In this guide
An abandoned-checkout email reaches a shopper who entered contact information and began checkout but did not complete the order. That is stronger intent than viewing a product or adding an item to a browser cart, but it is not proof that the shopper wants three reminders and a discount.
The sequence must restore the exact checkout, stop immediately after purchase, respect consent and channel rules, and address plausible friction without inventing urgency. This guide builds the trigger, exclusions, message ladder, testing plan, and failure diagnosis inside the wider ecommerce email lifecycle system.
Fast summary
- Browse, cart, and checkout abandonment are different events and should not trigger overlapping reminders.
- Purchase exclusion must be re-evaluated before every message and tested across order-source edge cases.
- Each message should restore, resolve, or help decide rather than repeat the same reminder.
- Use discounts selectively and evaluate contribution margin, repeat behavior, and incremental lift.
- Treat recurring recovery objections as evidence for storefront and checkout improvements.
Recommended platform
Some links are affiliate links. We may earn a commission at no extra cost to you. Disclosure
Separate browse, cart, and checkout abandonment
Browse abandonment begins with an identified visitor viewing a product. Cart abandonment begins when an item is added. Checkout abandonment begins after the checkout process and customer identity are captured. These are distinct intent levels and can fire close together. The latest event should normally suppress the earlier reminder.
Shopify's current marketing-automation documentation exposes separate automations for abandoned browse, cart, and checkout. Its abandoned-checkout documentation defines a checkout as abandoned after it remains incomplete for more than ten minutes after email information was provided, while also documenting cases where recovery email is not sent.
Do not call every recovery message a cart email when the trigger is Started Checkout. The distinction affects the restoration URL, personalization data, consent context, likely objection, exclusion event, and reporting denominator.
| Stage | Known intent | Best message job |
|---|---|---|
| Product browse | Viewed one or more products | Answer a product question or help compare |
| Cart | Selected an item and quantity | Restore considered products and surface cart-level concerns |
| Checkout | Entered checkout and provided identity | Restore checkout and resolve payment, shipping, trust, or timing friction |
| Payment failure | Attempted payment with an error | Provide accurate support and retry guidance rather than generic promotion |
| Purchase | Order completed | Stop recovery and move to post-purchase communication |
Swipe horizontally to compare every column.
Higher-intent events should replace lower-intent reminders, not stack on top of them.
Make trigger and exclusion logic the first deliverable
Use Started Checkout as the entry event when the platform integration supports it. Add a flow filter that excludes anyone who places an order after entering. Re-evaluate the exclusion before every message. A person can purchase during the delay, from another device, through a different campaign, or after contacting support.
Klaviyo's abandoned-cart flow guidance emphasizes recent-purchase exclusion and notes integration-specific edge cases. Shopify draft orders can create checkout events without a corresponding purchase event in some Klaviyo workflows, so test draft, test, subscription, marketplace, and Buy with Prime behavior if those order sources exist.
Exclude unavailable carts, unsupported shipping destinations, unresolved payment captures, internal test accounts, fraud-review states, and profiles without the permission required for the intended marketing send. If an external payment gateway captured money but failed to report it to Shopify, a recovery email can be especially damaging. Payment-event and support procedures belong in the QA plan.
- 1
Trigger a test checkout
Enter the correct customer identity and stop at several checkout stages.
- 2
Inspect event payload
Verify items, variants, quantities, price, currency, discount, URL, source, and timestamp.
- 3
Complete after entry
Purchase during each delay and confirm every remaining recovery message is skipped.
- 4
Test failure states
Use unavailable inventory, payment decline, shipping restriction, draft order, and cancellation scenarios.
- 5
Verify permission
Test subscribed, unsubscribed, suppressed, and jurisdiction-relevant profiles.
- 6
Inspect restoration
Open the email on another device and confirm the link restores a valid current checkout state.
Watch out
The most important abandoned-checkout test is not whether the email sends. It is whether every later email stops after an order is placed.
Use a message ladder that adds new value
The first message should make return easy: identify the store, show the considered items accurately, link to the recoverable checkout, and provide a support path. It can arrive while intent is fresh, but the exact delay should reflect purchase consideration, product complexity, and event reliability.
A second message should address the most likely unresolved question, not repeat the first with louder urgency. Use observed support and checkout data to choose shipping, returns, fit, compatibility, payment methods, guarantee terms, or product proof. A third message is justified only when it adds a distinct reason and frequency remains appropriate.
