Cold Email Follow-Ups: How Many, How Far Apart, and What Each One Should Say

By Laurel · Updated Oct 8, 2026

A cold email follow-up is any later message in a sequence, sent to a prospect who hasn't replied. Most vendor data lands between three and seven emails in total, spaced a few days apart, and each follow-up should add something new: proof, a different problem, a useful resource, or one direct question.

Follow-ups carry the risk: every step after the first goes to someone who already ignored you once. This article covers one email sequence, from count and spacing to copy, threading, stopping, and complaint rates. For the message itself, see how to write a cold email and cold email subject lines. The cold email cadences guide covers rep-level cadences across email, LinkedIn, and calls; the counts here are emails only.

Key Facts

• Belkins, an outbound agency, analyzed 7,530,489 emails from its 2025 client campaigns: steps 2 through 6 produced 58.6% of replies, and over 53% of email-sourced meetings came at step 3 or later (vendor data).
• Instantly's benchmark of its own users (1 January to 18 December 2025) found 58% of replies came from step one (vendor data). The two datasets disagree.
• Hunter's report on 31 million emails sent by its users in 2025 found three-message sequences drew a 6.8% reply rate against 3.3% for one message (vendor data).
• Google asks senders to keep user-reported spam below 0.10% and never reach 0.30%.
• The FTC says CAN-SPAM has no business-to-business exception and requires opt-outs honored within 10 business days.

What the Data Says, and Where It Disagrees

Follow-up numbers come from vendors and agencies analyzing their own campaigns. None is a census, and none shows that the follow-up caused the reply.

Four Datasets, Four Sequence Lengths

The advice differs by a factor of three, and the samples aren't comparable. Longer sequences may also go to different lists, so reply rate by sequence length isn't the effect of the extra emails.

Where Instantly and Belkins Disagree on Step One

Instantly says the first email captures 58% of replies. Belkins says it captures 41.4%, with steps 2 through 6 taking 58.6%. Several things can explain both.

• Length. The more steps follow the opener, the smaller its slice.
• Senders. Belkins runs managed outreach for clients and tracks booked appointments. Instantly's figures cover whatever its users send.
• Which steps count. Belkins counts steps 2 through 6. Instantly attributes its other 42% to "follow-ups" in one place and to steps 2 to 4 in another.
• Landing isn't causing. Someone who would have answered email one a week late shows up as a step-two reply.

The datasets agree on three things. A large share of replies arrives after the first email, reply rate per step falls after step one, and in Belkins' data meetings skew later than replies. Belkins' explanation, that prospects still in play at step 3 are the more engaged ones, is a hypothesis.

How Many Follow-Ups to Send

Plan two to four follow-ups, so three to five emails in total. That sits inside Belkins' range, under Instantly's, and above Hunter's. It's a starting point, not a measured optimum.

• Why not stop at one. A meaningful share of replies arrives later: 42% in Instantly's data, 58.6% in Belkins'.
• Why not keep going. Belkins reports volume falling sharply beyond step 5, with step 6 adding little, and suggests changing channel instead of extending the thread.
• When longer is worth it. A short list of high-value accounts can justify more touches, but the extra steps should change channel (see multichannel outreach sequences) or cite a new signal, as in signal-based outbound, not repeat the email.

Add a step only if the previous one produced positive replies that outweigh the unsubscribes and complaints it caused.

How Far Apart to Space Them

Instantly recommends spacing touches 3 to 4 days apart. lemlist suggests at least 2 days between the first two emails, 4-day gaps through the fourth, and 5 or more days after that. Belkins and Hunter report no gaps, and no source tests one gap against another.

Start near three business days, widen as the sequence goes on, and change one thing at a time. A tighter gap doesn't make a stalled prospect reply sooner. It only shortens the time between two unanswered emails.

A Sample Five-Step Schedule

This is a template to adapt, built from the ranges above, not a measured optimum. Day 1 is the first send. Shift any step that lands on a weekend or a local holiday.

Any reply or opt-out pulls the contact out early. Step 5 is the last email, not bait. A silent prospect leaves the sequence, and a later contact needs a new reason, such as a fresh signal, and starts as a new sequence.

What Each Follow-Up Should Say

Why "Just Bumping This" Fails

A bump adds only a reminder that email one went unanswered. Instantly tells senders never to send a follow-up that is just checking in, and Belkins says step 3 should read as a standalone message with a different hook, angle, and proof point. Someone who ignored email one has no new reason to answer email two unless email two supplies one.

Four Angles That Add Something New

• Proof. A documented result from a similar company, shared with permission, or a public fact about how others handle the problem. With no proof yet, skip this angle instead of inventing a number.
• A different pain. The first email picked one problem. If it missed, the same role probably owns another. Name it and ask who owns it.
• A useful resource. A one-page checklist or template, offered without asking for a call.
• A direct question. One question answerable in a word, such as who owns this or whether the timing is wrong.

Example Follow-Ups

These are illustrations with placeholders. Replace every bracket with something true.

Step 2, short, in the same thread:

Hi [First name],

One detail I left out: this usually comes up when [specific trigger, such as a new SDR hire] meets lead routing done by hand. If that isn't your situation, say so and I'll stop.

Is routing something you're looking at this quarter?

Step 3, proof, with a new subject ("How [similar company type] handles routing"):

Hi [First name],

Different angle from my first note. [Similar company type] in [market] had the same setup: [specific condition]. After [specific change], [documented result, shared with their permission].

That's their result, not a promise for [Company]. Want me to send how they set it up?

Reply in the Same Thread or Start a New One?

