Email list segmentation: from first segments to RFM and lifecycle

Email list segmentation means dividing subscribers into groups that share something: when they signed up, how they engage, what they buy, what they care about, where they live. Each group then gets email written for it instead of one blast for everybody. A small list can start with four groups. RFM, behavioral segments, lifecycle stages and cohorts become worth the effort once you have purchase and activity data to feed them.

What segmentation is and why it matters

A single email for the whole list is always a compromise. The same copy has to work for someone who joined yesterday, for a customer who orders every month, and for the person who grabbed a checklist three years ago and hasn't opened anything since. A subscriber in Dallas has no use for "same-day delivery in Chicago." A monthly buyer doesn't need your welcome sequence.

The number most people quote comes from a Mailchimp study. Mailchimp sampled about 2,000 of its users who segment and took roughly 11,000 segmented campaigns sent to almost 9 million recipients. It then compared them with the same users' unsegmented campaigns. Segmented sends got 14.31% more opens, 100.95% more clicks and 9.37% fewer unsubscribes. The page was last updated in February 2017, years before Apple started hiding opens from senders, so the open figure can't be reproduced the same way today, though the click figure can.

Deliverability is the other reason. Google's sender guidelines suggest you consider unsubscribing recipients who don't open or read your messages. Yahoo recommends monitoring inactive recipients and periodically sending them a reconfirmation email. Both start with putting inactive subscribers in a group of their own.

Where to start: 4 basic segments

Start small. Not twenty segments and an RFM matrix, but four groups you can build in half an hour in any email platform. Treat the time windows below as a starting point and adjust them to how often you send.

1. New subscribers (joined in the last 30 days). They remember signing up but don't know what to expect yet. Give them a welcome series with an introduction to the product, your best content and a first offer. It works best as its own sequence, separate from the regular newsletter.

2. Active (clicked or bought in the last 90 days). Your core audience. You can email them more often and offer more, within reason. Three emails a week is normal for an online store and far too much for most B2B products.

3. Fading (90 to 180 days without a reaction). Maybe their interests moved on, maybe the inbox is abandoned. Stop sending them everything. Either run a short "still there?" sequence or drop to one email a month with only your best material.

4. Long silent (more than 180 days). One last attempt to win them back, then remove whoever doesn't respond. The sequence itself is covered in our guide to reactivating inactive subscribers.

A caveat about "active." If you define it by opens alone, Apple Mail readers will blur the segment. Mail Privacy Protection stops senders from seeing whether a message was opened, so an open from those readers tells you nothing about whether they read you. Clicks, purchases and logins are safer signals. There is more on this in our article on email open rate.

Other segmentation criteria

Once the four basic segments are running, you can layer more on top.

• Geography. Different delivery times, promotions and stock. If your list spans time zones, location tells you when to send so the email arrives during the recipient's working hours. If it spans languages, it tells you which language to use.
• Signup source. Someone who downloaded a "cold email templates" lead magnet cares about sales, and a first email about DNS setup will miss. Someone who joined through a purchase is already a customer, so don't sell to them from scratch.
• Address type. A mailbox on a company domain (jane@company.com) and one on a free provider (jane@gmail.com) usually point to different audiences, often B2B and B2C. Tone, offer and the way you present pricing differ.
• Purchase behavior. Who bought in the last 30 days, 90 days, year. Who bought once and who bought three times or more. A separate offer for people who have already paid is the easiest way to drive repeat sales.
• Interests. If someone keeps clicking email-marketing articles and skipping SEO ones, let that shape what you send next.

Funnel stage is another option: lead, trial, customer, VIP. Leads need to see the value, trial users need help getting started, customers are ready for an upgrade, and VIPs need to be kept without being pestered. The fuller version of this idea is the lifecycle model further down.

What list verification adds to segmentation

Verifying a list does more than find dead addresses. Afterwards you also know which addresses sit on catch-all domains, which are role mailboxes like info@ or sales@, and which come from disposable services. All of it is useful for segments too.

