Email Segmentation for Ecommerce: The 5-Layer Model for DTC Brands

Email segmentation for ecommerce works best when you build it in layers—starting with engagement health, then purchase behavior, then RFM tiers, then product affinity, then predictive data. Each layer is only accurate once the layers below it are stable. Skip the foundation and your advanced segments contain the wrong people.

Here's what most articles on segmentation won't tell you: the order you build segments in matters as much as the segments themselves.

Every top result on this topic gives you a list. Thirteen segmentation ideas. Eleven tips. Seven strategies. What none of them give you is a system—something that tells you what to build first, what depends on what, and why a "champions" segment built on a dirty list is worse than no champions segment at all.

This is that system. The 5-Layer Segmentation Model is the framework we use to audit and rebuild retention programs for DTC brands. By the end of this article, you'll know exactly which segments to build, in what order, and what each layer is worth in revenue terms.

Why Does Segmentation Order Matter?

Segmentation has dependencies. You cannot build an accurate RFM "Champions" segment without first defining your active/inactive threshold, because an unfiltered list includes churned customers, bot signups, and engaged-but-never-purchased subscribers who will inflate your champion count and make your revenue benchmarks meaningless.

Think of it this way: an RFM model scores customers on Recency, Frequency, and Monetary value. If your "high recency" bucket includes subscribers who opened emails via Apple Mail's prefetch feature (which fires tracking pixels automatically, regardless of whether anyone actually read the email), you're not segmenting by recency—you're segmenting by noise.

The same problem cascades upward. A product affinity segment built on top of an uncleaned purchase segment will include lapsed customers who bought once two years ago. A predictive CLV model trained on a list with inactive subscribers will have skewed inputs.

This is why we use a layered model. Each layer creates the data integrity that the next layer depends on.

What Is the 5-Layer Segmentation Model?

The 5-Layer Segmentation Model is a dependency-sequenced framework for DTC email segmentation that tells operators not just what segments to build, but in what order. The five layers—Engagement Foundation, Purchase Behavior, RFM Tiers, Product Affinity, and Predictive and Zero-Party Data—form a hierarchy where each layer is only reliable once the layers below it are operational.

Here's the full structure:

The 5-Layer Segmentation Model

• Layer 1 — Engagement Foundation: Active vs. inactive split, suppression segments, deliverability protection
• Layer 2 — Purchase Behavior: Non-buyer, first-time buyer, repeat buyer, lapsed buyer
• Layer 3 — RFM Tiers: Champions, Loyal, At-Risk, Lost (Recency × Frequency × Monetary)
• Layer 4 — Product Affinity: Category-based segments for cross-sell and upsell
• Layer 5 — Predictive and Zero-Party Data: Predicted CLV, churn risk, quiz and survey responses

Each layer builds on the one below it. You don't build Layer 3 before Layer 1 is stable. Let's walk through each one.

Download the 5-Layer Segmentation Checklist — the exact audit framework we use across client Klaviyo accounts.

Why Is Layer 1 the Non-Negotiable Starting Point for Engagement?

Layer 1 is an engagement gate: it splits your list into people who are actively interacting with your emails and people who aren't. This layer exists to protect deliverability—every campaign you send to unengaged contacts quietly degrades your sender reputation and reduces inbox placement for the contacts who actually want to hear from you.

Active subscriber is a contact who has clicked an email in the last 30 to 90 days. Note that we use clicks, not opens. Since iOS 15, Apple Mail pre-fetches email content and fires tracking pixels automatically, which inflates open rates for Apple Mail users. Open rate is no longer a reliable engagement signal for 50 to 60 percent of most DTC lists. Click activity is the only clean signal you have.

In Klaviyo, your Layer 1 segments look like this:

• Engaged: Clicked email in last 30 days. Full campaign cadence. Receives all sends.
• Semi-Engaged: Clicked in last 31 to 60 days. Reduced cadence—your strongest campaigns only.
• Disengaging: No click in 61 to 90 days. Flows only. Entering winback territory.
• Unengaged: No click in 90-plus days. Suppress from all campaigns. Sunset flow only.

The suppression segment is the other side of Layer 1 that most brands ignore. Suppression isn't just about unsubscribes and hard bounces (though those matter). It also includes recent buyers who are mid-flow and shouldn't receive campaign interruptions, contacts in active abandonment flows, and refund-only customers whose one purchase ended in a return.

Solid email list hygiene practices are the prerequisite for Layer 1 working correctly. A list with 30 percent unengaged contacts dragging down your sender reputation will compromise every segment you build above it. Industry research from Klaviyo's deliverability team consistently shows that suppressing unengaged contacts improves inbox placement rates by 10 to 20 percent for most senders.

