
How to segment your email list by customer behavior
Engagement tiers, purchase-stage segments, RFM in practice and ready-to-build Klaviyo segment definitions, plus how to read people moving between them.
The Wizel Team
Email marketing strategist · · 9 min read
On this page
- Start broad, then add segments that earn their place
- Engagement tiers: who gets which sends
- Purchase stage and VIPs: RFM in practice
- Ready-to-build Klaviyo segments
- How customers move between segments, and what it predicts
- Send the same segments to paid social
- Customer questions worth asking every month
- How to set up behavior-based segments in Klaviyo
- Frequently asked
A list is everyone who ever gave you an email address. A segment is a group of people who did something: bought twice, clicked last week, went quiet after a first order. Sending them all the same email throws away what you know about each one. This chapter shows how to group customers by behavior in Klaviyo, what each group should get, and how to read people moving between groups.
40-60%
of revenue in many stores comes from VIPs, who are only 5-15% of customers
That split is a rule of thumb; yours will differ, but the shape holds almost everywhere. The guide to segments that drive revenue covers the basics and RFM in plain English explains the scoring.
Start broad, then add segments that earn their place
Most campaigns should go to one broad audience: people who can receive marketing email and have shown recent interest. The first split, if you split at all, is buyers and non-buyers. Keep a segment-specific campaign only when it adds total revenue.
Narrow sends flatter the per-recipient numbers: a win-back campaign can beat the whole engaged audience on click rate and still bring in a fraction of the money. The practitioner default is to reach 3-4 sends a week to the broad audience before adding win-back, new-subscriber or VIP campaigns, leaving that targeting to flows until then. Inside a broad send, dynamic product blocks and show/hide blocks keyed to a profile property (pet type, shopping for men's or women's) make one email personal.
Agencies disagree here. Some build buyer and non-buyer versions of every campaign. Others, including one running about 130 brands, send one version and compare results by buyer type afterwards: a second version of each send turns four emails a week into eight to make, and revenue doesn't double. Either way, use the same segment names and definitions in every client account, so reports line up across the book.
Engagement tiers: who gets which sends
Define engagement by what people do, not by opens. Apple Mail Privacy Protection fires automatic opens for a large share of most lists, so a segment built on Opened Email alone is inflated. Use clicks with bot clicks filtered out, site activity and orders in the last 90 days, and clone it at 30 and 60. Active on Site only sees visitors Klaviyo can identify, so a missing visit is weak evidence of lost interest.
Pick the core window from your buying cycle: 30-60 days for consumables like coffee or pet food, 60-90 for durable goods, 90-120 for high-consideration products, up to 180 for furniture. Then give each tier its job:
| Tier | Who | What they get |
|---|---|---|
| Core | Engaged inside the window that fits your buying cycle | Most campaigns, with cadence tested against revenue and complaints |
| Targeted | VIPs, high-intent non-buyers, buyers of a top category | 1-2 dedicated sends a month when there is something real for them |
| Extended | Engaged in the last 120-180 days | Relevant sends, while they add revenue without hurting delivery |
| Triggered | Cart, checkout, win-back and sunset audiences | Flows, coordinated with campaigns so nobody gets too much at once |
Judge an audience by the total revenue and orders it brings, with complaints, bounces and unsubscribes as the guardrail. A smaller segment nearly always shows a better click rate. That alone doesn't make it the better audience.
At the far end sit profiles with no clicks, visits or orders in a long time. Review them after twice your average purchase cycle plus 30 days (150 days if customers buy every 60), send one last re-engagement email, then suppress those who don't respond. Sunsetting at 30-60 days cuts customers who were only between purchases; never sunsetting drags down delivery for everyone. The deliverability chapter covers the inbox side.
Purchase stage and VIPs: RFM in practice
Order count is the simplest behavior segment, and each stage needs a different email:
- Prospects (no orders): brand story, proof and a reason to place the first order. Never send the welcome discount to someone who has already bought.
- First-time buyers (one order): how to use it, a review request and the natural next product.
- Repeat customers (two or more orders): cross-sells, loyalty perks and early access. They need less discounting than one-time buyers.
- VIPs: previews, early access and recognition. Access is the currency; most don't need a discount.
RFM adds recency and spend to that count. Recency is the strongest predictor, then frequency, then money. For a VIP tier, lead with frequency: one large order can make a one-time buyer look like a regular.
