Cold email follow up: how many, how far apart

The standard advice is that most replies come from follow ups rather than the first email, and it is broadly true. The conclusion usually drawn from it - that more follow ups are better - does not follow, and running it to its logical end is how senders burn domains.

Each additional message reaches a smaller pool of people who have already chosen not to reply, while the complaint risk per message rises. There is a point where the next follow up costs more in reputation than it returns in conversations.

How many follow ups

Three to four messages in total, including the first. Beyond that, marginal reply rate falls off sharply while the share of recipients who mark the message as spam keeps climbing.

The reasoning is worth stating plainly: someone who has ignored four messages is not undecided. Sequences of eight or ten exist because they produce a few extra replies in a spreadsheet, and the spreadsheet does not record the complaints that degrade the domain for every future campaign.

Spacing

Widening gaps work better than fixed intervals. Early follow ups catch people who were busy; later ones catch a change in circumstance, which takes longer to happen.

Avoid sending on the same weekday at the same hour each time. A precisely regular cadence is a machine signature, and it is visible to receiving systems in a way it is not to recipients.

What each follow up should contain

One rule covers most of it: never send a message whose only content is that you sent an earlier message. Bumping a thread, just circling back and checking in add nothing for the recipient and train them to ignore your name.

1. Second message: a different angle on the same problem, not a restatement.
2. Third message: new evidence - a comparable example, a specific result, a relevant change.
3. Fourth message: a short, genuine close that makes not replying easy.

The fourth message performs better than people expect, because it removes pressure rather than adding it. It should read as closing the file, and it should be true - if you say you will stop, stop.

Same thread or new thread

Reply on the existing thread for the second and third messages. It gives the recipient the context without restating it, and it looks like ordinary correspondence rather than a campaign.

One exception: if the original message may never have arrived, a fresh thread with a different subject is worth one attempt. But if you suspect delivery failure across a whole campaign, that is an infrastructure problem and no amount of resending fixes it - check placement first.

Stop conditions

A sequence needs rules for stopping that fire automatically, because relying on manual review at volume means they will not fire at all.

• Any reply, including an out of office with a named alternative contact.
• Any unsubscribe or request to stop, honoured immediately and permanently.
• A hard bounce, which should also remove the address from every future list.
• A soft bounce repeating across two consecutive sends.

Continuing to send after a request to stop is both a legal problem under CAN-SPAM and GDPR and the fastest route to complaints, which do more damage to a sending domain than almost anything else. Honouring opt-outs immediately is part of what separates cold email from spam rather than an optional courtesy.

Follow ups multiply your volume

This is the part most sequence planning misses. A four-message sequence means each prospect receives four emails, so a campaign to five hundred people is two thousand sends, not five hundred.

Since a Microsoft 365 mailbox safely handles around three cold emails a day, that campaign needs enough mailboxes to absorb the full sequence volume across its duration. Planning the sequence and the inbox count separately is how senders end up pushing mailboxes past a safe rate halfway through a campaign.

Follow ups also concentrate risk: the same recipients hearing from the same domain repeatedly. Spreading a sequence across several sending domains limits the damage when one of them degrades.