Klaviyo Winback Flow: Step-by-Step Setup

Short answer. Build a standard Klaviyo winback flow with Placed Order as the metric trigger, an initial delay slightly longer than the customer's normal buying cycle, and the profile filter Placed Order zero times since starting this flow. Use up to three emails, stop promotional steps after a new order, and branch one-time, repeat, or high-value customers only when the data supports different treatment.

A winback flow is for prior customers whose expected next purchase did not happen. It is not a generic re-engagement sequence for anyone who stopped opening email. The implementation starts with purchase data and timing, then uses content to answer one question: why should this customer return now?

This guide focuses on the Klaviyo build. For the policy that connects prevention, winback, sunset, and suppression, read Winback and sunset flows.

The correct Klaviyo winback architecture

Trigger: Placed Order
  -> Initial delay: slightly longer than normal repurchase cycle
  -> Profile filter: Placed Order zero times since starting this flow
  -> Optional customer-value or order-count split
  -> Email 1: relevant return reason
  -> Delay
  -> Email 2: proof, selection, or offer
  -> Delay
  -> Email 3: final clear choice
  -> No purchase: evaluate separately for sunset eligibility

This follows Klaviyo's current winback flow guide, updated October 2025 and verified July 16, 2026. Klaviyo recommends keeping the flow to three emails per recipient.

1. Calculate the delay from real order history

Do not begin with 90 days because it is common in templates. Measure when customers normally buy again.

A practical analysis

For customers with at least two completed orders:

1. Calculate the days between order one and order two.
2. Repeat for later consecutive orders.
3. Separate results by first product, category, subscription status, or customer group where useful.
4. Review the median and distribution, not only the mean.
5. Choose a delay slightly longer than the healthy repeat interval.

If the median repeat interval is 45 days and most repeat orders occur by day 60, testing the first winback message after that range is more defensible than waiting for a universal 90-day threshold.

Separate replenishment when appropriate

Klaviyo recommends a distinct replenishment flow for products with a predictable cycle. A reminder that a consumable may be running low is different from a winback message to a broadly lapsed customer. Do not make one flow perform both jobs.

For long-cycle products, consider whether another purchase is even the right outcome. Education, accessories, service, referrals, or category expansion may be more relevant than replacing the original item.

2. Verify the purchase metric

Open Analytics > Metrics and inspect Placed Order from the ecommerce integration you intend to use.

Check:

• the integration source;
• event volume and recent timestamps;
• order value and currency;
• item or category data needed for branches;
• cancellations, refunds, and test orders;
• profile identity across store and Klaviyo;
• any second order source such as POS, marketplace, or Buy with Prime.

If orders arrive from separate metrics, one flow may not cover the full customer relationship. Klaviyo specifically documents separate handling when Amazon Buy with Prime produces a distinct order event. Map every purchase source before relying on one profile filter.

3. Create the flow and choose Placed Order

In Klaviyo:

1. Go to Flows.
2. Choose Create flow.
3. Start from Klaviyo's winback template or build your own flow.
4. Select Metric > Placed Order as the trigger.
5. Confirm it is the ecommerce platform's actual order metric.
6. Add the required profile filter.

Use this filter:

What someone has done or not done
Placed Order
zero times
since starting this flow

The trigger records the original purchase. The delay moves the profile toward the winback window. The profile filter prevents a customer who buys again during that wait from receiving the next promotional action.

Klaviyo rechecks profile filters before each actionable step, not just at initial entry. A failed profile filter causes that action to be skipped, although the profile may continue to be evaluated at later steps. Review the exact behavior in Klaviyo's flow trigger and filter documentation.

Check re-entry

Metric-triggered flows allow re-entry by default in Klaviyo's current entry model. That makes sense because every new purchase creates a new potential buying cycle. The purchase filter should prevent the prior cycle from continuing after a subsequent order.

Still review re-entry and overlapping order sources. A duplicate Placed Order event can schedule duplicate paths. Test one real order end to end rather than assuming the integration deduplicates every edge case.

4. Add the initial buying-cycle delay

Place a time delay immediately after the trigger. Set it slightly longer than the selected repurchase interval.

Examples are only illustrations:

Do not branch into dozens of categories before the sample size supports it. Start with a global interval, document the compromise, and split the highest-volume category when the data shows a meaningful difference.

5. Branch customers only when the message changes

Useful branches include:

• one-time buyer versus repeat buyer;
• former VIP or high-value customer versus standard customer;
• product family with a distinct repeat cycle;
• recent site visitor versus no recent activity;
• customer with an available personalized product feed versus fallback products.

Avoid a branch based only on “opened the last email.” Apple Mail Privacy Protection can create automatic opens. If open behavior is needed, filter or report on the Apple Privacy Open property and combine it with clicks, site behavior, or purchase. Klaviyo explains this in its MPP guide.

Example branch plan

Every branch adds production and QA cost. If the body, timing, and offer remain identical, the branch is not doing useful work.

6. Build a three-email sequence

Email 1: make returning relevant

The first message should be the lightest touchpoint.

• Acknowledge the time since the relationship was active without guilt.
• Show a small number of best sellers, new releases, or category-relevant products.
• Explain a real improvement, new use case, or customer benefit.
• Give one primary CTA.

Klaviyo's official pattern includes an incentive as an option. It does not prove that every brand needs one. Test whether product relevance and service can work before making perpetual discounts the default.

Email 2: add proof or a controlled offer

After an appropriate delay, build on the first message:

• different products or a tighter category selection;
• customer reviews or useful proof;
• a time-limited offer when margin and positioning support it;
• a personalized product feed with a safe fallback.

Klaviyo product feeds can fall back to a secondary recommender when there is not enough profile data. New recommendation models and very recent events may take time to influence results. Preview several profiles and read the current product feed documentation.

