How we write cold email sequences

Most advice on cold email sequences is a list of rules: send this many emails, keep them under this many words, use this kind of subject line. Some of it is right. Very little of it tells you why, so the first time your situation doesn’t fit the rule, you’re guessing.

This is how we write the sequences we send for our clients and for ourselves, with the reasoning behind each choice. They’re defaults that work for us, not laws. Where your situation is different, you’ll know enough to break them on purpose.

Start with the list, not the copy

We don’t write a word until the list is built and every address on it is verified. That order sounds like admin, but it decides how good the copy can be.

The list tells you who you’re writing to: their role, their kind of company, the problem they’re likely to have this quarter. Copy written before the list exists is written to an imagined reader, and it reads that way. And any time spent on an email that goes to an address that bounces is wasted, along with a little of your domain’s reputation. How we keep sending healthy is its own subject; the short version is to verify first and write second.

Get the offer into one sentence

Before the first email, we write the offer as a single sentence with no jargon in it: what the reader gets, how soon, how much effort it takes on their side, and why it’s safer than what they’re doing now.

If that sentence won’t come, the sequence isn’t ready, and no amount of copywriting will fix it. A clear offer in plain words beats a clever email about a vague one every time. It also gives you something to check each email against: if a line doesn’t serve that sentence, it goes.

One email, one idea

Every email we write picks one focus. Usually that’s either the outcome (what they get: booked calls, a shorter close, a clean forecast) or the mechanism (how they get it: the process, the tool, the team). Selling both in one email muddies both, because the reader has to decide which part matters to them, and most won’t bother.

The same goes for the ask. One email, one question. Two questions give them two things to think about, and the easiest response to two things is neither.

Write to what the reader is living, not to the signal that found them

A lot of outbound is built on signals: a company raised money, posted a job, hired a new head of sales. Signals are a good way to decide who goes on the list and when. We almost never mention them in the copy.

There are two reasons. Between building a list and the first send, two weeks or more can pass while new mailboxes warm up, and “congrats on the raise” reads very differently a month after the raise. And “I noticed you’re hiring” tells the reader they’re being watched, which is not the first impression you want.

So the signal picks the list, and the email is written to the consequence: what the reader is dealing with because of it. A company that just raised is about to hire, spend, and be asked for results. Write to that, and the email still makes sense whenever it lands.

Three steps and what each one is for

Our default sequence is three emails over a week, on Day 1, Day 3, and Day 7. Many guides recommend four to seven emails over two to four weeks, often the ones written by tools that are paid by the email. Practitioners writing about it on Reddit and X tend to land closer to ours, at two to four emails.

Those days are the settings in the sequence editor, not a promise about the calendar. We don’t send on weekends, and mailbox limits, campaign settings, and a long queue can all stretch the gaps, so the third email often lands later than Day 7.

Each step has its own job:

• Step 1 starts a conversation on one idea and asks one question about interest. We usually put that question in the first line: whether the problem is something they’re dealing with, not whether they have 15 minutes. Two questions in a short email leave the reader unsure which to answer. Asking a stranger for their calendar in the first email is too much too soon. The meeting comes after the yes.

• Step 2 brings something step 1 didn’t have. Proof is our default: a result someone like them got. A different angle on the same problem, something useful you offer to send, or a smaller question all work too. What doesn’t work is step 1 again in different words.

• Step 3 routes or closes, and it’s the shortest. Most people who don’t reply aren’t saying no; the email reached the wrong person or the wrong week. “Is this yours, or should I ask someone else?” is the question we end almost every sequence with.

Longer sequences have their place. A small market where every account matters can justify more touches over more weeks. But that’s a choice you make for a reason, not a default.

For a full sequence with every line explained, see How to write a cold email sequence, annotated line by line.

Subject lines: short, lowercase, one thread

Our subject lines are one to three words, lowercase, in the reader’s own words for the topic: pipeline, renewals, month-end. Most guides agree on short and lowercase, and we do too.

Where we differ is on two points. We don’t personalize the subject, because a subject with their company name in it looks like a mail merge before they’ve opened it. And we use the same subject on every step, so the whole sequence sits in one thread in their inbox and reads as one conversation. That’s also why we never fake a “Re:” on a follow-up: the real thread is already there.

Follow-ups that read like a person wrote them

Everyone agrees a follow-up should add value and never say “just checking in”. The harder part is the line-level detail that makes a follow-up read as written rather than sent. These are the edits we make on almost every draft.

The greeting tightens as the sequence goes on. Step 1 has its greeting on its own line. Step 2 starts Hi again {{firstName}}, and carries straight on into the body on the same line. Step 3 often drops the greeting word: {{firstName}}, is this yours.... It’s how a real thread gets shorter as two people get used to each other.

We delete throat-clearing lines. “Following up on my last email.” “Quick follow-up.” “One last note from me.” They take the most valuable line in the email, the first one, and carry no information. Hi again already says it’s a follow-up.

We delete labelling phrases. “Here’s how it works:” and “The part that matters is” announce something instead of saying it. Say the thing.

We cut setup-and-contrast. “Most agencies hand you a report at the end. We hand you something else:” makes the reader wait through a straw man to reach the point. Give them the point.

We cut the summarising payoff. If the email has just shown the benefit, a closing line explaining it (“which means more time for what matters”) only makes it weaker.

Two smaller habits sit alongside these. Contractions are fine, because people use them. And we never put “quick” in front of an ask. “Quick call”, “quick question”, and “quick back and forth” are everywhere, and the word apologises for the ask before the reader has heard it. Name a real number (“Worth 15 minutes?”) or ask about interest (“Worth a look?”).

Two variations, one idea per pair

Every step goes out in two versions, and each pair tests one thing: a question against a statement, proof against an offer, asking who owns it against naming the colleagues you’d ask instead. When a pair differs in one way, the result tells you something. When it differs in five, you’ve learned which email did better, not why.

Keep each pair small enough to read side by side, and give it enough sends to mean something before you call a winner.

These are defaults, not rules

Everything above is what works for us on cold audiences, people who have never heard of us. Some of it changes when the situation does.

• Warm audiences can take a meeting ask in step 1. Someone who downloaded your guide or met you at an event already knows who you are.

• Small markets can justify longer sequences and more personal research per account.

• Other channels are written differently. A LinkedIn message is more conversational than an email, and it isn’t an email pasted into a DM.

The point of knowing why a default exists is that you can tell when it doesn’t apply.

Templates you can use

Every template in our library is built on this method, with two variations per step and the reasoning behind each line:

• B2B cold email sequence template: the default three steps, the one every other template builds on.

• Follow-up after no reply: the email that goes out when the first one gets no reply, in three variations that each bring something new.

• Breakup email: the last email in the sequence, three ways to ask the routing question, and what to do with the replies.

More are on the way in the template library. If you’d rather we write and run it for you, that’s Full Outbound.