To segment an ecommerce email list, you need about ten groups built around three questions: where the contact is in their journey (new subscriber, customer, lapsed customer), how engaged they are and what they're interested in. Everything else is either for analysis only or can be ignored. Below you'll find the difference between a list and a segment, how to define an "engaged" subscriber now that opens are inflated, a table that separates the segments you send to from the ones you keep only to read the numbers, and the real case of one customer.
Lists vs. segments: the difference that matters
- A list is a static set of contacts who have agreed to receive your emails. It only changes when someone subscribes, unsubscribes, or is added or removed by hand. It tells you who you can contact and where they came from: the pop-up on your site, a signup page, a lead generation campaign.
- A segment is a dynamic group defined by rules. "Has bought at least once", "clicked in the last 60 days", "viewed the shoes category". It updates itself: people who meet the conditions join, people who stop meeting them drop out.
In practice, lists are for collecting contacts and for triggering flows that people go through only once, like the welcome series. Segments decide who gets each campaign, trigger flows when someone changes status (after their first order, for example) and show you how your contact base moves over time.
How do I segment an ecommerce email list?
Start from three axes, not fifty rules.
- Stage. New subscriber, subscriber with no orders, one-time customer, repeat customer, customer who hasn't bought in a while. Each stage has a different goal: the first order, the second, the comeback.
- Engagement. Who interacts with your emails and your site, and how recently. This axis decides how many emails someone gets, not what they get.
- Interest. The category or product they viewed, added to cart or bought. This axis decides what you tell them.
Every useful segment comes from combining these axes. "Engaged subscribers with no orders who viewed the herbal tea category" is a segment with a clear goal. "Women aged 25 to 34" almost never is, because the data is often incomplete and says little about what they want to buy.
What "engaged" means after Apple Mail Privacy Protection
For years, engagement was measured with opens. With Mail Privacy Protection turned on, though, Apple Mail downloads the images in every email in the background, including the open pixel: the open is recorded even if the email was never read.
Some platforms flag these opens. Klaviyo, for example, shows on every open whether it came from a device with this protection; in that case, though, it can't know whether a real read also happened. Where that data exists, leave those opens out of your engagement rules.
A sturdier definition of "engaged in the last N days" combines several signals:
- a click on an email;
- a site visit from an identified contact (recognized by the platform's cookie or logged in);
- a product added to cart or a started checkout;
- an order;
- an open, but only if it wasn't recorded automatically.
New subscribers are the exception. Anyone who signed up recently should count as engaged even with no history, or they'd be left out right when they're paying the most attention.
The segments you need and the ones for analysis only
The rules are written in plain words so you can rebuild them on any platform. Adapt the time windows: shorten them if your customers buy every three weeks, lengthen them if you sell mattresses.
| Segment | Typical rule | What it's for | Use |
|---|---|---|---|
| Recent subscribers | Subscribed in the last 14 days | Key messages for new subscribers, excluding them from signup pop-ups | Sending |
| Subscribers with no order | Subscribed more than 14 days ago, no orders | Content that explains product, shipping and returns; first-purchase offers | Sending |
| Unused welcome code | Received the first-purchase code, no orders, code not expired | A reminder before it expires | Sending |
| Orders without a code | At least one order, never used a discount code | Seeing how much your sales depend on discounts; leaving them out of promotional emails when they don't need them | Sending and analysis |
| Category interest | Viewed or bought products in a category in the last 60 days | New arrivals and guides for that category | Sending |
| Awareness | Engaged with emails, no recent product page views | Emails that introduce the brand, tips, useful content | Sending |
| Consideration | Viewed products or added to cart in the last 30 days, no orders | Products, launches, comparisons, reviews | Sending |
| Engaged 30 / 60 / 90 days | At least one signal among click, visit, cart and order in the window | Deciding send frequency | Sending |
| Expected date of next order | Next order predicted within the next 15 days | Replenishment, complementary products | Sending |
| High value | Historic or predicted lifetime value above a threshold | Early access, loyalty program, review requests | Sending |
| Churn risk | High probability of not ordering again | Targeted win-back emails, before they stop buying altogether | Sending |
| Never engaged | Subscribed more than 90 days ago, no click, visit or order ever | Reconfirmation, then exclusion | Limited sending |
| Inactive | No signal in the last 180 days | Reconfirmation, exclusion, possibly deletion | Analysis |
| Non-subscribers with no orders | In the database, but no consent and no orders | Checking how much they weigh on your total contacts | Analysis only |
The last segment never gets marketing emails, because these people never opted in. If you sell in the EU, that consent is a legal requirement. In the US, the CAN-SPAM Act works on an opt-out basis, but emailing people who never asked to hear from you is a quick way to collect spam complaints. The segment shows how many contacts you're storing without being able to use them, such as shoppers who started checkout without subscribing.