Email 3: close the attempt

The last email should make the next choice clear:

• return before a genuine deadline;
• update preferences;
• unsubscribe;
• take no action and leave the winback path.

Do not threaten to delete customer data or claim an account will close unless that is actually the policy. Winback is a marketing sequence, not a data-retention notice.

7. Configure pressure and exclusions

Review each flow message for Smart Sending. Klaviyo's current default email Smart Sending window is 16 hours, although settings and message-level behavior can differ. A skipped message is not automatically rescheduled.

For winback, consider:

• other campaign or flow emails in the same period;
• active checkout, cart, or browse recovery;
• post-purchase messages after a new order;
• SMS or push sent under separate channel windows;
• repeated winback attempts after later purchases.

Do not exclude someone from winback merely because they are in an abandonment flow segment if the stronger flow already has a short lifetime. Instead, define which message wins and how long the exclusion should last. Permanent broad exclusions can hide customers from relevant lifecycle messaging.

See Klaviyo's Smart Sending guide for current channel behavior.

8. Test content and flow logic

Previewing an email proves only part of the implementation. Test the trigger, delay, filter, branch, content, and exit.

Required QA profiles

Use a QA copy with shortened delays, then restore production timing and have another person review it. Put messages in Manual status when you need to inspect real qualifying profiles before automatic delivery. Check Recipient Activity and skip reasons before moving each action to Live.

Klaviyo's preview and test guide notes that some dynamic coupons, preference links, and tracking behavior do not function like a live send in a simple preview. Complete a controlled live-path test where required.

9. Back-populate carefully

A newly created metric-triggered winback flow normally handles orders that occur after activation. Existing customers whose qualifying purchase happened months ago may require Add past profiles or a controlled segment campaign, depending on the setup.

Before adding past profiles:

1. Calculate how many profiles will be scheduled at each step.
2. Confirm current marketing eligibility.
3. Recheck recent orders from every source.
4. Review Smart Sending and other active flows.
5. Use Manual status or a small sample first.
6. Document whether past profiles should receive the full sequence or only a current message.

Do not back-populate an entire historical customer base simply because the button is available.

10. Test one improvement at a time

Klaviyo supports A/B testing for individual flow emails. You can test subject line, discount, or content and choose or automatically determine a winner based on the configured metric.

Start with a business question:

• Does an incentive create enough incremental margin to justify its cost?
• Does category personalization increase qualified clicks?
• Does a shorter sequence convert as well with fewer complaints?
• Does a different initial delay match the actual buying cycle better?

Use one variable per test. For purchase outcomes, select the relevant conversion metric and allow the measurement window to complete. Klaviyo's flow email A/B test guide explains the current setup.

Open rate can be a diagnostic, but it should not be the only win condition. Track purchase conversion, revenue per recipient, margin after discount, unsubscribes, complaints, and longer-term repeat behavior.

11. Hand non-buyers to sunset carefully

Completing winback without purchasing does not automatically mean a profile should be suppressed. Separate these outcomes:

Build the sunset audience from sustained multi-signal inactivity, not from “did not buy in winback” alone. Then follow the Klaviyo sunset flow implementation.

Launch checklist

• The Placed Order metric comes from the correct integration.
• The initial delay is based on observed repurchase behavior.
• Placed Order zero times since starting this flow is a profile filter.
• All purchase sources can stop or exclude the flow where necessary.
• Three emails or fewer are planned for each recipient.
• One-time, repeat, and VIP branches have distinct content.
• Dynamic product blocks have tested fallbacks.
• Smart Sending and cross-channel pressure are documented.
• Manual QA covers a mid-flow purchase.
• Back-population volume and eligibility are reviewed.
• Conversion, margin, unsubscribe, and complaint metrics have owners.

FAQ

What trigger should a Klaviyo winback flow use?

For the standard ecommerce pattern, use the actual Placed Order metric from the store integration. Add a delay slightly longer than the normal buying cycle and the profile filter Placed Order zero times since starting this flow.

Should I use a date property for Klaviyo winback?

Not for the standard flow described by Klaviyo. Its official setup uses the purchase metric plus a delay. A date- or segment-based design can serve a specialized use case, but it creates different entry, re-entry, and maintenance behavior that must be tested.

How many emails should a Klaviyo winback flow include?

Klaviyo currently recommends keeping the flow to three emails per recipient. A smaller sequence can be better when the audience is very inactive or already receives frequent campaigns. Test message count against purchases, margin, complaints, and unsubscribes.

Should the first winback email include a discount?

It can, but it is not mandatory. Test a relevant new-product or service message against an incentive. Evaluate incremental margin, not just attributed revenue. Former VIPs or premium customers may respond better to access, curation, or service than a broad percentage discount.

Why did a customer skip the second winback email?

Inspect Recipient Activity. A new Placed Order may have caused the profile filter to fail, Smart Sending may have skipped the message, the profile may no longer be eligible for marketing, or another filter may not match. A purchase-related skip is usually correct behavior.

Can the same customer enter winback again?

Metric-triggered flows can allow re-entry after later purchases. Review the flow's re-entry settings and duplicate order sources. A new purchase should begin a new buying cycle, while the purchase filter prevents the old cycle from continuing.

Build winback around the buying cycle

The reliable setup is simple: verified purchase event, defensible delay, purchase filter, relevant content, and controlled handoff. Deliver can analyze repurchase intervals, build the Klaviyo flow, QA past-profile entry, and design the matching sunset policy.

Review your Klaviyo winback flow with Deliver.

Continue with:

• Design the complete winback and sunset strategy
• Build the Klaviyo sunset flow
• Create an ecommerce retention plan

Charlotte Rodrigues, CRM Lead at Deliver. Updated July 16, 2026.