Key takeaway
An email nurture campaign is a sequence that answers the questions somebody needs answered before they'll buy, at their pace rather than yours. An agent can now draft the whole sequence for you, which leaves two jobs it can't do: deciding what "ready" means, and turning the flow on. Write "ready" as a segment before you write email one, because a segment is the only form of that definition the send path and the agent can both act on. Judge a nurture on how many people got ready, not on opens, and a lead who filled in a form hasn't subscribed to your list.
What an email nurture campaign is.
An email nurture campaign is a sequence sent to somebody who gave you their address for one reason, answering the questions they need answered before they'll buy, at their pace rather than yours. Nurture is the only word in this corner of email named after what it's trying to do rather than how it works, and it's the one everybody defines by how it works instead.
Nurture emails teach more than they sell, they run without anybody pressing send, and they're aimed at a person who is interested and not yet ready. That much is conventional and it's correct. What separates a nurture from a broadcast is the timing: a broadcast has a send date and one audience, and a nurture has an entry condition and an audience of one person at a time, each moving at their own speed.
Stopping there is what lets "nurture" quietly absorb a discount code, a cart reminder, and the monthly newsletter.
Nurture, drip, and workflow aren't three words for the same thing.
A drip is named for its schedule, so it's defined by when it sends. A workflow is named for its structure, so it's defined by how it's put together. A nurture is named for what it's trying to do to a person, which makes it the only one of the three that can succeed or fail at something.
That isn't a semantic point, it's a difference in what you're allowed to judge. A drip gets judged on whether it ran, a workflow on whether it's built correctly, and a nurture on whether somebody who wasn't ready became ready. Most nurture campaigns are built in a workflow builder and reviewed as workflows, which is exactly why they end up judged on whether they ran.
The same sequence is usually all three at once, and the words genuinely do overlap in every tool's interface. The question worth asking about your own is which of the three you'd defend it as.
Decide what "ready" looks like, then write it as a segment.
Readiness is an observable. It's something that happens in your own data on a particular day: a demo booked, a reply to a real question, a trial started, a pricing page opened twice in a week, or a form filled in that only a buyer would bother with.
The goal always turns out to be more sales. That's a wish, not a goal. If you can't name the event you'd see, you can't tell whether the sequence worked, you can't tell when a person is done with it, and you can't answer the question everybody asks next, which is how many emails to write. The answer to that one is however many it takes to get somebody to the signal, and not one more.
Prose is where most nurture plans stop, and prose is where they stay unfalsifiable. The test worth applying is whether the definition survives being typed into a filter builder. Our own documentation draws the line in three sentences: "Saved views remember how you like to inspect the contacts table. Segments define an audience that can be reused for sending, automation, and agent workflows. If a filter set should affect delivery, save it as a segment." A readiness definition affects delivery, because it decides who stops receiving the sequence. So it is a segment. If it won't compile into one, you don't have a definition of ready, you have a hope.
Compiling it costs you vagueness, which is the point. A segment is built from filter groups: profile, list membership, engagement, deliverability, and behavior. Behavior is the one that carries event history rules, and the documented rules are performed, not performed, count at least, and count at most. Date filters take relative windows such as "within the last 30 days" and negative windows such as "has not opened in the last 30 days." Nearly every honest readiness definition turns out to be one positive behavior rule plus one negative window, and writing it in that form forces two answers you would otherwise leave open: what counts, and how recently.
Signals are specific to the business and usually smaller than people expect. A B2B software seller might take a second visit to the pricing page inside a week. A consultancy might take a reply that names a date. A course seller might take a syllabus download, because that's the one page nobody opens idly. A signal you have to ask a salesperson about isn't a signal, it's an opinion with a delay attached.
Some purchases give you nothing at all until the purchase itself, and there the honest answer is a proxy you've chosen on purpose and written down, rather than a proxy you inherited from a template.
Where a readiness segment goes wrong.
Two failures make a readiness rule quietly return the wrong people, and neither of them shows up anywhere in the sequence.
The first is opens. An open reads like the cheapest engagement signal available and it's the one I'd trust least. I've audited accounts where opened activity was so badly undercounted that it sat at near parity with clicks, when opens should outnumber clicks by an order of magnitude. Any segment built on "opened" was unusable there, and we had to lean entirely on clicks for engaged. If your entry rule or your exit rule rests on an open, it rests on a measurement that fails silently rather than loudly, and you will read the resulting segment as a fact about your buyers.
