Daily sending limits that keep you invisible to filters

The working limit for cold email is 25–40 messages per inbox per day — far below the caps the providers publish. Spam filters do not care what your plan technically allows; they care whether your sending pattern looks like a busy human or a small broadcaster. Stay inside the human band, ramp gradually, and scale by adding inboxes rather than raising volume per inbox.

Why is the real limit so far below the official cap?

Google Workspace permits up to 2,000 emails a day per account, and Microsoft 365's caps are similarly generous. Those numbers exist for legitimate internal and transactional mail, not as a safe target for outreach. The filters run a separate calculation entirely: a real employee sends a few dozen external emails a day, receives replies, and mixes new threads with ongoing ones. An account pushing hundreds of similarly shaped messages to strangers, with few replies coming back, deviates from that profile in a way reputation systems are built to notice.

25–40 a day is the band where cold volume hides inside normal human behaviour. Volume discipline is one layer of a larger stack — domains, authentication, warm-up, verification, monitoring — all covered in the Cold Email Deliverability guide. But it is the layer firms break first, because the temptation to "just send more" is constant.

What happens when you exceed the band?

The mechanism is a chain, and every link is measurable. When volume jumps beyond your established pattern, filters flag the deviation and route more of your mail to spam. When more mail lands in spam, engagement falls — nobody replies to what they never see. When engagement falls, your sender reputation falls with it, so the next day's sends score worse even at the old, sensible volume. When reputation drops far enough, recovery is not a matter of easing back for a day or two; rebuilding trust typically takes weeks of low-volume, high-engagement sending.

That is the trap: the penalty for one aggressive week commonly outlasts any pipeline it generated. Note what triggered it — volume behaviour, not vocabulary. Word choice gets the blame far more often than it deserves, as covered in spam trigger words: mostly myth, partly real.

How does 25–40 a day translate into pipeline?

Less directly than most founders assume, because follow-ups consume most of the budget. We run sequences of 4 emails over 14 days, so at steady state roughly three-quarters of each day's sends are follow-ups to earlier contacts. Call it 30 sends a day: that is roughly 8–10 new prospects entered per day, or roughly 150–200 new contacts per inbox per month. At a healthy ~4% positive-reply rate, one inbox produces a handful of warm conversations a month.

That arithmetic is why a single inbox is a pilot, not a pipeline — and it is the sizing logic behind the full system laid out in the UK B2B Outbound Playbook.

How do you scale without raising per-inbox volume?

Horizontally. Add inboxes — typically two or three per secondary domain — and add secondary domains as you go; your primary domain never sends cold email. Each new inbox holds the same 25–40 daily band, so no individual sender ever looks like a broadcaster. Want roughly 300 sends a day? That is eight to ten inboxes across three or four domains, not one inbox at 300.

The structure also contains failure. When one domain's reputation is damaged, you retire it and its two or three inboxes while the rest of the system keeps sending. One overworked inbox on your main domain offers no such firebreak.

How do you ramp a new inbox?

Slowly and boringly. A fresh inbox typically needs several weeks of warm-up — automated low-volume exchanges that generate opens and replies — before it earns a reputation worth protecting; a well-warmed mailbox can score 91/100 on reputation tools. From there, increase cold volume in small steps toward the 25–40 band rather than starting there on day one. Consistency then matters as much as the ceiling: the same rough volume every working day beats bursts and silences, because the pattern itself is the signal.

Which numbers tell you the limit is working?

Watch three weekly: bounce rate (commonly held under 2% on a verified list), spam complaints (as close to zero as makes no difference), and positive replies (~4% is the working expectation; below 3%, fix the campaign rather than the volume). And do not rely on inference alone — inbox placement testing shows you directly where your mail actually lands, before the reply rate tells you the hard way.

---

Next step: the Growth System Audit — £450, seven days, credited against any build — maps where your growth system leaks and what to build first.