means 30 outbound messages total or 30 initiated threads. In some tools it's the latter — 30 threads your mailbox starts, plus however many replies it generates to incoming warmup mail, which can push real outbound volume to 45 or 50. On a 21-mailbox setup, a 30-message assumption that's really 50 adds 420 unplanned sends a day across the account. That's most of a second campaign's worth of volume you never budgeted for.
The fix is mechanical: pull the actual sent-folder count for one mailbox over a full week and divide by five. That's your real daily warmup volume. Trust the mailbox, not the dashboard setting.
{"h2": "The per-minute throttle is the sharper edge"}
Daily quotas are generous. Burst limits are not. Exchange Online's roughly 30-messages-per-minute throttle is easy to trip when warmup fires a batch at the same moment your campaign queue releases a block of sends and three follow-up steps come due on the same morning. Gmail is less rigid about per-minute rates but will start deferring connections if a mailbox hammers it.
What you get is a 4.x temporary error, retries, delayed delivery, and in the worst case a chunk of the day's sends pushed into a window you never intended — like 9pm in the recipient's timezone, which is exactly when a cold email looks most like spam. Spreading sends across the working day with randomized gaps solves most of this, but only if the warmup scheduler and the campaign scheduler are aware of each other. If they're separate systems writing to the same mailbox, they aren't, and you get collisions. There's more on this failure mode in our guide to email throttling and rate limits.
{"h2": "How to budget warmup vs. campaign volume per mailbox"}
Set a total daily ceiling per mailbox first, then split it. Warmup takes the larger share early and shrinks as campaign volume grows — but it should never drop to zero while the mailbox is actively cold-sending.
{"table": {"headers": ["Mailbox age", "Total daily ceiling", "Warmup sends", "Campaign sends"], "rows": [["Week 1-2", "10-15", "10-15", "0"], ["Week 3-4", "20-25", "15-18", "5-8"], ["Week 5-6", "30-35", "15", "15-20"], ["Week 7-8", "40-45", "12", "28-33"], ["Mature (3+ months)", "45-50", "10-12", "35-40"]]}}
A reasonable steady state for a mature mailbox is 40-50 total sends a day: 30-40 campaign, 10-15 warmup. If you need more campaign volume than that, add mailboxes rather than raising the per-mailbox ceiling. Ten mailboxes at 35 campaign sends each is dramatically safer than five at 70, even though the account total is identical — the risk concentrates per sending identity, not per account.
We've argued elsewhere that you should keep warmup running after launch, and this budgeting math explains why that advice gets ignored so often: people need the capacity, and warmup is the easiest thing to cut. Cutting it buys you ten extra sends a day and costs you the only reliably positive engagement signal in the mailbox.
{"h2": "Worked example: 500 a day across 21 mailboxes"}
Take the setup from the top of this post. 500 campaign emails a day, 7 domains, 3 mailboxes each. Campaign load works out to 24 emails per mailbox per day. Warmup at steady state adds 12 initiated messages plus roughly 8 replies, so about 20. Total outbound per mailbox: around 44 a day — comfortable, well inside both hard and soft limits. Account-wide that's about 920 messages a day leaving your infrastructure, of which 420 are warmup. Nearly half your sending volume is mailboxes talking to each other, and that's fine, as long as you planned for it.
Now change one assumption. Set warmup to 40 a day because "more warmup is safer." Per-mailbox outbound jumps past 64, you start bursting against Microsoft's per-minute throttle during overlap windows, and 45% of everything your infrastructure sends goes to warmup peers instead of prospects. Reputation doesn't improve proportionally — warmup hits diminishing returns well before 40 a day, and past a certain point you're just adding synthetic traffic that looks nothing like your real sending pattern.
One more thing the 24-per-mailbox figure hides: that's new sends only. Follow-ups are additional. A four-step sequence at full steady state roughly triples daily send volume per mailbox compared to day one of the campaign, because you're sending step one to new leads while steps two through four fire for everyone added in previous weeks. If you're building something at this scale, the full arithmetic is in our 500 cold emails per day setup guide.