Thresholds must come from your own store; copying them from another category is the most common way RFM goes wrong. A borrowed $500 VIP bar takes in nearly every buyer at a sofa store and almost nobody at a candle shop. Set it by how often your category repeats:
| Store type | Examples | One-time buyers | Champion bar |
|---|---|---|---|
| Low repeat | Perfume, jewelry, furniture | 80%+ | 2-3 orders |
| Medium repeat | Fashion, beauty, home goods | 40-80% | 4-5 orders |
| High repeat | Supplements, subscriptions | Under 40% | 6-8 orders |
Check that the bar holds roughly your top 5-15% of customers: not empty, not half. For a spend cut, take your own top 5-10% by total spend. Keep champions out of deep-discount sends built for lapsed customers. And people who spend a lot but rarely (the buyer of a $300 espresso machine) want a monthly note at most; a long gap between their orders is normal, not churn.
Ready-to-build Klaviyo segments
Build these as dynamic segments, not static lists. A segment rechecks its conditions continuously, so a first-time buyer who orders again leaves First-time buyers within minutes; a list goes stale the day you pull it.
| Segment | Definition | What to send them |
|---|---|---|
| Engaged 90 | Clicked an email (bot clicks excluded), active on site or placed an order at least once in the last 90 days | Regular campaigns; clone at 30 and 60 days |
| New, never bought | Profile created in the last 30 days and placed order zero times | Only the welcome flow for 7-14 days, then brand and bestseller content |
| First-time buyers | Placed order exactly once over all time, at least once in the last 60 days and zero times in the last 30 | Cross-sell, review request, refill nudge, or an offer that expires in 14 days. Use 20-40 days for consumables. |
| VIP | Placed order at least 3 times, or historic lifetime value at or above your own top 5-10% cutoff | Early access 24-48 hours ahead, previews, surprise perks |
| At-risk | Has ordered, but no clicks, site visits or orders in the last 60 days | Your best content and new arrivals. No discount yet. |
| Win-back | Ordered in the last 180 days, not in the last 90, and no clicks in the last 60 | A short series that opens with your strongest offer |
| Browse, no buy | Viewed a product in the last 14 days with no checkout or order in that time | Low-pressure category content and reviews |
| Cross-sell | Ordered where item categories contain X, never where they contain Y | Category Y, with proof from other X buyers |
| Sunset candidates | No opens, clicks, site visits or orders in 180 days, and not subscribed within the welcome flow's length plus one day | One last re-engagement email, then review for suppression |
At-risk means email and site activity went quiet; win-back means time since the last order. Escalate value before price: only win-back gets the offer, because a coupon at 60 days of silence teaches people that ignoring you pays. Set the win-back window to 2-3 times your median reorder interval. The 90-180 days above suit a store whose customers reorder every month or two; a mattress brand using them would chase customers who were never lost.
Then save four exclusions for every campaign's Don't send to box: the sunset segment, profiles created in the last 7-14 days (still in the welcome flow), buyers from the last 3-7 days when the send is a promotion, and anyone in a win-back flow. Klaviyo already skips unsubscribes, bounces and spam complaints.
One trap: category conditions only work if your Shopify integration passes categories in the Placed Order event. If a cross-sell segment comes back near empty, check one profile's order event before the logic, and use contains rather than equals.
How customers move between segments, and what it predicts
Sizes tell you where people are; movement tells you where the money is going. Watch three moves each month: subscribers becoming buyers, first-time buyers becoming repeat buyers, and regulars sliding toward lapsed. On the slide, timing matters (a rough rule of thumb for win-back response, at a store with a one-to-two-month buying cycle):
2-4%
come back at 91-180 days since their last order
0.5-2%
come back at 181-365 days
So trigger a flow when a regular buyer first crosses into that window, rather than waiting for a quarterly win-back campaign. To size a move, multiply the people who could move by the extra share you expect the email to move, then by what each extra order leaves after costs. Don't use the whole gap between two segments' values: repeat buyers are worth more partly because they already have more orders.
Two checks catch problems early. First-time buyers usually outnumber repeat buyers by roughly 60/40 to 70/30; past 85% first-time, you have a retention problem, not an email problem. And group customers by the month of their first order: if newer groups produce fewer top customers, look at who your ads bring in before you rewrite a flow. For long buying cycles (furniture, fine jewelry, electronics), don't push frequency; move people toward referrals, accessories and service.
