Email marketing

How to Build an Intent-Based Audience Segmentation Strategy for B2B Email

In B2B email, what a contact did with your last campaigns says more than their job title. Build four intent tiers from clean click data, then add firmographics on top.

A laptop screen displaying digital marketing analytics charts and graphs
Photo by Goumbik

Picture two contacts on the same client list. One is a finance manager at a 40-person freight firm who clicked the pricing link in the last two campaigns. The other is a head of operations at a 2,000-person manufacturer. On paper they're a perfect fit, but they haven't opened anything since October. Segment by job title and company size and the manufacturer gets the heavier nurture flow. Segment by what each person did and the finance manager gets a booking link today.

That's the argument of this guide. In B2B email, what a contact did with your last campaigns tells you more about buying intent than their title, headcount or industry. Firmographics tell you who could buy. Interaction history tells you who's looking right now. Interaction segments only work under three conditions, though. You strip machine opens out of the data first. You rank signals by how much effort they cost the reader. And you treat firmographics as a filter on top of intent, never as the segment itself.

You'll get seven steps below. Each one gives you the action, what you should see and the mistake to avoid, and after them comes a way to prove the result to a client. The timing matters. It's February 2024, and the major mailbox providers began enforcing their bulk sender rules this month, so mailing people who never open now costs you more than wasted sends. We use SendTo, our white-label email marketing platform, as the worked example.

Finance manager at a freight company office checking shipping paperwork, with a warehouse visible through the window
Photo by ernestoeslava

By the end you will have four intent tiers running, and you need 90 days of clean campaign history to start

You'll finish with four live tiers for each client: clickers, openers who didn't click, non-openers, and new subscribers with no history yet. Each tier gets one message plan. You'll also run a control test that shows whether the tiers beat the old lists. That's audience segmentation based on email interaction history, built so a client can see why it works.

If the term is new to you, a segment in email marketing is a part of a list chosen by a rule. That part gets a message written for it instead of for everyone. The rule can be a job title, a location or a past action. This guide makes past actions the main rule and keeps everything else as a second layer.

Check four things before you start. If a client is brand new and has no history, put everyone in the new subscriber tier and send three or four campaigns first. Tiers built on two sends will mislead you.

  • At least three to five past campaigns, or about 90 days of regular sends. With less than that, one lucky subject line decides who looks engaged.
  • A verified list. Dead addresses skew every rate, and SendTo verifies addresses before each send.
  • Click tracking switched on for every campaign in the window. Without clicks you only have opens, which is the weakest signal.
  • A platform that can segment by behaviour on past campaigns, not only by contact fields.

Step 1: Strip machine opens out of your history before you trust a single segment

First, rebuild the engagement history so it counts people, not software. Corporate security gateways open messages and follow their links to check them before a person ever sees the inbox. Privacy features in some mail apps load images in advance, which records an open whether anyone read the email or not. Both show up as engagement unless you filter them out.

Once you filter, expect the open rate on a B2B list to drop, sometimes sharply. A client who's used to seeing 45% might now see something in the low twenties. The campaign didn't get worse. It's the first honest number they've had, and every tier you build sits on it.

SendTo's reports exclude security scanner opens, so the opens and clicks in the client portal and partner view come from people. Opens from privacy image loading are harder to separate from real ones on any platform, and that's the main reason the next step leans on clicks over opens. For the full mechanics, read our piece on how security scanners distort email open tracking metrics.

The mistake here is building a "hot openers" segment from raw data. On a list full of corporate addresses, that segment fills up with firewalls. You end up sending your most direct offer to machines, and the people who actually read get nothing written for them.

Why this matters more in B2B than in consumer email

Consumer inboxes rarely sit behind a link scanner. Business inboxes often do, and the bigger the company, the more likely it is. A scanner follows every link as the email arrives, so one contact can look like they clicked all six links in a newsletter within seconds of delivery. No person does that.

So treat two patterns as warning signs, not intent: a click within a minute or so of delivery, and a contact who clicks every link at once. Real clicks come later and usually hit one or two links.

Step 2: Rank audience segmentation based on email interaction history by effort, not by volume

Next, rank the signals. The rule is simple: a signal is worth what it cost the reader to give it. Opening an email costs almost nothing. Clicking through to a pricing page costs attention and a little commitment. Booking a call costs time in their calendar.