Gmail keeps messages in one conversation until the subject line changes, then starts a new one. So the choice is really about the subject.

Stay in the thread when the follow-up points back at the earlier email, like the step 2 nudge. The reader sees the earlier note below and doesn't have to reconstruct the context. Instantly says step 2 emails written to feel like replies, not reminders, beat formal follow-ups by about 30%, with no method given for that figure.

Start a new subject when the follow-up has a new angle that stands alone. A proof point under "Re: lead routing" is a mismatch, and the FTC says the subject line must accurately reflect the content of the message.

Two rules hold either way. Send every follow-up from the mailbox that sent the original, because rotating senders breaks the thread and makes a "Re:" false. And never put "Re:" on a message that isn't a reply: Google's sender guidelines say not to start subject lines with Re: or Fwd: unless the message is an actual reply or forward. A follow-up threaded under your own earlier email is a real reply, and a "Re:" typed onto a first email isn't.

No source cited here compares same-thread and new-subject follow-ups head to head, so treat this as reasoning and test it.

When to Stop, and the Breakup Email Done Honestly

Stop Rules

Pull a contact out of the sequence on any of these:

• Any reply, including "no," "wrong person," and "not now." A reply needs a human answer, not step 4, and handling replies and booking meetings covers that.
• An unsubscribe, an opt-out request, or a spam complaint.
• A hard bounce, or the final step going out.

An out-of-office reply should pause the sequence until the person is back, not count as engagement. Check that your sales engagement platform can stop on reply and pause on out-of-office.

The Breakup Email, Done Honestly

A breakup email is the final message, the one that says the sequence is ending. Claims that it's the highest-replying email are common, but Belkins and lemlist report reply rates falling after the first email, and none of the four reports above breaks out a final one. Treat it as a clean ending, not a trick.

Honest means four things.

1. It's true. If it says last email, it is. A "final" note followed by another teaches the reader to ignore the sender.
2. No pressure. No fake deadline, no guilt ("I guess this isn't a priority"), and no "closing your file" unless the contact really is closed in your CRM.
3. Three easy exits. Not now (and when to check back), wrong person (and who), or no (and you'll remove them).
4. It ends. Contact them again later only with a new reason, in a new sequence.

Subject: should I close this out?

Hi [First name],

I've written a few times about [topic] and haven't heard back, so I'll stop after this one.

If the timing is wrong, tell me when to check back. If someone else owns this, point me to them. If it's a no, a one-word reply is plenty and I'll take you off my list.

[Your name, title, company, postal address] Reply "no" and I won't write again.

How Follow-Ups Affect Spam Complaints

Google asks senders to keep user-reported spam below 0.10% and never reach 0.30%, and Yahoo says to stay below 0.3%. Google defines the Postmaster Tools spam rate as the percent of messages delivered to an engaged recipient's inbox and then marked as spam. If 1,000 of your messages reach Gmail inboxes, three reports put the rate at 0.30%.

Follow-ups raise the risk for a structural reason. Step 1 goes to a whole list, but every later step goes only to people who haven't replied, a pool that leans toward the uninterested. That's reasoning, not a measurement. The four reports above publish reply rates by sequence or step, and none breaks out complaints by step. Hunter reports an overall average unsubscribe rate of 1.6%.

Four controls matter most.

• Stop on reply, every time. A contact who writes "no" must never get step 4.
• Honor opt-outs before the next send. Google recommends fulfilling unsubscribe requests within 48 hours and Yahoo says 2 days. The FTC sets 10 business days as the legal limit for US commercial email. A queued step should check suppression when it sends, not only at enrollment. This is general information, not legal advice.
• Give the uninterested a quieter exit. Google requires one-click unsubscribe for marketing messages from senders above 5,000 messages a day to Gmail and says unwanted messages without it are more likely to be reported as spam. Below that volume, a plain opt-out line in every step does the same job.
• Watch complaints by step. Postmaster Tools shows the domain rate, and Google points to its Feedback Loop dashboard for campaign-level spam rates. If late steps draw far more complaints than early ones, the sequence is too long for that segment.

Authentication and warm-up sit upstream of all this, and cold email strategy covers them.

Adjusting for Southeast Asia

No dataset cited here breaks out follow-up results by country, so treat every number as untested in Singapore, Malaysia, the Philippines, Indonesia, and Vietnam until your own sends say otherwise.

In relationship-first, hierarchical business cultures a bare "following up" can read as curt, so keep the formal greeting and soften the ask. Relationship-first vs task-first cultures and the Southeast Asia business culture guide explain the patterns. And count gaps in business days around national holidays such as Tet, Hari Raya, and Lebaran, because a step that lands in a holiday week can meet a pile of out-of-office replies. Split tests by country.

Measuring Whether Follow-Ups Pay

Judge each step by positive replies per 100 delivered plus its unsubscribes and complaints, and the sequence by meetings booked per 100 contacts. Belkins advises measuring each step's cumulative contribution to the reply pool, not its per-step rate alone, since a low-rate step can still add a large share of replies. Allow a few hundred contacts per variant before comparing. Outbound pipeline math treats sequence length as a funnel input.

Whatever sends the email, record the outcome on the contact in a CRM, so a "not now" becomes a dated follow-up and a "no" becomes a suppression. Rework's Sales Ops CRM is one place to keep that record. Inbound leads run on different rules, since they asked to hear from you: see lead follow-up best practices.

Read the Sequence Like the Recipient

Before a sequence goes live, put all the emails in one document and read them top to bottom as the person who ignored the first. If steps two through five could swap places unnoticed, the sequence is one idea repeated, and that is the first thing to fix.