Customer Segmentation for Ecommerce: Types and Examples

Every customer segmentation guide on page one of Google opens with the same four boxes: demographic, geographic, psychographic, behavioural. They are correct and almost useless for an online store, because a store’s most valuable segments do not come from who someone is. They come from what they have bought, when, and how often.

This guide covers the theory quickly, then spends its time on what matters for ecommerce and DTC brands: behavioural segmentation examples you can build in Klaviyo this week, and how to decide which segments deserve their own messaging.

---

What is customer segmentation?

Customer segmentation is splitting your customer base into groups that share traits or behaviours, so each group gets messages and offers that fit it. For ecommerce brands, the most useful segments are behavioural and value-based, built from purchase history, browsing and engagement, because past behaviour predicts the next order better than demographics do.

The payoff is relevance. A first-time buyer, a five-order regular and a customer who lapsed a year ago need different emails. Sending all three the same weekly campaign wastes the regular’s attention, discounts to someone who would have paid full price, and ignores the lapser’s actual reason for leaving.

---

The types of customer segmentation (and how they apply to a store)

Type	Based on	Ecommerce example	How useful for DTC
Demographic	Age, gender, income, life stage	Menswear vs womenswear buyers	Moderate: useful for catalogue split, weak predictor of the next order
Geographic	Country, region, climate, time zone	Southern hemisphere shoppers get winter range in June	Useful for send time, shipping offers and seasonality
Psychographic	Values, interests, lifestyle	Vegan buyers, performance athletes, “clean beauty” shoppers	High when you collect it directly (quizzes, preference centres)
Behavioural	Purchases, browsing, engagement, loyalty	Bought twice in 90 days, viewed a category three times	Highest: the core of ecommerce segmentation
Value-based	Revenue, margin, lifetime value	Top 10% by predicted CLV	High: decides who gets VIP treatment and who never needs a discount
Needs-based	The problem the customer is solving	Sleep vs energy vs gut health supplement buyers	High for health, wellness and skincare

Two practical notes. First, psychographic and needs-based data is hard to infer but easy to ask for. That is what zero-party data is: a quiz answer or a preference-centre choice, stored on the profile. Second, value-based segmentation is best done with a scoring model. Our RFM segmentation guide covers recency, frequency and monetary scoring in full, so we will not repeat it here.

---

Behavioural segmentation examples for ecommerce

Behavioural segmentation groups customers by what they do. These are the segments we see earn their keep in most ecommerce accounts, grouped by the question they answer.

Purchase behaviour

1. Prospects. Subscribed, never purchased. They need proof and a reason to buy, not loyalty perks.
2. First-time buyers. One order, inside the last 60 to 90 days. The most important segment in the account, because this is where the second order is won.
3. Repeat buyers. Two or more orders. They trust you; sell them the next logical product.
4. VIPs. Top decile by spend or predicted CLV. Early access and recognition, not blanket discounts.
5. Discount-only buyers. Every order used a code. Useful for deciding who hears about sales first, and for keeping discount-driven messaging away from customers who buy at full price.

Product affinity

6. Bought A, never bought B. The simplest cross-sell segment and often the highest revenue per recipient. Cleanser buyers without a moisturiser; protein buyers without creatine.
7. Category buyers. Customers who only buy from one category, for category-specific launches and content.
8. Price tier. Customers who buy premium SKUs versus entry SKUs. They respond to different offers.

Timing and occasion

9. Due for reorder. Past the median replenishment window for the product they bought. The trigger for replenishment flows.
10. Seasonal or event buyers. Bought only during BFCM, Mother’s Day or a summer sale. Brief them before the event, rest them outside it.
11. Lapsing. Past their own expected next order date, not a blanket 90 days. Win-back material.

Engagement

12. Engaged in the last 30, 60, 90 days. Opened or clicked (with care, given Apple Mail Privacy Protection) or visited the site. The foundation of send volume decisions.
13. Unengaged 180+ days. Suppress or sunset. They cost you on the Klaviyo invoice and in inbox placement.
14. High-intent browsers. Viewed the same product or category several times in a week without buying. Tailor content or trigger browse abandonment.

Most behavioural segmentation guides stop at “cart abandoners”. In practice, the segments above are where the money is. The exact Klaviyo definitions for the core seven are in our Klaviyo segmentation guide, and the supplement-specific versions (purchase cycles, product affinity, predictive) are in the supplement segmentation guide.

---

How to segment customers: a six-step process for DTC brands

1. Start from a revenue question

“Segment our list” is not a goal. “Lift second-purchase rate from first-time buyers” is. So is “stop discounting to customers who would pay full price”. Each goal points to the segments you need and the metric that proves they worked.

2. Audit the data you already have

Most stores are sitting on more than they use: order history, product and variant, discount codes used, browse events, form responses, SMS consent, support tags. Check what flows into your email platform and what is stuck in Shopify, your subscription app or your reviews tool.