The second failure lives in the negative half, and it is the one nobody warns you about. A condition written as "source is not import" looks like it means everyone who didn't arrive from an import. It doesn't. A comparison against a missing value isn't false, it's unknown, and unknown rows don't come back. The PostgreSQL manual puts it plainly: "Ordinary comparison operators yield null (signifying 'unknown'), not true or false, when either input is null. For example, 7 = NULL yields null, as does 7 <> NULL." So a negative filter on a field that is blank for some contacts silently drops them rather than keeping them.
That bites hardest on exactly this kind of list. On a migrated account the genuine organic signups are usually the contacts with no source recorded at all, so the intuitive filter throws away the people you most wanted to nurture, and the segment count looks plausible enough that nobody checks. Write the negation with the blank case spelled out, and read the count before you point a sequence at it.
A segment can't start the flow, and that's the right way round.
Having decided readiness is a segment, the next thing people go looking for is the trigger that fires when somebody joins it. There isn't one. The documented flow triggers are three: a contact added to a list, an event received, and a tag added. A segment is the audience for a campaign and a condition inside a flow step, not a starting gun.
That reads like a limitation for about a minute. Then it stops, because a nurture sequence doesn't want to start when somebody becomes ready. It wants to start when somebody enters, and stop when they become ready. Entry is a moment, which is what a trigger is for. Readiness is a state, which is what a condition is for. Build it that way and the segment you wrote becomes the exit, re-evaluated every time the flow reaches a split, instead of a trigger that fires once and then has to be undone.
The event trigger is the one worth learning, because it lets readiness come from your product rather than from your email data. A product event posted to /v1/my/events can start or personalize a flow, and you choose the event name and the JSON fields it carries: names are lowercase letters, numbers and underscores up to 64 characters, and the data object goes to 32 KB. Nitrosend doesn't inspect your database or infer a schema from it, so the contract is yours to keep stable. If a flow expects event.order.total, keep sending that path with the same meaning. A readiness signal that already exists in your product, a trial started or a quote accepted, belongs there rather than in a filter that tries to infer it from clicks.
The five questions a nurture sequence answers, in order.
Every guide to this subject hands you a list of email types: welcome, educational, product explainer, social proof, webinar invite. Type is the wrong axis. A case study sent to somebody who doesn't yet know what the product does fails on position rather than on copy, and rewriting the copy won't save it. What a reader is really doing is working through a set of questions, and they arrive in a stubborn order.
- Why should I care? the problem, named in their words, before anything at all is offered.
- Is this for someone like me? the shape of the person it works for, so the wrong reader can leave early.
- How does it actually work? the mechanism, plainly, because a buyer who can't explain it internally can't buy it.
- Has it worked for anybody else? proof, which only lands once the first three have been answered.
- What does it cost me to try? the ask, and the smallest one that gets you a real answer.
The order is the content plan, and it's the only part of a nurture campaign that transfers between businesses. Your product, your proof, and your price are yours. The sequence of questions is close to universal, because it's how a person decides anything that costs money. When the order is wrong the symptom is unmistakable: the sequence gets rewritten sentence by sentence, quarter after quarter, and never gets fixed, because the problem was never in the sentences.
A returning buyer arrives already holding the first three answers, and sending them anyway is how a nurture campaign insults the person it was built to keep.
One email, one question, one ask.
An email in a nurture sequence has one job: answer one of those questions, and earn the open of the next one. "One call to action per email" is that job stated as a rule, with the reason for it left out.
The constraint isn't aesthetic. Two asks in one email means you can't read the click, so the single number that would have told you which question landed is gone. Cut the second ask, even when it looks free. Treat each email as a chapter rather than a summary, and you get a usable test for a draft you're unsure about: a chapter ends where the next one has to begin.
Two things don't belong in a nurture sequence and both get put there constantly. A discount is a price cut, so it ends the sequence rather than advancing it, and what you learn is that somebody will take money off rather than that they were convinced. A cart reminder is a transaction prompt, which is a different program with a different trigger and a different clock.
The last email in a sequence is allowed to break all of this, because a sequence that never asks for anything is a newsletter.
A lead who filled in a form didn't subscribe to your list.