Interaction signals ranked by the effort they cost the reader
SignalEffort for the readerWhat it tells youWeight
Booked a call or repliedHighReady to talk now5
Clicked a pricing, demo or booking linkHighComparing options4
Clicked a case study or blog linkMediumResearching the problem3
Opened three or more campaigns, no clicksLowPaying attention, not acting2
Opened onceVery lowBarely aware of you1
No human opensNoneNot looking, or not receiving0

You should end up with a short ranked list of three to five signals for each client. Keep it short. A software firm's top signal might be a click on the demo link. An accountancy practice might rank a reply above everything else, because its buyers write back instead of booking. Write each client's list down before you build anything, so the tiers come from the list.

The mistake is giving opens and clicks the same weight because opens are more plentiful. A big number feels like evidence, but a thousand opens with no clicks tell you less than forty clicks on a pricing link. Even after Step 1, some of those opens are still privacy image loads.

Step 3: Build four tiers and expect the top one to be small

Now create the four segments. Clickers from your recent campaigns make up the top tier. Openers who didn't click come second, and non-openers come third. New subscribers who joined too recently to have any history get a tier of their own, so you don't mistake them for cold contacts.

SendTo builds segments from clicks, unclicked opens and non-openers on past campaigns. You pick the campaigns, pick the behaviour and save, which takes a few minutes per client. Before you use the clicker tier, check it for the instant-click pattern from Step 1.

Expect a small top tier. On a typical B2B list, clickers might be 5 to 10% of contacts, yet they bring in the majority of replies and bookings. That's the 80/20 rule in email marketing: a small share of subscribers drives most of the results. The split is rarely exactly 80/20, but the shape holds. A small top tier means your filter worked.

The mistake is padding the top tier to make it look bigger, either by adding people who opened once or by stretching the lookback window. A large hot tier full of lukewarm contacts gets the direct offer and ignores it. That teaches your client that segmentation doesn't work.

How far back to look

Set the window by how often the client sends. Weekly senders should use the last six to eight campaigns, roughly two months. Fortnightly senders should use four to six. Monthly senders should use the last three to four, roughly a quarter.

Longer windows dilute the signal. Someone who clicked once last spring isn't looking now, and counting them as a clicker lets one old click outweigh months of silence. If a client sends at irregular intervals, count campaigns instead of days.

Step 4: Layer firmographics on top of intent, never underneath it

This step is the heart of the guide. Firmographics are company facts like industry, company size and the contact's role. They still matter, but they decide how you talk to someone inside a tier, never which tier they belong to. The tier sets the offer and how often you send. Firmographics set the examples, the proof and the wording.

The tier sets the offer, and firmographics shape the message
TierOffer and how often you sendWhat the firmographic layer changes
ClickersA direct next step, up to weeklyAn industry case study, with pricing framed for their size
Unclicked openersOne clear call to action with proof, fortnightlyA proof point from a company their size
Non-openersA re-engagement attempt, monthly at mostA subject line in their role's language
New subscribersA short welcome sequenceA welcome example from their industry

You should end up with that matrix for each client. A clicker at a ten-person firm and a clicker at a thousand-person firm both get the booking link this week. The first reads about a small firm that got set up in an afternoon. The second reads about a rollout across departments. When those clickers turn into ongoing work, it feeds straight into a retainer strategy for B2B email marketing clients.

The mistake is sending the "enterprise" segment one nurture flow whatever its engagement. Cold enterprise contacts get pushed hard, so they complain or go quiet. Meanwhile the warm ones get the same slow drip as everyone else. A job title tells you someone could buy. It can't tell you whether they're looking.

Step 5: Give each tier one job and write the message to that job

Write four briefs, one per tier, and give each brief a single job. A tier with two jobs gets an email that does neither. The job comes from where the contact sits on the effort ladder from Step 2, so the brief writes itself once you know the tier.