3. Choose the model

Pick one primary model and a few overlays rather than fifty ad-hoc segments:

• Lifecycle stage (prospect, first-time, repeat, VIP, lapsing, lapsed) for campaign routing. See lifecycle marketing.
• RFM for value and recency scoring.
• Product affinity for cross-sell.
• Stated preferences (quiz or preference-centre data) for content and offer choice.

4. Build it where you send from

Segments that live in a spreadsheet do not get used. Build them in the platform that sends the email and SMS so they update in real time. In Klaviyo that means dynamic segments on placed-order counts and dates, product properties, profile properties and, once your account qualifies, predictive fields such as predicted CLV and expected date of next order. Klaviyo’s help centre lists the threshold (as of September 2026): at least 500 customers with orders, 180 days of order history with orders in the last 30 days, and some customers with three or more orders.

5. Activate with genuinely different messages

A segment that receives the same email as everyone else is not a segment, it is a filter. Change at least one of: the product featured, the offer (or lack of one), the proof used, the send frequency, the channel.

6. Measure by segment and prune

Compare revenue per recipient, conversion rate and unsubscribe rate by segment. Merge segments that behave the same. Split any segment where one subgroup clearly outperforms. Revisit definitions every quarter, because reorder windows and catalogue mix change.

---

What to send each core segment

A segment only earns its place if it gets something different. This is the starting playbook we use for the core ecommerce segments.

Segment	Message focus	Offer approach	Frequency
Prospects (never bought)	Proof, bestsellers, reasons to believe	First-order incentive if you use one	Moderate; tighten as engagement drops
First-time buyers	How to get results, the next logical product	Usually none; save it for later if needed	Driven by the post-purchase flow first
Repeat buyers	New products, cross-sell by affinity, content	Bundles rather than blanket discounts	Your main campaign audience
VIPs	Early access, launches, recognition	Perks (early access, gifts), not percentage off	High for launches; they want to hear from you
Due for reorder	Reminder of what they bought, convenience	Subscribe-and-save offer where relevant	Triggered, not scheduled
Lapsing	What’s new, what they’re missing	Escalate only if value-led emails fail	Low, then sunset
Discount-only buyers	Sales, bundles, clearance	Their natural home is sale sends	Around promotions
Unengaged 180+ days	A last “still want to hear from us?”	None	Sunset, then suppress

The same logic applies to SMS, with tighter frequency. Reserve SMS for the segments where timing genuinely matters: due for reorder, restocks for past buyers, VIP launches.

B2C versus B2B segmentation

Most segmentation guides cover both. B2B segmentation groups accounts by firmographics (industry, company size, revenue) and buying stage, and usually lives in a sales CRM. B2C ecommerce segmentation groups individual shoppers by behaviour and value, and lives in the platform that sends email and SMS. If you sell both, as Popl did with individual and corporate orders, keep the two sets of segments and journeys separate rather than tagging one list two ways.

---

The five tests of a useful segment

Classic marketing theory says an effective segment is measurable, substantial, accessible, differentiable and actionable. Translated into ecommerce terms:

Criterion	The question to ask
Measurable	Can you define it from data you actually hold, and does it update automatically?
Substantial	Is it big enough that a separate email or flow branch is worth building?
Accessible	Are these people reachable, with email or SMS consent and a healthy inbox placement?
Differentiable	Does this group behave differently from the rest (conversion, AOV, reorder timing)?
Actionable	Can you write a genuinely different message or offer for them?

If a proposed segment fails any row, fold it back into a larger one.

---

What good segmentation looks like in practice

At Better Body Co., a supplements brand with a 10+ person in-house email team, the account had been segmenting by broad lifecycle stages only (new, active, lapsed). The audit found 60-70% of the available segmentation surface unused: predictive properties, product-affinity tagging and behavioural cohorts. Activating those segments, alongside fixing flow-versus-campaign reporting and attribution, was part of how email’s share of revenue moved from about 10% to the 20-25% target range after a three-month consulting engagement.

The lesson generalises. Most accounts do not need more segments; they need the right five or six, built properly and actually used.

---

Segmentation mistakes to avoid

• Segmenting by demographics you cannot act on. Age bands rarely change what you send.
• Hundreds of static lists. Build dynamic segments that update themselves.
• Ignoring the engagement layer. A perfectly targeted segment still hurts deliverability if half of it has not opened in a year.
• Over-trusting opens. Apple Mail Privacy Protection inflates opens. Use clicks, site activity and orders to define engagement where you can.
• Never measuring by segment. If you cannot say which segment earns the most per recipient, you cannot improve the model.

---

Next step

Segmentation is the engine inside every retention marketing programme: it decides who gets the replenishment reminder, who gets VIP early access and who never needs a discount. If you want to know which segments your account is missing and what they are worth, our free Klaviyo audit maps your current segments against your order data and sends a written 90-day plan.

---

Sources

• Klaviyo Help Center, predictive analytics requirements