Once Layer 1 is stable—meaning your campaigns are going to contacts with genuine click-based engagement—you have the data integrity to move to Layer 2.

Where Does Retention Actually Live in Purchase Behavior Segmentation?

Layer 2 splits your engaged subscribers by purchase history. This single split is the most impactful segmentation decision most DTC brands never make correctly: buyers and non-buyers require fundamentally different messages, and sending the same campaign to both groups serves neither well.

The four segments in Layer 2:

• Non-Buyer: On the list, zero orders all time. These are conversion targets. Your welcome flow, browse abandonment, and cart abandonment flows own this segment.
• One-Time Buyer: Exactly one order all time. This is the most critical retention gap in DTC. The jump from one purchase to two is where most brands leak the most lifetime value.
• Repeat Buyer: Two to three orders all time. Growth segment. These customers have validated the brand. Cross-sell, upsell, and bundles do the work here.
• Lapsed Buyer: One or more past orders, but nothing in the last 90-plus days (adjust based on your product's replenishment cycle). Recovery target. Enter a winback flow for lapsed customers before the window closes.

The first-to-second purchase rate is the key metric at this layer. According to Blossom's benchmark data, a strong DTC program converts 30 to 40 percent of one-time buyers to repeat buyers. Below 20 percent, fixing the second-purchase problem is more urgent than any segmentation work above Layer 2.

In Klaviyo, Layer 2 segments use the "Has placed order" condition with order count and date filters. For Shopify brands, purchase data flows directly into Klaviyo profile properties through the native integration, so these segments stay current without manual updates.

Once your purchase-tier segments are stable and your welcome flow is targeting new subscribers correctly—with sequences tailored for first-time visitors—you're ready to build Layer 3.

How Does RFM Segmentation Work in Klaviyo?

RFM segmentation scores customers on three dimensions—Recency (when they last bought), Frequency (how many times they've bought), and Monetary value (how much they've spent)—and groups them into tiers based on their combined scores. In Klaviyo, this is implemented using "Has placed order" conditions, date-of-last-order properties, and order-count filters layered together into named tier segments.

RFM segmentation is not a single segment. It's a matrix of conditions. Here's how the four primary tiers map to Klaviyo logic:

RFM Tier Definitions in Klaviyo

• Champions: Placed 3-plus orders AND last order within 60 days AND total revenue above your top-quartile threshold. These are your highest-value, most recently active customers.
• Loyal: Placed 2-plus orders AND last order within 90 days. Engaged and returning, not yet in champion territory.
• At-Risk: Placed 2-plus orders BUT last order was 91 to 180 days ago. They were repeat buyers who are starting to slip.
• Lost: Placed 1-plus orders, no order in 180-plus days, no email click in 90-plus days. At this point, the cost of re-engagement likely exceeds the expected return for most of this group.

The critical implementation note: you cannot build accurate At-Risk or Lost segments without first completing Layer 1 and Layer 2. If your engagement gate isn't clean, your At-Risk segment will contain both genuinely at-risk customers and unengaged contacts who never had strong engagement to begin with. These two groups need completely different interventions.

According to Blossom's benchmark data, RFM Champions—typically 8 to 12 percent of a healthy DTC list—drive 20 to 30 percent of email-attributed revenue. That's the highest revenue concentration of any segment type in the entire architecture. No other segment produces that ratio of list share to revenue contribution.

This is why building Layer 3 correctly is worth the effort. A clean Champions segment is the highest-ROI target in your entire program. A dirty Champions segment gives you false confidence and wrong revenue attribution.

Not sure if your Champions segment is clean? We audit Klaviyo segmentation architectures every week — here's what we look at in yours.

How Do You Build Cross-Sell Segments That Actually Convert?

Layer 4 uses purchase history to group customers by the categories or product lines they've bought from, enabling cross-sell and upsell campaigns that feel relevant rather than random. Product affinity segments are only worth building once your purchase-behavior baseline from Layer 2 is stable—otherwise you're building affinity models on incomplete data.

Product affinity segmentation is the practice of grouping customers by their purchase patterns across your catalog, then using those groups to send targeted campaigns for complementary products.

A skincare brand might build:

• Cleanser buyers who haven't bought moisturizer (cross-sell opportunity)
• Moisturizer buyers who haven't bought SPF (routine completion)
• Buyers of the entry-level product who haven't tried the outcomes tied to your specific list premium version (upsell opportunity)

In Klaviyo, these segments use the "Has placed order at least once where items contain" condition filtered by product title, SKU, or collection. For Shopify brands with well-structured collections, collection-level filtering is usually cleaner than SKU-level filtering.