A nurture list is built from people who handed over an address for a reason that wasn't this: a whitepaper, a webinar, a pricing inquiry, a badge scan at a conference. That makes it the shape of program most likely to be mailing somebody who never asked, which has a permission answer and a deliverability answer, and they're different answers.
A teaching email sent to move somebody toward a purchase is marketing, whatever it teaches. That matters because the messages a nurture sequence is genuinely not marketing are short and enumerable: completing something the person already asked for, a warranty or safety notice, a change to the terms of an account they hold, and delivering something already owed. Everything else in the sequence is marketing mail and carries what marketing mail carries.
Permission also travels badly. The soft version most senders lean on, that a download or a webinar registration is consent to a sequence, is narrower than they assume almost everywhere, and it is narrowest where the address was collected for something other than being sold to. A download is not a negotiation for a sale.
The mailbox providers ask a blunter question. Google's sender guidelines say not to send messages to people who didn't sign up, and they name the consequence: those recipients mark the mail as spam, and future messages get marked as spam too. Yahoo asks senders to verify they're only mailing users who specifically requested it, and both put the ceiling at a spam rate below 0.3%. I'd treat that 0.3% as generous rather than strict: the automatic flag I ran at 0.5% at my last company was already five times too high by the time it caught anything.
Permission is the floor, and none of it makes the sequence wanted. The practical version is asking for the nurture at the moment you take the address, which costs one checkbox and settles all three positions at once.
What a nurture campaign is judged on.
A send has metrics and a program has a result. The two get confused because the platform reports the first one automatically and nobody reports the second. What actually describes a nurture campaign is a smaller set of measures, and not one of them is an open rate.
- Readiness rate: the share of people who entered the sequence and later matched the readiness segment you defined.
- Time to ready: how long that took, measured from entry rather than from the first send.
- Question drop-off: which of the five questions people stop opening at, which names the email to rewrite.
- Complaint and unsubscribe rate: the cost of the program, and the number that decides whether the next one reaches anybody.
Open rate and click rate are diagnostic rather than results. They'll tell you which email is weak once you already know the program is working, and they can't tell you that it is. The uncomfortable half is worth saying plainly: a nurture campaign with excellent engagement and no movement in readiness rate is a well-written newsletter, and it can run for years looking healthy.
A readiness rate needs something to compare it against before it means anything, and the honest comparison is the leads who never entered the sequence at all, which most senders don't keep.
Composing the sequence is one call. Turning it on is another.
A nurture campaign has traditionally been stored in three places. The timing is in a builder, the copy is in a doc, and the reason each email sits where it does is in somebody's head, usually somebody who has left. The reasoning is the part worth keeping. The questions in order, the segment that ends the sequence, and what each email is for will fit in a paragraph, and a paragraph can be reviewed by somebody who wasn't in the room, argued with, rebuilt for a second segment, and compared against what ran last year. A canvas can't be any of those things.
That's the part we built Nitrosend around. In 2026 the job changed: agents build the campaign, segment the audience, draft the copy and stage the send, and the marketer is the director rather than the operator. Every capability is an API endpoint and an MCP tool before it's a screen, so campaigns, flows, contacts, lists and segments are things an agent creates, changes and explains back to you. A sequence described precisely enough to review is described precisely enough to build.
Note the word stage. Composing a flow and activating it are deliberately two different calls: the compose tool creates or edits an automation-flow draft and never activates it, and activation is a separate call made after a review step. That split isn't a missing feature, it's the safety model, and it exists because of something I did to myself. I was testing a re-send, the flow triggered on its own, and it went to dummy contacts rather than real people purely on the timing. What I wrote that afternoon was that we needed a hard safeguard stopping Claude or any other agent, even one running with its permission prompts turned off, from firing a full blast without explicit consent from a person. Humans approve. Agents operate.
The honest state of this today is that agents are good at composing flows and the drafts pile up unactivated. Closing the draft-to-activation gap is probably the biggest unlock left in agentic email, and it isn't a copywriting problem. It's the same problem as the rest of this page: nobody wrote down what the sequence was supposed to achieve, so nobody could decide it was ready to run. A readiness segment is what makes that decision reviewable in a minute rather than a meeting.