• Catch-all domains accept mail for any address, so the specific mailbox can't be confirmed. Keep those addresses in their own group, send to them less often and watch bounces for that group alone.
• Role-based addresses such as info@ or support@ are read by a team, which may not include the person who subscribed. Leave them out of marketing campaigns and keep them for transactional mail.
• Addresses with an uncertain status can get a small test send, and the bounces decide what happens to them.
• Disposable inboxes stop working within hours or days, so put them straight on the removal list.

uChecker sorts a list into three groups. Good means the mailbox exists and accepts mail, bad means it doesn't exist or the server rejected the check, and unknown means deliverability couldn't be determined. For bad and unknown addresses the report gives a reason, such as a typo, no MX records, a server rejection, a disposable service, a catch-all domain, a role address or a blocklist hit. Those reasons are what you filter on when you build groups in your email platform. Company versus free-provider addresses you can split yourself with a domain filter in any spreadsheet.

RFM analysis: recency, frequency, monetary

Once the basic segments work and you have purchase data, RFM is the next step. A profile field tells you who someone is. RFM tells you what they do and what they're worth. Two subscribers of the same age in the same city can be opposites: one ordered three times last month, the other hasn't opened an email in half a year.

Recency is the time since the last purchase or other action, Frequency is how many times the person bought in a period, and Monetary is how much they spent. Each subscriber gets a score on every axis, and the three scores make a code such as 5-3-4. A spreadsheet is enough to build it.

1. Export from your email platform or CRM, per customer, the date of the last purchase, the number of purchases over the past 6 to 12 months and the total spent.
2. Split each metric into five equal buckets. The fifth with the freshest purchases gets a recency score of 5, the oldest gets 1. Do the same for frequency and monetary value.
3. Join the three scores into a three-digit code.
4. Collapse the codes into six to eight groups; nobody needs 125 segments. Everyone with R, F and M at 4 or above is a top customer, for instance. R at 2 or below with F at 3 or above is a regular buyer who has gone quiet and needs a win-back.

Some platforms do the math for you. Klaviyo has a dedicated RFM report that scores each axis from 1 to 3 and places customers in six groups: Champions, Loyal, Recent, Needs attention, At risk and Inactive. To get the report, an account needs at least 500 customers who have placed an order, at least 180 days of order history and orders within the last 30 days.

The model describes the past and doesn't forecast anything. It does show you right away the customer who used to buy often and big and then went silent.

Behavioral segmentation

RFM fits businesses with orders and receipts. A SaaS product, a media site or a B2B service may have no money axis at all, or everyone pays the same $49 a month. Then you segment by actions. Not "woman from Chicago" but "clicked in three of the last five emails, followed the pricing link and visited the pricing page twice this week."

Signals people usually track:

• Email engagement. Which links they click, on which topics, at what time of day.
• Site and product activity. Pages they view, features they use. This needs your site or app connected to the email platform through a tracking script or the API.
• Content preferences. Someone who has clicked only analytics articles for three months doesn't need design case studies.
• Buying signals. A pricing page visit, an abandoned cart, a downloaded price list, a demo request.

For a SaaS company this might turn one weekly newsletter into four streams: people exploring the product without paying, active paying users, paying users who log in less than they used to, and readers who open emails but have never logged in. The jobs differ too: convert, retain and upsell, prevent churn, warm up. Emails that fire on an event are called triggered emails, and our overview of triggered email campaigns lists the common types.

Lifecycle segmentation

The third model adds time to behavior. Every subscriber is at some stage, and the email should fit it. The question changes from "what do I send this week?" to "what is this person's next step?"