Do not claim inventory is reserved, a cart expires, or a discount ends unless the systems enforce that statement. Klaviyo's flow guidance recommends urgency only when it is applicable. Manufactured countdowns can increase short-term pressure while weakening trust and complaint performance.
| Message | Primary job | What it must not do |
|---|---|---|
| 1: Restore | Return to the exact checkout and offer help | Invent a problem or lead with an automatic discount |
| 2: Resolve | Answer the most likely decision blocker | Repeat identical product tiles and copy |
| 3: Decide | State a truthful final reason, expiry, or support option | Use false scarcity or continue after purchase |
Swipe horizontally to compare every column.
Every additional message must earn its place with a different customer job.
Diagnose the checkout objection before writing copy
Review checkout funnel steps, payment errors, shipping-zone failures, customer messages, onsite searches, returns, and session behavior where consent permits. If shipping cost appears late, fix checkout disclosure and product-page expectations before writing a clever email. Recovery should not become a recurring patch for preventable storefront friction.
Segment only when the difference changes the message. First-time shoppers may need trust and policy clarity. Returning customers may need convenience. High-consideration products may need compatibility or consultation. International checkouts may need duties and delivery context. Do not create branches for every profile attribute if the content remains identical.
Use direct copy: identify the item, explain the next action, and state support availability. Include delivery and return summaries only when current and market-appropriate. The recovery link should be the primary action; links to generic social pages or unrelated collections distract from the task.
- Shipping cost or delivery estimate shown later than expected.
- Payment decline, authentication failure, or gateway confusion.
- Unclear returns, exchange, warranty, or cancellation terms.
- Fit, compatibility, ingredient, material, or care uncertainty.
- Discount code failure or unexpected promotion conditions.
- Inventory loss or variant changes during checkout.
- Trust concern about the store, fulfillment origin, or support access.
Test discounts as economics, not as a default second email
A recovery discount can convert price-sensitive shoppers and train repeat visitors to abandon checkout. Before offering one, calculate contribution after product cost, fulfillment, payment fees, expected returns, and the discount. Exclude products or customers where the economics do not work.
Test no offer against a targeted offer using mutually exclusive branches and a meaningful sample. Compare incremental completed orders, margin, average order value, repeat purchase, return rate, and future full-price behavior. Platform-attributed revenue alone can over-credit high-intent messages.
Consider alternatives that solve the actual objection: transparent shipping, payment support, a sizing answer, extended exchange clarity, a bundle that improves value, or a free-shipping threshold. If the store needs a discount to recover most checkouts, inspect pricing, acquisition promises, and checkout friction upstream.
| Offer strategy | Potential advantage | Risk to measure |
|---|---|---|
| No discount | Protects margin and reveals natural recovery | May not move truly price-sensitive shoppers |
| Conditional free shipping | Addresses a common checkout surprise | Threshold and shipping economics |
| Targeted percentage | Can resolve price friction for selected profiles | Habit formation, margin, and coupon leakage |
| Fixed amount | Clear value on suitable basket sizes | Excessive discount on low-value carts |
| Support or consultation | Solves fit and compatibility uncertainty | Operational response capacity |
Swipe horizontally to compare every column.
Evaluate recovered contribution and customer behavior, not only coupon revenue.
QA the sequence and measure incremental recovery
Test on mobile and desktop with single-item, multi-item, discounted, multi-currency, unavailable, and long-title checkouts. Verify image, variant, quantity, price, currency, checkout URL, policy links, unsubscribe, reply path, dark mode, and plain-text content. Confirm that the restored checkout recalculates current price and stock honestly.
Measure eligible checkouts, messages delivered, clicks, completed orders, time to purchase, margin, discount use, unsubscribe, complaints, and support contacts. Separate people who clicked from those merely attributed after an open. Apple's privacy behavior makes open-based attribution particularly weak.
Where volume permits, hold out a random eligible group from one or more messages. Compare completed orders and contribution within the same observation window. A flow that claims substantial revenue but produces little lift against holdout may be capturing natural return behavior rather than causing it.
- 1
Verify eligibility
Count who entered, who should not enter, and why profiles were skipped.
- 2
Verify message truth
Check product, variant, price, currency, stock, discount, timing, and urgency.
- 3
Verify exits
Purchase, unsubscribe, suppression, inventory, and support states must stop or change later messages.
- 4
Separate attribution
Report click-through recovery separately from open-influenced platform attribution.
- 5
Test incrementality
Use holdouts or controlled branches when the eligible volume can support a useful comparison.
- 6
Repair upstream causes
Feed recurring shipping, payment, policy, and product objections back into the storefront.
Frequently asked questions
Start with one reliable restoration email. Add a second only when it addresses a distinct objection, and a third only when it has a truthful purpose and acceptable frequency. Every message must stop after purchase.