Two limits are worth knowing before you point an agent at a twelve-email nurture. A long flow for a brand with a lot of assets is arithmetically impossible to push through one tool call, so compose by intent rather than pasting finished emails, and break the work into smaller missions with a defined goal per flow instead of asking for full autonomy. And what comes back is set by what you give it: generic prompts receive generic results, and the fix for mediocre AI-generated email isn't better prompts, it's better references. Nitrosend's Brand Kit learns a brand's colors, fonts and tone from a site, so an agent drafting message four of a nurture doesn't start from a blank template.
A canvas is genuinely easier for seeing a long sequence at a glance, and for one sequence that never changes, none of this pays for itself.
If your nurture campaign is one nobody can state the purpose of, the fix starts with writing that purpose down and then compiling the readiness half of it into a segment. Nitrosend takes a flow with its triggers, steps and branches as an agent command and builds what you described, as a draft you approve. Contacts are unlimited on every plan including Free, which keeps a list you don't control out of the billing conversation. Start free and write the argument for one sequence.
Go deeper
Sources
- Google, Email sender guidelines: not sending messages to people who didn't sign up, and keeping spam rates reported in Postmaster Tools below 0.3%.
- Yahoo, Sender Best Practices: verifying you're only sending mail to users who specifically requested it, and a spam rate below 0.3%.
- PostgreSQL, Comparison Functions and Operators: ordinary comparison operators yield null rather than true or false when either input is null, which is why a negative filter drops blank rows.
- Nitrosend docs, Segments: saved views versus segments, the filter groups, the behavior rules, and relative and negative date windows.
- Nitrosend docs, Flows: the three flow triggers, the step types, and the draft, active, paused and archived states.
- Nitrosend docs, Events API: posting a product event to start or personalize a flow, the event name rules, and the 32 KB data limit.
Common questions
An email nurture campaign is a sequence sent to somebody who gave you their address for another reason, answering the questions they need answered before they'll buy. What makes it a nurture rather than a broadcast is where the timing comes from: a broadcast goes out on a date you pick, and a nurture starts when a person enters it and moves at the speed of their own decision.
There's no universal number. Write as many emails as it takes to answer the questions this buyer actually has, and end the sequence at the readiness segment rather than at a count. If you can't compile that definition into a segment, the length question has no answer and whatever number you pick is a guess.
They're named after different things. Drip names the delivery: the emails go out on a fixed schedule, and the word says nothing about their purpose. Nurture names the purpose: the sequence exists to take somebody from interested to ready to buy. Most sequences fit both descriptions at once, and only the second one can be right or wrong, because only the second one claims a result. That's why it's worth deciding which label yours is really earning.
One question, answered properly, with one ask at the end of it. The questions a buyer works through are why they should care, whether it's for someone like them, how it actually works, whether it's worked for anybody else, and what it costs to try. Pick the one this email is for and leave the rest to the emails around it.
It ends when the reader matches the readiness segment you defined before writing email one, such as a demo booked, a trial started, or a reply that names a date. It also ends when they've made clear it never will. A segment can't start a flow, so readiness belongs in a condition inside the sequence rather than in its trigger, and if you can't say what would end it, the sequence has no finish line.
That depends on what they agreed to when they gave you the address. A teaching email sent to move somebody toward a purchase counts as marketing rather than as a transactional message, so it needs the consent marketing needs. In the UK the soft opt-in is narrower than most senders assume: it wants details taken during a sale or negotiations for one, marketing for similar products or services, and a free way to refuse at collection and in every message. Asking for the nurture on the form itself settles it.
Two numbers answer that. Count how many of the people who entered went on to do the thing you decided counts as ready, and count how long that took them from the day they entered. Opens and clicks can't settle it: they diagnose one email at a time, and a sequence can post healthy engagement for years while nobody in it gets any closer to buying. If readiness isn't moving, the sequence isn't doing its job, however well the emails read.
It can draft the whole thing, and it deliberately can't switch it on. Composing a flow and activating it are two separate calls in Nitrosend: the compose tool creates or edits an automation-flow draft and never activates it, and activation happens afterwards, once a person has reviewed the draft. That split is why an agent composing a twelve-email nurture is safe to run. The two decisions still yours are what "ready" means, written as a segment, and whether this draft goes live.
No, though they overlap at the start. A welcome sequence sets expectations and confirms what somebody signed up for, and a nurture answers the questions standing between them and a purchase. In practice the first nurture email often does both jobs, and knowing which job each email is doing is more useful than drawing a hard line between the two.