The revenue logic here is straightforward: a customer who bought Product A and has now seen a targeted campaign for the logical Product B is far more likely to convert than the same customer seeing a generic bestsellers email. The message matches where they are in the product journey.

Product affinity segments also feed directly into your campaign calendar built around segment tiers. Once you know which customers are in which affinity buckets, you can build campaign weeks around specific cross-sell sequences rather than one-size-fits-all promotions. Research from Shopify's commerce insights team shows that product-affinity-targeted campaigns outperform generic promotional emails by 2 to results that vary by program on revenue per recipient across most DTC categories.

When Should You Add Predictive and Zero-Party Data to Your Segmentation?

Layer 5 uses Klaviyo's predictive analytics features and explicitly collected zero-party data to build forward-looking segments—predicting which customers are about to churn, which are approaching their next purchase window, and which have stated preferences that enable personalized offers. This layer is only worth building once Layers 1 through 4 are operational and generating clean data.

Customer Lifetime Value (CLV) is the total predicted revenue a customer will generate over their relationship with a brand. Klaviyo's predictive analytics model calculates predicted CLV, predicted next order date, and churn risk for every profile with sufficient purchase history. These become segment conditions at Layer 5.

Layer 5 segments include:

• High Predicted CLV (not yet VIP): Customers predicted to become top-tier buyers who haven't crossed the threshold yet. Treat them like VIPs early to accelerate the journey.
• High Churn Risk: Customers Klaviyo's model predicts are unlikely to purchase again. These need aggressive winback intervention before the window closes.
• Predicted Next Order Window: Customers whose next purchase is predicted within the next 14 days. This is your replenishment segment—reach them before they search elsewhere.
• Zero-Party Data Segments: Contacts who completed a quiz, survey, or preference center, grouped by their stated preferences (skin type, fitness goal, flavor preference, pet breed).

Zero-party data is information a customer explicitly shares with a brand—quiz answers, survey responses, preference center selections—as opposed to behavioral data inferred from their actions. It's the cleanest personalization signal available because the customer told you directly what they care about.

In practice, zero-party data powers the most targeted campaigns in a mature program. A quiz that asks "What's your biggest skincare concern?" gives you a personalization signal on day one, before any behavioral data exists. Welcome flow variants built around quiz answers convert at higher rates than generic sequences because the product recommendations actually match what the customer said they need. Combined with a thoughtful welcome offer strategy for new subscribers, this approach can significantly lift first-purchase conversion from the welcome flow.

Why Is the Segment Build Order Non-Negotiable?

Each layer of the 5-Layer Model has prerequisite conditions from the layers below it. Skipping layers doesn't save time—it produces segments that appear to work but contain the wrong people, generating misleading performance data that causes bad optimization decisions downstream.

Here's the dependency chain in concrete terms:

1. You cannot build accurate RFM segments (Layer 3) without a clean purchase behavior baseline (Layer 2).
2. You cannot build a clean purchase behavior baseline (Layer 2) without first suppressing unengaged contacts from your campaign sends (Layer 1).
3. You cannot suppress unengaged contacts reliably (Layer 1) without click-based engagement conditions, because open rates are corrupted by Apple Mail Privacy Protection.
4. You cannot build useful product affinity segments (Layer 4) without stable RFM tiers (Layer 3), because affinity modeling on a list that includes churned customers and inactive subscribers produces false affinity signals.
5. Predictive CLV (Layer 5) requires clean purchase history data flowing from all prior layers—Klaviyo's model trains on order data, so dirty order data produces wrong predictions.

The practical implication: before you ask "what segments should I build," ask "what is the cleanest layer I have right now?" Then build the next one up. If Layer 1 isn't stable, everything above it is built on sand.

How Often Do Segments Need to Be Refreshed?

Segments expire. A Champions segment built in Q4 can contain 30 to 40 percent churned customers by Q2 if you don't account for engagement decay—and sending to it as if everyone is still an active champion means overpaying for sends, inflating apparent engagement metrics, and eventually damaging deliverability. Most segments need a maintenance cadence, not just a creation date.

Segment decay is real and underappreciated. The Klaviyo segment engine updates membership dynamically based on current conditions, so your segment definitions stay accurate as long as the conditions are correctly configured. The failure mode isn't that Klaviyo forgets to update—it's that operators set conditions that don't account for time-based decay.

Common decay patterns to watch:

• Champions decay: A champion who hasn't purchased in 120 days is no longer a champion by any reasonable definition, but if your Champions segment only filters on order count and monetary value (not recency), they'll stay in that segment indefinitely.
• Engagement window drift: As your list ages, more contacts drift toward the semi-engaged and disengaged buckets. If you don't tighten your engagement window as this happens, your "Engaged" segment gradually fills with people who aren't actually engaged.
• Over-segmentation on small lists: A 20,000-person list split into 40 micro-segments means some segments receive sends so infrequently that Klaviyo has no recent engagement signal to work with, which can affect sender reputation signals over time. There's an optimal segmentation depth for each list size—and the minimum viable send segment is typically 500-plus recipients to generate statistically readable performance data.