Send the same segments to paid social
Klaviyo syncs lists and segments to Meta, Google and TikTok as custom audiences and keeps them current as people enter and leave. Expect 40-60% of profiles to match. Meta syncs hourly and needs at least 100 profiles in an audience; Google and TikTok can take up to 48 hours. Four jobs are worth setting up:
| Job | Segment to sync | How to use it |
|---|---|---|
| Stop paying for buyers | Bought in the last 14-30 days, VIPs, active subscribers | Exclude them from acquisition ads |
| Retarget intent | Started checkout in the last 7 days with no order, repeat product viewers | Ads for the products they looked at |
| Reach people email can't | Lapsed buyers and customers who unsubscribed from email | New arrivals and bestsellers that aim for a purchase, not a resubscribe |
| Find more of your best | VIPs, or your top 20% by margin or 12-month value | Lookalike seed of 1,000+ profiles: start at 1%, widen to 2-5% once it performs |
Advertising consent is separate from email consent: in the EEA and UK, Google requires it through your consent tool, outside Klaviyo.
Customer questions worth asking every month
Report every campaign by buyer type, one group per person: non-buyers, active one-time buyers, active repeat buyers and lapsed buyers (active meaning an order within 2-3 times your median reorder interval). A blended result hides which angles win first orders and which bring people back. Then ask:
- Which campaigns convert non-buyers? Their angle belongs in the welcome flow.
- Which move one-time buyers to a second order? Reuse them in post-purchase sends.
- Who unsubscribed? Losing a buyer costs far more than losing someone who never bought.
- Did first-order revenue fall while repeat revenue held? Look upstream: one agency traced it to a pop-up that stopped firing on a new landing page.
- Where did your top 20% come from: which channel, first product and first discount?
- Are Active on Site, Viewed Product, Added to Cart, Checkout Started and Placed Order all firing? A missing event means an empty segment and a missing flow.
Your next step: check those five events, build the engaged, first-time, repeat, VIP and win-back segments with windows set from your own reorder interval, and save the four exclusions. The 15 questions to ask your email data turn the reporting into a monthly routine. With segments in place you can decide who gets what across a month, which is the next chapter: how to plan a month of email.
How to set up behavior-based segments in Klaviyo
- 1
Check your tracking
Confirm that Active on Site, Viewed Product, Added to Cart, Checkout Started and Placed Order all fire for a test visit and order.
- 2
Build engagement tiers
Create Engaged 30, 60 and 90 from bot-filtered clicks, site activity and orders, then pick the core window from your buying cycle.
- 3
Build purchase-stage segments
Create segments for no orders, one order, two or more orders and VIPs, with the VIP bar set from your own customers.
- 4
Set the lapse windows
Build at-risk, win-back and sunset segments, with the win-back window at 2-3 times your median reorder interval.
- 5
Save campaign exclusions
Save the sunset segment, new profiles, recent buyers and win-back flow members as segments to add to every campaign's Don't send to box.
- 6
Report by buyer type
Each month, compare every campaign's results across non-buyers, active one-time buyers, active repeat buyers and lapsed buyers.
Frequently asked questions
How do I build an engaged segment in Klaviyo after Apple Mail Privacy Protection?
Base it on clicks with bot clicks filtered out, site activity and orders in the last 30, 60 or 90 days, not on opens. Automatic opens from Apple Mail inflate any segment built on Opened Email alone.
How do I set a VIP threshold?
Start with order count, set to how often your category repeats, with an order in the last 180 days. For a spend-based cut, sort customers by total spend and take your own top 5-10%, never a borrowed dollar figure.
Should I send separate campaigns to buyers and non-buyers?
Only if the separate version adds total revenue. Many teams send one version with dynamic product blocks and compare the results by buyer type afterwards.
Can I use Klaviyo segments for Facebook and Instagram ads?
Yes. Klaviyo syncs segments to Meta, Google and TikTok as custom audiences that update as people move. Use them to exclude recent buyers, retarget high intent and seed lookalikes from your best customers.
When should I stop emailing unengaged subscribers?
Review profiles with no clicks, visits or orders after about twice your purchase cycle plus 30 days, send one last re-engagement email, then suppress those who don't respond. Cutting at 30-60 days drops customers who were only between purchases.