• New subscriber. Doesn't know the product yet. Needs the welcome series: who you are, why you're useful, what the emails will be about.
• Engaged lead. Opens, clicks, visits the site, hasn't bought. Needs answers to objections, so case studies, reviews and an FAQ.
• First-time buyer. Help getting started and recommendations that build on the first order.
• Repeat customer. Loyalty perks and early access, fewer promos and more substance.
• Top customers. Buy often and recommend you to others. A referral program and early access to new products suit them; generic blasts mostly annoy them.
• At risk. Engagement is dropping, RFM scores are sliding. Time for a win-back sequence.
• Lapsed. One final email, and if there's no answer the address leaves the list.

Lifecycle stage says where someone is, RFM says what they're worth, behavior says what to write about. Take a repeat customer with an RFM code of 4-3-5 (buys rarely but spends a lot) whose last three clicks were all on automation articles. That person should get an email about your automation tier.

Cohort analysis

A cohort is everyone who joined in the same period, say January or February. A list-wide average hides change, and comparing cohorts shows it. Suppose February sign-ups click noticeably less three months in than December sign-ups did at the same age. Then something changed at the entry point: the traffic source, the signup form, or the welcome series.

Paired with RFM, cohorts tell you where to look. If recent cohorts slide into the at-risk group faster than older ones, the problem is acquisition or the first emails. If older cohorts hold steady among repeat customers, the product itself is fine.

How to set up segments technically

It depends on your tool, but the logic is the same everywhere. In Mailchimp a segment is a set of conditions: "signup date in the last 30 days" gives you new subscribers, a condition on campaign activity gives you the active group. Brevo, Klaviyo and ActiveCampaign offer similar filters on activity, dates and profile fields.

Segments are either dynamic or static. A dynamic one recalculates on its own, so a subscriber who was silent for six months and then clicks leaves the win-back flow without anyone touching it. A static list stays as it was. If your platform only has static lists, refresh them by hand at least once a month, or someone who came back will keep getting "still there?" emails for weeks.

For more advanced setups, CRMs like HubSpot or Salesforce hold deal and touchpoint data, which you can use for cross-sell segments such as "bought product A but not product B."

Mistakes I see most often

Too many segments too fast. A company with 3,000 subscribers creates 15 segments of 200 people, and each needs its own copy. Two marketers can't keep up, and within a month every segment gets the same email again. A group of 50 also won't give you any result you can trust. Start with two to four groups and add more once those run reliably.

Segments that never update. A subscriber landed in the at-risk group a month ago, has bought three times since, and is still getting win-back emails. Recalculate segments before each send, or at least weekly.

Segmenting a dirty list. Dead mailboxes, typos and spam traps end up in every segment and drag its numbers down. You see weak results in the "active" group and decide segmentation doesn't work, when part of that group has been dead for months. In RFM the same addresses inflate the lapsed group. So the order is: verify the list, then segment, then send.

Segmenting only by demographics. Gender and age aren't behavior. A 35-year-old in Denver might be your most loyal customer or someone who subscribed by accident and never clicked once. Clicks and purchases tell you more than a profile field.

Segments without different content. The list is neatly divided and everyone still gets the same email. That's just different labels on identical boxes.

And the most common mistake is not segmenting at all and sending one campaign to the whole list.

Segmentation with a small list

"I have 500 subscribers. Why would I segment?" I hear this a lot, and I get it: splitting a list that size feels like overkill.

It works the other way round. With a small list every subscriber counts, and losing 50 people to irrelevant emails is 10% of the list. Even a simple new-versus-active split makes sense once you have a few hundred addresses. Mistakes are also cheap at this stage, so send two groups different content and compare clicks while the list grows.

Quick checklist

If you haven't segmented your list yet, here is a plan for the next week:

• Run the whole list through a validator. Remove invalid addresses and set the unknown ones aside.
• Create two segments: active (clicked or bought in the last 90 days) and everyone else.
• Send your next campaign to the active segment only and check the metrics.
• Prepare a short win-back sequence of two or three emails for inactive subscribers.
• Remove anyone who doesn't respond to it.
• A month later, add new subscribers as a third segment with their own welcome series.
• If you have purchase data, build RFM next.