  • Clickers: one direct next step. Book a call, see a demo or reply with a yes. Keep it short, because they already know what you sell.
  • Unclicked openers: one clear call to action, backed by one piece of proof. They're reading but not convinced, so give them a number or a named result from a company like theirs.
  • Non-openers: change the approach, not the volume. Try a new subject line style, a different send time, or a three-line plain-text check-in that reads like a person wrote it.
  • New subscribers: a short welcome sequence of three or four emails that says what they'll get and asks for one small action early.

You should end up with four one-page briefs per client. In SendTo, you feed each brief and the client's brand preset into the designer and get an HTML draft back in seconds, then remix until the tone fits the tier. The built-in reviewer checks subject lines, links and structure before anything goes out, which matters when four versions go to one list. Our piece on using AI email reviewers to standardise quality control covers how to run that check across a team.

Clients often ask about the 5 C's of email. There's no single standard list, and different trainers use different words. A common version is clear, concise, compelling, credible and a call to action. That set works fine as a checklist for each brief. Just notice that a tiered plan already makes the hardest one easy. Writing to one tier with one job is what keeps an email concise.

The mistake here is writing one master email and trimming it four ways. The clicker version still carries the long explainer, and the non-opener version still has the same subject line they ignored last time. Start each brief from the tier's job, not from last week's newsletter.

Step 6: Set suppression and re-engagement rules for non-openers before the mailbox providers set them for you

Put two numbers in writing for each client. The first is how many campaigns in a row a contact can go without a human open before they move into re-engagement. The second is how many more silent sends move them into suppression.

Here are thresholds we'd start with. Weekly senders move a contact to re-engagement after eight unopened campaigns, then suppress after three more re-engagement attempts with no response. Monthly senders use four unopened campaigns, then two attempts. That answers the question of how many unopened emails is too many. It's about two to three months of silence, then a short last try, then stop.

The timing isn't optional any more. This month the major mailbox providers began enforcing their bulk sender requirements. Senders have to authenticate their mail with SPF, DKIM and DMARC. Every marketing email needs an easy unsubscribe, including a one-click option, and opt-outs have to be honoured quickly. And the spam complaint rate has to stay well below 0.3%. Non-openers are where complaints come from. Someone who hasn't read you in four months is the person most likely to hit the spam button instead of hunting for the unsubscribe link.

So yes, the new rules change how you treat non-openers. Before, mailing them wasted sends. Now it risks the client's whole domain. SendTo verifies addresses before every send and applies suppression across the platform, so a suppressed contact stays suppressed even if the client imports the same spreadsheet again next quarter. Spreading volume across several brand-matched sending domains also stops one domain carrying all the risk. Our article on multi-domain distribution for high-volume client campaigns explains the mechanics.

You should see the non-opener tier shrink over a few months while complaint rates stay flat or fall. If the tier stays the same size, contacts are going into re-engagement but nothing is moving them out.

The mistake is deleting non-openers instead of suppressing them. A deleted contact comes back on the next CSV upload and starts receiving full campaigns again, as if nothing happened.

Step 7: Refresh the tiers on a schedule, because intent goes stale

A click in November says very little about February. Review or rebuild each client's tiers before every major send, and at least once a month. Because the segments in SendTo are built from rules on past campaigns, a refresh means updating which campaigns count, not rebuilding from scratch.

You should see contacts moving both ways. Some openers start clicking and move up. Some clickers go quiet and drop to the opener tier. Both are good news. Movement means the tiers track real behaviour, and a tier that never changes is a list someone forgot to update.

Expect the seasons to show in your numbers. Public interest in email marketing peaks in November, December and January and hits its low in June. A client's top tier will often look healthiest in January and thinnest in early summer. Don't panic in June and widen the lookback window to fill the tier back up. Keep the rules steady and let the size move.

The mistake is building the tiers once as static exported lists. They feel finished, and by the third month they're sending the booking link to people who stopped reading in autumn.

Check it worked by sending against a control, not against last month

A control group is the only result a client can't argue with. Before the next campaign, hold back a random 10% of each tier and send that slice the client's old, untiered email. Or run one campaign to the old demographic list alongside the tiered version. Then compare the two groups from the same send on the same day.

Measure per contact, not in totals. Compare real human clicks, replies and bookings for every hundred contacts, plus unsubscribe and complaint rates. SendTo shows human opens, clicks and bookings in both the client portal and the partner view, with scanner opens already removed. That means you and the client look at the same honest figures. For how those numbers are separated, see our guide to tracking real human open rates.

