Four messages across two to three weeks, spaced roughly day one, day four, day nine and day sixteen. Each message carries something the previous one did not. The sequence stops on any reply, including a no.
Published by emailcampaign.ai, which runs cold sending infrastructure. The volume figures below are the ones we recommend to our own customers.
The shape
| Step | Day | Job | Length |
|---|---|---|---|
| 1 | Day 1 | Why them specifically, and one clear ask | 50 to 125 words |
| 2 | Day 4 | A different angle on the same problem, not a nudge | 40 to 80 words |
| 3 | Day 9 | Proof: a comparable customer, a number, something concrete | 40 to 80 words |
| 4 | Day 16 | A clean close that makes no reply an acceptable answer | 20 to 40 words |
Two things about that table matter more than the days. Each message gets shorter, which signals that you respect the silence. And each has a distinct job, which is what stops the sequence reading as four copies of one request.
What each step should actually contain
Message one. One sentence establishing why you wrote to this person specifically, one or two sentences on the problem, and one ask. If the first sentence could have been sent to anyone, rewrite it before worrying about anything else.
Message two. The most wasted message in most sequences, because it usually says "just bumping this". Use it for a genuinely different angle: a different consequence of the same problem, or the same point aimed at a different priority. Never open with an apology for writing again.
Message three. Evidence. A similar company, a specific result, a short observation that shows you understand their situation. This is where a case is made rather than asserted.
Message four. The close. Say plainly that you will stop, and make silence a valid answer. Something like: "I will leave it there. If this becomes relevant later, the door is open." It is the most replied-to message in many sequences, because it removes pressure rather than adding it.
Spacing, and why the gaps widen
Same-day and next-day follow-ups read as pressure, and pressure from a stranger produces complaints rather than replies. Complaint rate is the metric with teeth: Google publishes the threshold at below 0.1 percent, never reaching 0.3 percent, and that is one complaint in a thousand.
Widening gaps mirror how people actually chase something politely. Three or four days, then five, then a week. It also spreads the sending load, which matters when each mailbox is limited to eight to ten sends a day and your sequence steps are competing with new first-contacts for that budget.
Why longer sequences do not pay
Sequences of eight or ten messages exist because a tool made them easy to build. The response curve does not support them: the large majority of replies a sequence will ever produce arrive by the third or fourth message.
What continues past that point is complaint risk. Someone who has ignored four messages and receives a fifth, sixth and seventh is being given a reason to click report rather than delete. You are trading a diminishing chance of a reply against a rising chance of damage to the sending domain, and the domain is the asset that all future campaigns depend on.
Stopping rules
- Any reply stops it, including "not interested". Continuing after a decline is how a polite no becomes a complaint.
- Any opt-out stops it permanently, across campaigns, with a suppression record so a later list does not reintroduce them.
- An out of office does not count as a reply. Pause and resume after their return date rather than treating it as engagement.
- A hard bounce stops it and suppresses the address. A policy rejection, the 5.7.x family, is about your sending rather than the recipient and should never suppress the person.
Threading
Keep messages one and two in the same thread so the recipient has context. After that, a fresh thread with a new subject often performs better, because a thread they have already skipped twice carries that history into the inbox preview.
Questions, answered straight
- What is a cold email sequence?
- A planned series of messages to the same person, sent over days or weeks, where each message adds something rather than repeating the last. It stops the moment they reply, positively or negatively.
- How many emails should a cold sequence have?
- Three to four is the range where the return justifies the cost. A first message and two or three follow-ups captures most of the response that a sequence will ever produce, and longer sequences mostly generate complaints rather than replies.
- How long should I wait between follow-ups?
- Three to four days between the first and second, then widen: roughly day one, day four, day nine, day sixteen. Same-day or next-day follow-ups read as pressure and are a common source of complaints.
- Should follow-ups be in the same thread?
- Threading the first two is natural and gives the recipient the context. After that a fresh thread with a new subject is often better, because a thread they have already ignored twice carries that history.
- When should a sequence stop?
- Immediately on any reply, including a negative one, and after the final planned message. Continuing after a decline converts a polite no into a complaint, and complaint rate is the metric that costs you the domain.
Published 20 September 2026. Updated 20 September 2026. Written by the team that runs the infrastructure; numbers come from the platform's own provisioning and sending, and from the providers' published documentation at the time of writing.