· 9 min read
Abandoned cart emails: what actually works
Cart abandonment is the most-run lifecycle program in ecommerce, which means every ESP ships a template, every vendor sells a best-practice blog post, and most teams run the default with modest tweaks. The default works — barely — because the program has high inherent signal. A user with an item in their cart is telling you something. The question is whether your program reads that signal or ignores it.
Justin Williames
Founder, Orbit · 10+ years in lifecycle marketing
The three-message structure that consistently wins
1 hr
Typical time-to-first-send that outperforms longer delays.
24 hr
Second send. Enough time for natural abandon-and-return patterns to complete.
72 hr
Third send. After this, the cart is cold — the user moved on.
Email 1 (1 hour post-abandon):Gentle reminder. Cart contents, product photos, straightforward resume link. No discount. Most carts recovered here weren't abandonments in the spiritual sense — the user got distracted, the checkout timed out, their toddler needed attention. The reminder itself is enough.
Email 2 (24 hours):The user didn't come back. Add context: an FAQ or social proof specific to the items in the cart. Still no discount. The purpose is addressing a reason for hesitation — return policy, sizing, shipping speed — not discounting past it.
Email 3 (72 hours):Last attempt. This is the message where a discount (if you use one) shows up. Keep it modest (10–15%), time-limited, and tied to the specific cart so it doesn't train the user to abandon future carts strategically.
The discount decision most teams get backwards
A discount in email 1 teaches users to abandon carts on purpose. A discount in email 3 captures price-sensitive users without reprogramming your base.
The instinct is to discount early and heavy to rescue the sale. The operator view is the opposite: early discounts train the base to abandon carts as a coupon-extraction tactic. Every cart-save analysis I've run shows the same pattern — programs that discount in email 1 have higher abandonment rates 90 days later than programs that reserve discounts for email 3 only.
The price-testing guide covers why this training effect is real and how to measure it. For cart rescue specifically: no discount email 1, maybe a supporting-content email 2, discount only in email 3 if you use one at all.
Channel mix beyond email
Email is the slowest-reactive channel in cart rescue. A cart abandoned at 8pm doesn't benefit from an email scheduled for 9am the next day. Push notifications (where opted in) close the timing gap — a 30-minute-post-abandon push has higher recovery rate than an hour-one email, because the user is still in the mental state that led to the abandonment.
SMS fits the rescue pattern well but has higher consent and cost barriers. Worth running for high-value carts (above a threshold) where the send economics work. The Multi-Channel Orchestration skill handles the channel selection and frequency governance to prevent a user from receiving email + push + SMS for the same abandoned cart within an hour.
The push specifics are covered in the push notification copy guide. Cart-rescue push is one of the highest-performing push use cases — treat it with the design care it deserves.
What to measure beyond recovery rate
Recovery rate (% of abandoned carts that convert via the rescue sequence) is the obvious metric. Two metrics that matter more:
Incremental recovery rate: % of carts that converted because of the rescue, not carts that would have converted anyway. Measure via a holdout — a random 5–10% of abandoners who receive no rescue sequence at all. Compare their natural return rate to the sequence-targeted rate. The delta is your true program lift.
Trained-abandonment rate:are users abandoning more often over time? If yes, the discount policy is probably too generous. Segment abandonment rate by user cohort (users who've received the rescue sequence multiple times vs first-timers). If rescued users have higher abandonment over time, you're training the behaviour.
,
When not to run a cart rescue at all
Three contexts where cart rescue doesn't pay off:
Subscription products with low AOV. The cost of the send + the discount usually exceeds the incremental revenue. Focus on the first-purchase path instead.
Products where abandonment is often a shopping behaviour (bookmarking items, comparison shopping across sessions). Users who abandon to come back later don't need a rescue; they need a favorited-items feature.
Products where the rescue cost per send exceeds 5% of AOV. At that point, the program is a marketing expense masquerading as recovery revenue.
Frequently asked questions
- How many cart abandonment emails should I send?
- Three is the operator consensus. Email 1 at ~1 hour, email 2 at 24 hours, email 3 at 72 hours. More emails produce diminishing recovery and increased unsubscribes. Less than three misses the price-sensitive segment captured by the final message.
- When should I include a discount?
- Email 3 only, if at all. Early discounts train users to abandon carts as a coupon-extraction tactic. Reserve discounts for the last-attempt message, keep them modest (10–15%), and tie them to the specific cart so they're not reusable.
- What's a good cart recovery rate?
- 10–15% is typical; 15–20% is strong. But raw recovery rate misleads without an incrementality measurement. A holdout group showing 7% natural return rate and a sequence showing 14% means the program is contributing 7 percentage points, not 14.
- Should I send push notifications as well as email?
- Yes, where push consent exists. A 30-minute-post-abandon push often outperforms a 1-hour email because the user is still in the mental state that produced the abandonment. Coordinate across channels so the user doesn't get email + push + SMS for the same cart within an hour.
- How do I measure if cart rescue is training users to abandon?
- Segment abandonment rate by user cohort. Users who've received rescue sequences multiple times vs first-time abandoners. If rescued users have higher abandonment rates over time, the discount policy is too generous and is reshaping the base's behaviour.
- Should I include reviews or social proof?
- In email 2, yes — reviews specific to the items in the cart. Generic social proof ('customers love us') is weak by this point in the sequence. Item-specific reviews answer a real question the abandoner might have had.
This guide is backed by an Orbit skill
Related guides
Lifecycle marketing for flat products
Lifecycle playbooks assume a product with clear stages and a steady engagement cadence. Most real products aren't like that. This guide is about designing lifecycle programs for products with flat, infrequent, or irregular usage — where the standard textbook doesn't apply.
The cadence question: how often should you email?
Everyone asks how often to email and almost nobody answers it properly. This guide is about what 'cadence' actually means, why the right number isn't a single number, and the five inputs that settle the debate for your specific program.
Personalisation that doesn't feel creepy
There's a line between personalisation that earns a user's trust and personalisation that breaks it. This guide is about where the line actually is, how lifecycle programs cross it without noticing, and the specific patterns that keep you on the right side.