Segmenting your email list: who gets which message?
Segmenting an email list means splitting your recipients into groups that each receive a different message, based on what they bought, opened or told you themselves. The goal is not to send more email but to send less irrelevant email: every message that does not fit costs you an open, a click or eventually an unsubscribe.
At 2Mail you manage your contact lists centrally, split them up and see per send what each group did with it — so you never have to guess which segment is working.
Why segmenting pays off more than sending more often
Four effects you only see once you stop sending one message to everybody.
Relevance is the real alternative to frequency
Anyone wanting their numbers to go up tends to send more. That works for an edition or two, after which the unsubscribes arrive. Putting the same effort into relevance lasts longer: fewer emails, but to people they actually fit.
Your reputation rides on behaviour, not volume
Mailbox providers watch what recipients do with your message: open it, ignore it, delete it, mark it as spam. An email to a group with no interest in it drags that average down — and with it the odds that your next message reaches the inbox.
Without segments you are measuring an average
A campaign that scores respectably on average often hides one group that responded superbly and one that did nothing at all. Only by measuring separately do you see which of the two deserves more.
Unsubscribing is a symptom, not an accident
Most unsubscribes do not happen because somebody no longer wants to hear from you, but because the content has failed to be about them three times running. A group that gets less but better-aimed mail stays.
From one list to usable segments in four steps
Start small. Two or three segments you genuinely use in every campaign are worth more than twenty you created once.
Formulate the question first, the segment second
Not "which groups can I make?" but "which message do I want to word differently, and for whom?". If you have no different message for a group, you do not need that segment. That one question cuts half the ideas straight away.
Look at the data you already hold
Purchase date, sign-up date, product bought, language, and what someone did with earlier emails: it is usually already there. Start from that rather than sending out a survey first — data you have no concrete use for is data you need not collect.
Build two or three segments and use them immediately
The classics, which almost always pay off: people who bought recently, people who bought once but not for a long time, and people who signed up without ever buying. Each of those three deserves a different tone, and you have the data for it.
Measure per segment and prune
Look at opens, clicks, unsubscribes and bounces per group rather than across the whole send. A segment that after three campaigns does not diverge from the average adds nothing and can go back into the main list.
Which data may you use to segment?
Segmenting is a processing of personal data, even when you split on purchase history alone. Two practical rules keep it manageable: collect nothing you have no concrete use for, and make sure people know what you are splitting on where that is not self-evident.
Splitting on what someone bought from you, or on the language they chose, is generally unproblematic. Once your segment becomes a profile — inferred interests, behaviour across channels, sensitive categories — the question changes and you should review your legal basis and your duty to inform first.
Where that processing happens counts just as much: segmentation data is personal data like any other.
Read about processing inside the EUSix ways segmentation falls apart
How you go about it at 2Mail
Segmentation is mostly editorial work; the platform should not get in your way while you do it.