{"h2": "What hitting the cap actually looks like"}
You rarely get a clean warning. On Gmail you'll see 550 5.4.5 "Daily user sending limit exceeded," and sending is disabled for up to 24 hours — the mailbox is dead weight for the rest of the day. On Exchange Online it's usually 432 4.3.2 throttling responses, or messages sitting in the queue with escalating retry delays. Inside your sending platform, the tell is a spike in failed or pending sends clustered at a specific time of day, almost always afternoon, after the morning warmup batch and the campaign queue have both drained the quota.
The downstream damage is subtler. Your sequence timing breaks. A step-two follow-up scheduled for day three goes out on day five. Reply rates sag, and the sag looks like a copy problem or a targeting problem, so you rewrite the email instead of fixing the queue. If you ever see a daily quota error, don't start with campaign settings — open the sent folder and count everything that left the mailbox, warmup replies included.
{"h2": "When to cut warmup instead of campaign volume"}
Warmup is maintenance, not insurance, and there are situations where trimming it is the right call. A mailbox that's six months old with consistent campaign volume and clean placement can safely drop to 8-10 warmup sends a day. If you're running a short, high-value campaign to 200 named accounts and need the headroom for follow-ups, pause warmup for that week and restart afterward. And if warmup replies are polluting your unified inbox badly enough that you're missing real prospect replies, reduce volume until the filtering is fixed — a missed reply costs more than a marginal reputation gain.
The opposite case matters more. If a mailbox starts showing placement problems, cut campaign volume and raise warmup, not the reverse. Reputation recovers through sustained positive signal, not through going quiet. A mailbox sitting idle for two weeks comes back looking like a dormant account that suddenly woke up — which is precisely the pattern filters are trained to distrust.
{"h2": "How Sendvanta handles the overlap"}
Sendvanta treats warmup and campaign sends as a single per-mailbox budget rather than two systems competing for the same quota. Warmup runs inside your own Gmail, Microsoft 365, or SMTP mailbox using real labeled messages, so it does count against your provider limits — and the scheduler knows it. You set a total daily ceiling per mailbox, and warmup and campaign volume are allocated inside that number, with sends spread across the day to stay clear of per-minute throttles.
Layered deliverability health scoring watches mailbox, domain, DKIM, and IP signals together, and when risk thresholds are crossed the platform slows or pauses campaign sending first, leaving warmup running — which is the order you want. You can see how the pieces fit on the features page.
{"faq": [{"q": "Do warmup emails count toward Gmail's 2,000-per-day limit?", "a": "Yes. Warmup messages sent from your Google Workspace mailbox count against the same daily recipient quota as campaign emails. In practice the 2,000 figure is irrelevant for cold email — your safe ceiling is 40-60 total sends a day, warmup included."}, {"q": "How much warmup volume should a mature mailbox run?", "a": "About 10-15 messages a day, roughly 25-30% of total outbound. New mailboxes start at 100% warmup and taper over six to eight weeks as campaign volume ramps."}, {"q": "Do warmup replies count against my limit too?", "a": "Yes — any message leaving your mailbox counts, including automatic replies to incoming warmup mail. Check your sent folder for the real number rather than trusting the warmup setting, which often counts threads instead of messages."}, {"q": "Can I pause warmup to free up capacity for a campaign?", "a": "For a short burst on an established mailbox, yes. For new mailboxes or any mailbox showing placement issues, no — cut campaign volume instead and keep warmup running."}, {"q": "What happens if I exceed the daily limit?", "a": "Gmail returns 550 5.4.5 and disables sending for up to 24 hours. Exchange Online returns 432 4.3.2 throttling errors and delays delivery. Either way your sequence timing breaks and remaining sends land at odd hours."}]}
Provider limits cited here reflect published Google Workspace and Microsoft 365 documentation as of February 2026; verify your own plan's caps in the admin console, since they vary by tier, billing history, and account age.
Ready to send outbound that lands?
Create your free Sendvanta workspace — no credit card required.
Start free