Cold email warmup takes about two weeks before you send a single real email, then three to four more weeks to ramp to full volume. Call it six weeks from buying the domain to sending at your target rate. During the first two weeks you send nothing but automated warmup traffic. After that you start at 5–10 emails per mailbox per day and roughly double each week until you hit 25–30.
That’s the schedule. The more useful part of this guide is what the schedule is actually for, which parts genuinely matter, and the limit that stopped us before deliverability ever did, because it wasn’t the one we’d planned around.
The schedule
| Period | Warmup emails | Real cold emails | What you’re doing |
|---|---|---|---|
| Days 1–14 | 5–10/mailbox/day, automated | Zero | Building a send history from nothing |
| Week 3 | Continue | 5–10/mailbox/day | First real sends, small and slow |
| Week 4 | Continue | 10–20/mailbox/day | Watching bounce and spam rates |
| Week 5+ | Keep running | 25–30/mailbox/day | Steady state |
Two things people get wrong here.
These numbers are per mailbox, not per domain. Four mailboxes on a domain sending 25/day each is 100 emails/day, and each mailbox follows its own ramp. This is why outreach setups use several mailboxes instead of pushing one inbox harder. You scale sideways, not upward.
Warmup never stops. It’s not a two-week setup task you finish. The warmup traffic keeps running underneath your real campaigns permanently. Turn it off after a good month and your ratio of “cold strangers” to “people who reply” collapses, which is the exact signal that got you filtered in the first place.
What warmup is actually doing
Warmup tools send mail between real inboxes in a shared network. Those inboxes open your message, mark it “not spam”, and reply. You’re manufacturing a history of positive engagement so that when your first real cold email arrives, the receiving server has seen your domain behave like a domain people want to hear from.
That’s it. It’s a reputation prior, nothing more. It does not make bad copy land, it does not fix a domain that already got burned, and it does not buy you exemption from the authentication rules below.
Google’s own sender guidelines are unglamorous and say the same thing in one line: “Start with a low sending volume to engaged users, and slowly increase the volume over time,” and avoid sudden volume spikes if you have no history of sending large volumes.
Authentication is the part that isn’t optional
Warmup gets attention because it’s the interesting part. Authentication is boring and matters more.
Google requires all senders to have SPF or DKIM set up, valid PTR records, and TLS. If you send 5,000 or more messages per day to Gmail addresses you’re classified as a bulk sender, and then you need SPF and DKIM and DMARC, plus one-click unsubscribe headers.
Here’s the nuance nobody writes down: at cold outreach volumes you are almost certainly not a bulk sender. Four mailboxes at 30/day is 120 emails, and that’s 2% of the threshold. Most warmup articles quote the bulk-sender rules at small senders as though they’re all mandatory. They’re not.
Set up all three anyway. DMARC costs an afternoon, and the one rule that does apply to you regardless of volume is the spam rate: keep reported spam complaints below 0.30%, with Google recommending you stay under 0.10%. Cross 0.30% and your domain stops being eligible for Gmail’s delivery mitigation until you’ve held it below that line for seven straight days. At 120 emails/day, 0.30% is less than one complaint. The margin is thin enough that “we’re too small for the rules” is a bad bet.
One-click unsubscribe isn’t required below the bulk threshold either. Include a plain unsubscribe line regardless. A recipient who can’t find one clicks the spam button instead, and that button is the only number that can actually kill you.
The limit that stopped us wasn’t deliverability
We planned our own outreach ramp at 60 emails/day in week one, 90 in week two, 120 at steady state (four mailboxes, ceiling of 30 each), comfortably inside every deliverability guideline above.
We never sent 60. The sending platform’s plan allowed 1,000 contacts per month, which works out to about 46 emails a day. The billing ceiling sat below the deliverability ceiling, and it was the one that bound. We cut week one to 45/day, not because the domain couldn’t take more, but because the plan couldn’t.
This is a genuinely common trap and it’s invisible until you’re mid-launch. Outreach tools price on contacts, not emails. A three-step sequence to one person is one contact and three emails. So the contact allowance divided by your sequence length, not the email allowance and not your mailbox capacity, is your real daily number.
Work it out before you buy domains. We now treat the plan limit as a hard input to the ramp: the sending schedule refuses to advance to a volume the subscription can’t cover, no matter how clean the deliverability week looked. Every campaign day is planned against both ceilings.
While we’re on unglamorous mistakes: check what your tool’s step delays actually mean. On some platforms the delay belongs to the step before the gap, so a “0 day” delay on step one fires step two immediately. We shipped that misconfiguration once. Nothing about your warmup saves you from sending someone two emails in the same minute.
What matters less than people claim
Buying eight lookalike domains. Useful at real scale. At one or two campaigns it’s mostly cost and admin. Start with one alternate sending domain, never your main company domain, so a burned reputation can’t take your actual email with it.
Chasing a perfect warmup-tool score. The score measures the warmup network, not your campaign. A 100% score with a 4% bounce rate on real sends means your list is bad and the score is telling you nothing.
The exact ramp curve. The difference between doubling weekly and adding 30% weekly is noise. The difference between ramping and not ramping is everything.
When you don’t need any of this
If you’re sending fewer than about 20 emails a day, to people who know who you are, from your normal company inbox, skip the whole apparatus. No warmup tool, no separate domain, no ramp. Send the emails. Warmup infrastructure exists to solve a volume problem, and at that volume you don’t have one. Buying a warmup subscription to send 15 emails a day is a tax on a problem you don’t have.
The apparatus starts earning its cost somewhere around 50 emails a day to strangers, sustained.
FAQ
How long before I can send my first cold email? Two weeks of warmup with zero real sends, then start at 5–10 per mailbox per day. Full volume around week five or six.
Can I speed it up? Not meaningfully. The thing you’re building is a history, and history takes time by definition. Buying an aged domain skips some of it, but you inherit whatever reputation it already carries, sometimes worse than starting clean.
Do I need a separate domain for cold email? Yes, if you’re sending real volume. Use an alternate domain, not your primary. If it gets burned you throw it away instead of losing the address your clients and invoices run through.
My emails still land in spam after warmup. What now? Check in this order: authentication actually passing (not just configured), bounce rate on your list, then copy. Spam placement after a clean warmup is usually a list-quality problem wearing a deliverability costume.
Warmup gets the email delivered. What happens after someone replies is the other half, and it is where most of the money leaks: see what CRM setup costs in 2026, or what we build so no enquiry goes unanswered.