Behavioural segmentation groups subscribers by what they did rather than by who they are. It outperforms demographic segmentation consistently, for a simple reason: past behaviour predicts future behaviour, while demographics predict a stereotype.
The practical form is a set of triggers that fire per person rather than per campaign, which is why behavioural segmentation and automation end up being the same subject.
The eight worth building, in order
1. Engagement recency
Opened or clicked in the last 90 days, 90–365 days, or over a year. The first segment to build and the only one that is not optional, because it protects inbox placement for everything else. Suppress the third group after one win-back attempt.
Data needed: last open, last click. You already have both.
2. Purchase recency and frequency
One purchase or several, recent or not. The recent single-purchase group is the most convertible segment most businesses have, and most send it the same newsletter as everyone else. Formalised, this is RFM segmentation.
Data needed: order count, last order date, total spend.
3. Cart or checkout abandonment
Started and did not finish. The highest-converting email most businesses can send, because the intent is unambiguous and recent. See abandoned cart best practices.
Data needed: a cart event from your commerce platform.
4. Category or product affinity
What someone actually buys or browses within your range. Turns a generic promotion into a relevant one, and it is the input for recommendation blocks. See dynamic content.
Data needed: product or category on each order.
5. Email content affinity
Which topics someone clicks. If you write about five things and a subscriber only ever clicks one, that is a segment and a preference they never told you about. Underused because it requires tagging your links by topic, which takes a convention rather than a tool.
Data needed: click data with topic tags on links.
6. Site behaviour
Viewed a page and did not convert, visited pricing three times, read the same article twice. Powerful and requires connecting your analytics to your email platform, which is the highest-effort item so far.
Data needed: page-view events pushed to the subscriber record.
7. Product usage
For software, the richest source available: activated or not, hit a limit, used a feature, stopped logging in. Every meaningful lifecycle email in a SaaS business is triggered from this. See email marketing for SaaS.
Data needed: product events pushed to the subscriber record via API or webhook.
8. Support and service history
Raised a ticket, left a review, returned an item. Small volumes, high sensitivity — the value here is mostly in suppression, since sending a promotional campaign to someone with an open complaint is the kind of mistake people remember.
Data needed: ticket status on the subscriber record.
The order matters
Build them in the order above, because it runs from most valuable per unit of effort to least. Items one and two need data you already hold. Items six and seven need an integration.
Most senders would get more from doing one and two properly than from attempting six.
Making it work mechanically
Behavioural segments are queries against fields on the subscriber record, so the whole thing reduces to keeping those fields current. The seven that unlock most of this list: last open, last click, order count, last order date, total spend, last product category, and signup source.
Update them on a schedule from your commerce or product data. Any platform with custom fields and dynamic segments supports it — in IGSendMail you can push updates through the API or the MCP server from a nightly job, which means the segments are queries against current data rather than static lists that quietly go stale.
The failure mode to avoid: building a segment as a one-time export. A static list is a snapshot of who qualified last month, and it will keep emailing people who no longer belong in it.
The limit
Every segment is a version of an email someone has to write. Four behavioural segments you genuinely customise beat twelve that all receive the same content with a different opening line.
Build in order, stop when the writing capacity runs out, and revisit when it grows. The broader framework is in email segmentation strategies and lifecycle email marketing.
Frequently asked questions
What is behavioral segmentation in email marketing?
Grouping subscribers by what they did — opened, clicked, bought, browsed, used a feature — rather than by who they are. It outperforms demographic segmentation because past behaviour predicts future behaviour more reliably than attributes do.
Which behavioral segment should I build first?
Engagement recency, because suppressing long-dormant subscribers improves inbox placement for everyone else and therefore improves every other segment's results at the same time.
What data do I need for behavioral segmentation?
Seven fields cover most of it: last open, last click, order count, last order date, total spend, last product category and signup source. The first two you already have; the rest come from your commerce or product system.
How is behavioral segmentation different from automation?
They overlap almost entirely. A behavioural segment defines who qualifies; an automation is what happens when they do. In practice you build them together, because a segment nobody sends anything to achieves nothing.