The maintenance cadence we recommend: review Layer 1 engagement thresholds quarterly. Review Layer 3 RFM tier conditions after every major sales event (Q4 purchase surges can inflate champion counts temporarily). Review Layer 4 product affinity segments when you significantly change your catalog.

Frequently Asked Questions

What are the best email segmentation strategies for ecommerce?

The best email segmentation strategies for ecommerce build in dependency order: engagement health first, then purchase behavior, then RFM tiers. Most brands skip directly to advanced segments like RFM champions or product affinity without building the engagement foundation that makes those segments accurate. A clean Layer 1 engagement gate—built on click activity rather than opens—is the non-negotiable starting point for any segmentation architecture that will produce reliable revenue data.

How does Klaviyo segmentation work for DTC brands?

Klaviyo segmentation works by applying conditions to contact profiles—properties like order count, last order date, email click activity, predicted CLV, and custom fields from Shopify or quiz integrations. Segments update dynamically as profiles meet or stop meeting conditions. For DTC brands, the most effective Klaviyo segments combine engagement conditions (clicked email in last 30 days) with purchase conditions (has placed order, order count, date of last order) to create layered targeting that sends the right message to the right person at the right lifecycle stage.

What is RFM segmentation in email marketing?

RFM segmentation groups customers by Recency (when they last purchased), Frequency (how many times they've purchased), and Monetary value (how much they've spent). In email marketing, RFM tiers—Champions, Loyal, At-Risk, Lost—determine campaign targeting and flow triggers. In Klaviyo, RFM segments are built using combinations of "Has placed order" conditions, date-of-last-order filters, order count thresholds, and total revenue properties. RFM is only reliable once a clean engagement foundation and purchase behavior baseline are in place.

How many email segments do I need for a DTC brand?

Most DTC brands need five to eight active campaign segments at Layer 3 maturity: Engaged Non-Buyer, One-Time Buyer, Repeat Buyer, VIP, Lapsed Buyer, and a suppression segment with standard exclusions. More segments than your team can produce differentiated content for create complexity without value. The minimum viable send segment is 500-plus recipients—below that, performance metrics are too noisy to optimize against. Start simple and add precision only when list volume and team bandwidth support it.

How do I measure the performance of email segments?

Measure email segment performance using Revenue Per Recipient (RPR) as the primary metric, not open rate. Revenue per recipient (RPR) is the average revenue generated per email sent, calculated by dividing total email-attributed revenue by emails delivered. RPR lets you compare segment performance across different list sizes—a 500-person Champions segment with $4.50 RPR is clearly outperforming a 5,000-person general send at $0.60 RPR, even though the smaller segment generates less total revenue. Track RPR by segment monthly and compare against your Layer 3 benchmarks to identify which segments are over- or under-performing their expected contribution.

The Architecture Argument

Segmentation isn't a feature you turn on. It's an architecture you build—and like any architecture, the foundation determines what's possible above it.

The brands that get the most out of their Klaviyo segmentation aren't the ones with the most segments. They're the ones whose layers are clean from the bottom up. Their engagement gate is tight. Their purchase behavior segments are accurate. Their RFM tiers reflect actual customer value. That foundation is what makes their Champions campaigns hit the way they should.

If you've been building segments ad hoc—grabbing what's easy without a sequenced system—the 5-Layer Model gives you a way to diagnose where the architecture breaks down and what to fix first.

Key takeaways:

• Build Layer 1 (engagement foundation) before Layer 3 (RFM) or your champion segments will contain inactive subscribers and your revenue benchmarks will be wrong.
• Use click-based engagement conditions in Klaviyo, not opens—Apple Mail Privacy Protection makes open data unreliable for 50-plus percent of most lists.
• RFM in Klaviyo is a matrix of conditions, not a single segment—it requires a clean purchase-behavior baseline from Layer 2 to produce accurate tiers.
• Segment decay is real—time-based conditions in your RFM segments prevent churned customers from staying in your Champions bucket indefinitely.
• There's an optimal segmentation depth for each list size—over-segmentation on lists under 20K creates send frequency gaps and noisy performance data.

Want us to build this architecture for your brand? Book a free strategy call — we'll map out which of the 5 layers to tackle first based on your list size and revenue targets.

Get tactics like this in your inbox every week. Subscribe to our newsletter →