Klaviyo segmentation gets underused in a specific way: most accounts lean almost entirely on lists and a handful of basic filters, when the platform is built to support dynamic, behaviour-based segments that update in real time. A Klaviyo segmentation guide worth following starts by separating what Klaviyo actually offers from the narrower subset most accounts default to.

Lists vs Segments in Klaviyo

Lists in Klaviyo are static groups, built by explicit sign-up (a form, an import) and changed only when someone is manually added or removed. Segments are dynamic: built from conditions against properties, events, and predictive analytics, and membership updates automatically as a customer's behaviour changes. A customer can fall out of a segment the moment they stop meeting its conditions, with no manual work required.

This distinction is the single most important thing to get right in Klaviyo segmentation, because using a list where a segment is needed, or the reverse, produces messaging that is either stale or missing the dynamic behaviour that makes Klaviyo's targeting worth using in the first place. Sign-up and consent management belongs on lists. Anything based on behaviour, purchase history, or engagement belongs on segments.

What Klaviyo Segments Can Actually Be Built On

Klaviyo segmentation conditions fall into a few categories, and combining them is where the real precision comes from. Profile properties cover static or slow-changing attributes like location or acquisition source. Metrics, Klaviyo's term for tracked events like Placed Order, Viewed Product, or a custom event pushed from your backend, are what make behavioural segmentation possible, filtered by how many times an event occurred and within what window.

Predictive analytics, available on accounts with enough purchase history, adds Klaviyo-calculated fields like predicted customer lifetime value, expected date of next order, and churn risk, which can be used as segment conditions directly without building the underlying model yourself. Combining a metric condition (has not placed an order in 45 days) with a predictive field (historically high CLV) produces a segment far more precise than either alone, and this combination is usually where the highest-value win-back audiences live.

Building an RFM-Style Segmentation Model in Klaviyo

Step 1: Define your core segments before opening the segment builder. Decide on the segment names that matter for the business, commonly something like Champions, Active, At Risk, and Lapsed, based on the same logic behind any RFM analysis, before translating them into Klaviyo's condition builder.

Step 2: Build recency and frequency conditions off Placed Order. A typical Active segment condition combines "Placed Order at least once in the last 60 days" with a frequency threshold appropriate to the product's purchase cycle. Keep the time windows consistent across your named segments so a customer's position is unambiguous.

Step 3: Layer in engagement, not just purchase behaviour. For accounts where email or SMS engagement matters as a leading indicator, combine purchase-based conditions with an engagement metric, such as opened or clicked in the last 30 days, to catch customers whose behaviour is shifting before it shows up in purchase data.

Step 4: Use exclusion conditions to keep segments mutually distinct. Without explicit exclusions, a customer can technically qualify for both an Active and an At Risk segment depending on how the conditions are written. Build segments so a given customer lands in exactly one lifecycle-stage segment at a time.

Step 5: Connect segments to flows, not just campaigns. A segment used only for occasional manual campaign sends captures a fraction of its value. The same segments should gate entry and exit conditions on flows, so a customer's lifecycle stage actually changes what they are automatically sent, not just what they are manually included in.

Segment Templates Worth Building First

A handful of segment types cover most of the value in a typical Klaviyo account, and building these first gives the rest of the lifecycle programme something real to trigger on.

VIP or Champions. Top decile by predicted lifetime value or actual spend, combined with recent activity. This segment typically gets early access, better support, and is excluded from generic discount campaigns that would otherwise train high-value customers to wait for a markdown.

Browse abandoners. Viewed a product multiple times in a window without placing an order. Browse abandonment converts at a lower rate than cart abandonment but still represents intent worth a single, low-pressure follow-up rather than being ignored entirely.

Recently active, never purchased. Engaged with email or the site repeatedly but has not converted. This segment often responds better to a friction-removal message, addressing a specific objection, than to a discount, since price is rarely the only barrier for someone who is already paying attention.

Win-back by recency band. Rather than one dormant segment, split lapsed customers by how long they have been inactive, 30 to 60 days, 60 to 90, 90-plus, since the right offer and tone differ meaningfully across those bands, and treating them identically undersells the closer band and overspends on the group least likely to return.

Segmentation Mistakes Specific to Klaviyo Accounts

Rebuilding the same logic as a list instead of a segment. This is the most common mistake in accounts migrating from a simpler platform, where list-based thinking gets carried over and static exports get used where a dynamic segment would update automatically.

Ignoring predictive analytics because the underlying model feels like a black box. The predicted fields are calculated from the account's own purchase history and improve as more data accumulates. Accounts with enough order volume are usually leaving real precision on the table by not using them at all, in the same way segmented sends consistently outperform broadcast campaigns once the underlying data actually supports the split.

Building segments that never get audited. A segment built eighteen months ago, before a pricing change or a new product line, may no longer represent the group it was named for. Segmentation, like any retention model, needs revisiting against current data, not treated as permanent once built.

Too many overlapping segments with no shared naming convention. Accounts that have been active for years often accumulate dozens of segments built by different people for one-off campaigns, with no consistent structure. This makes it difficult to know which segment is the current source of truth for a given lifecycle stage.

Klaviyo's segmentation engine is capable of far more precision than most accounts use. The platform is rarely the constraint. Most of the gap comes down to whether the account has a defined lifecycle model translated into segment logic, or whether segments were built ad hoc, one campaign at a time, with no shared structure underneath them. Auditing the current segment list against the templates above is usually a faster way to find the gap than starting from a blank canvas, since most accounts already have the raw data needed; what is missing is the structure connecting it to what actually gets sent.

Frequently Asked Questions

What is the difference between a list and a segment in Klaviyo?

A list is a static group that changes only through manual sign-up, import, or removal. A segment is dynamic, built from conditions against properties, events, and predictive fields, and membership updates automatically as customer behaviour changes.

Can Klaviyo segments be based on purchase behaviour?

Yes. Segments can be built on metrics like Placed Order, Viewed Product, or custom events, filtered by frequency and time window, which is what makes behaviour-based and RFM-style segmentation possible inside Klaviyo without external tools.

What are Klaviyo's predictive analytics fields used for?

Predictive fields like predicted lifetime value, expected date of next order, and churn risk are calculated automatically from an account's purchase history and can be used directly as segment conditions, adding precision that would otherwise require building a model outside the platform.

How many segments should a Klaviyo account maintain?

Enough to cover the core lifecycle stages that drive different messaging, commonly four to six, rather than a large number of narrow, overlapping segments built for individual one-off campaigns that are hard to maintain and audit over time.

Should Klaviyo segments be used for flows or just campaigns?

Both, but flows are where the ongoing value comes from. A segment used only for occasional manual campaigns captures a fraction of what it is capable of compared to using the same segment logic to gate entry and exit conditions on always-on flows.