Comparing against last month misleads you in two ways. Seasonality moves engagement on its own, so a January campaign beats a December one whatever you did. And list growth changes the mix, because a month with lots of new sign-ups pulls the averages down. A control sent alongside the test cancels both out.

The control also shows a client the five benefits of market segmentation in their own data. Messages fit the reader better, more people respond, fewer people unsubscribe or complain, the sending domain stays healthier, and the client learns which of its buyers are actually in the market. That last one is the benefit agencies tend to undersell.

If the tiers do not beat your old lists, the problem is usually one of four things

A tiered send that ties with the old list isn't proof the method fails. In our experience the cause is one of four, and you can check them in this order.

  1. Machine opens still inflate the top tiers. Tighten the click rules so instant clicks and every-link clicks don't count, and drop open-only signals from the clicker tier entirely.
  2. The lookback window is too long. Cut it back to the ranges in Step 3. One click from last spring shouldn't earn anyone the booking link.
  3. Every tier still gets the same message. Put the four emails side by side. If the subject lines and calls to action match, the tiers are only labels.
  4. The list is too small for one send to show a difference. On lists under a few thousand contacts, a top tier of 5% is too few people for one campaign to prove anything. Run the control across four to six campaigns and judge the total.

On small lists, segmentation still pays. Writing to non-openers differently protects the domain whatever the list size. You just need more sends before the numbers mean anything.

If you write about email for agencies and freelancers, this is a method you can teach and recommend with real numbers behind it. Our affiliate programme pays you 20% of the partner subscription of every agency you refer, every month that agency pays SendTo. You never take a share of what that agency's own clients pay it. Attribution lasts 90 days from the first click, the first touch wins, and commission is paid monthly by PayPal.

Recommend SendTo to the agencies you write for and earn 20% of their partner subscription every month they pay.

Join the SendTo affiliate programme

Frequently asked questions

What's a segment in email marketing?

It's a part of your list chosen by a rule, like a job title, a location or a past action. That part gets a message written for it instead of for everyone. In this approach, past actions are the main rule and everything else is a second layer.

What does the 80/20 rule mean for email marketing?

A small share of your subscribers drives most of your results. On a typical B2B list, clickers often make up 5 to 10% of contacts but bring in most replies and bookings. The split is rarely exactly 80/20, but the shape holds. A small top tier means your filter worked.

Is segmenting by behaviour really better than by job title or industry in B2B?

Yes, for deciding who gets which offer. Firmographics tell you who could buy. Interaction history tells you who's looking right now. Use company size, role and industry to shape the examples and wording inside each tier, never to decide the tier.

Can I trust open data when so many B2B contacts sit behind security scanners?

Not raw open data. Scanners open emails and follow links before a person sees them, so strip machine opens first and lean on clicks. SendTo's reports exclude security scanner opens, but privacy image loads are hard to separate from real opens on any platform.

How far back should I look when deciding who counts as engaged?

Match the window to how often the client sends: the last six to eight campaigns for weekly senders, four to six for fortnightly, three to four for monthly. Longer windows let one old click outweigh months of silence.

How much campaign history do I need before building intent tiers?

You need at least three to five past campaigns, or about 90 days of regular sends with click tracking on. With less, one lucky subject line decides who looks engaged. Put brand-new clients in the new subscriber tier and send three or four campaigns first.

Do the bulk sender rules change how I should treat non-openers?

Yes. The major mailbox providers began enforcing bulk sender rules in February 2024, so mailing people who never open now costs you more than wasted sends. Send non-openers a re-engagement attempt once a month at most, and set your own suppression rules before the providers set them for you.

How do I prove to a client that the tiers beat their old lists?

Send against a control, not against last month. Keep a group on the old list treatment, run the tiered sends alongside it over the same period and compare the two. A before-and-after comparison can't tell you whether the tiers did the work.

Everything this blog teaches, on a platform with your name on it.

Email campaigns, websites, landing pages and print for your clients, made from a short brief and sent under your brand. Create your account and look around with a sample workspace loaded.

No contract · Cancel any time · Explore everything before you pay