Engagement and send frequency
The main reason to segment by engagement isn't writing different copy. It's deciding how often to email each group. Here's a starting point to calibrate on your own data:
- Engaged in the last 30 days: they get every campaign in your content calendar.
- Engaged 31 to 90 days ago: one or two campaigns a week, only the strongest.
- Engaged 91 to 180 days ago: only the big moments, like seasonal sales or Black Friday.
- No signal for more than 180 days: no regular campaigns. Two or three reconfirmation emails a few days apart, with a clear choice between staying subscribed and leaving, then exclusion.
A real number gives you the scale. In a footwear store we work with, the contacts unengaged for three months (on the list for at least 90 days, at least five emails received, no opens or clicks in the last 90) number 14,999, and that's an underestimate because of Apple's automatic opens. Three campaigns a week to the whole list would mean writing often to almost fifteen thousand people who haven't reacted in months: they're the ones who mark you as spam and drag deliverability down for everyone. For thresholds and domain reputation, see the article on why emails go to spam.
On list cleaning: a flow that automatically unsubscribes inactive contacts is convenient but hard to keep under control. Many stores prefer to review inactive contacts every two or three months and decide by hand, so they know who leaves and why.
Predictive segments: useful only with enough data
Some platforms estimate, for each customer, the date of their next order, their predicted lifetime value and the probability that they won't come back. That way email follows the customer's calendar: instead of emailing everyone 45 days after purchase, you email each person when they're probably about to reorder.
The limit is the amount of data. Klaviyo, for example, only calculates these predictions if at least 500 customers have placed an order, there are at least 180 days of order history with orders in the last 30 days, and some customers have ordered three or more times. A young store won't get there, and that's fine: a simple rule ("last order 40 to 60 days ago, no order since") does a similar job.
Segments and advertising
Many platforms sync segments with Meta and Google: recent customers to exclude from acquisition campaigns, subscribers with no orders to reach with retargeting, high-value customers as the seed for lookalike audiences, unengaged contacts to reach on Meta instead of with more emails. The footwear store's four segments, with rules and numbers, are in the article on email and Meta Ads.
Segments you can ignore
- Segments that are too small. A group of 80 people won't give you readable results and doubles the writing work. If a segment stays under a few hundred contacts, merge it with a neighboring one.
- Segments built only on opens. Since Mail Privacy Protection, they include people who don't read.
- Segments created for one campaign and never used again. Delete them every so often, or nobody will remember what they were for.
From segment to person: one customer's case
The segment decides who to email, not what will persuade each of the people in it. To see this, you have to go down to a single profile: a customer of the same footwear store, with no personal data. It's a single case, so an example of the mechanism and not a statistic.
What happened. On October 1 she starts a checkout without completing it and gets three recovery emails. On October 9 a cart and a viewed product show up, each with its own emails. That same day she orders, 2.9 hours after the cart email: €264. Before, she behaved like the store's potential buyers; after the order she moves into the group that bought in the last 45 days and leaves the acquisition campaigns.
The angles tried. Each email leans on one argument, the angle, and each angle was used once or twice.
| Moment | Angles tried | Outcome |
|---|---|---|
| Viewed product | Social proof, season, belonging, problem to solve, urgency | No order |
| Cart | Product benefits, urgency, aspiration, novelty, authority, discount | Order, attributed to novelty |
| Checkout | Curiosity, belonging, season | No order |
The resulting profile. Novelty is at the top, because it's the only angle that drove an order. Belonging, season, social proof and urgency sit in the middle (53 out of 100), aspiration, authority and discount lower down (26). These are relative scores, not conversion rates.
The subject line has a history too: of her 16 opens in the last 180 days, 12 were on short subject lines (under 35 characters) and 4 on medium ones, and the ones with her name work better. The next one will be short and include her name. Opens have the limits described above: here they decide the shape of the subject line, not whether to email her.
The next move. Novelty has never been tried at checkout, so that's where it gets tried. Next in line are purchase history, recent visits, product benefits and her birthday, if the store collects it.
What it teaches, with caution:
- Store and person can match, but it's not guaranteed. Across all the store's recovery emails, novelty has the highest order rate, as the article on Meta Ads shows. For her it went that way; for another customer, social proof might win.
- Low engagement doesn't rule out a purchase. Her score is 30 out of 100. Engagement decides frequency, it doesn't predict who buys.
- With little data per person, try several angles. After one order, no angle is "hers" forever.
So segments decide who to email and how often; what to write is chosen on the profile when there's a history, and on the segment when there isn't.
To see how segmentation fits into the rest of your email program, from automated flows to the content calendar, start with the ecommerce email marketing strategy.
