Email and Retention · focused work
Abandoned cart email flows
An abandoned cart flow writes to someone who added a product and left without paying. It is usually the highest-earning automation a store runs, and it is also the one most likely to be credited with orders that were arriving anyway. The work is building it so it recovers sales, and then knowing how much of what it reports is genuinely extra.
This is the right work if
- Checkout abandonment is the only trigger running and browsing and cart activity go unused
- The first message sends the following day, long after the shopper decided somewhere else
- Every message in the sequence carries a discount, so regulars have learned to abandon deliberately
- One shopper receives an email, a text and a display ad about the same cart
- The flow has never been run against a holdout, so its reported revenue has never been checked
What it is
What the flow actually contains
Three different behaviours are usually collapsed into one flow. Browsing without adding anything is mild interest. Adding to a cart and leaving is a decision deferred. Reaching checkout, entering an email and stopping is an interruption or an objection at the last step. They deserve different timing, different content and different frequency, so the build separates them and gives each its own trigger conditions, exclusions and message set.
What decides performance is less about copy than about three things. Identification, because a store can only write to visitors it recognises, which makes sign-up placement and early email capture at checkout part of this project rather than a separate one. Timing, taken from the store's own distribution of how long people take between last cart activity and an order. And honesty in measurement, because the reporting inside an email platform credits the flow whenever a recipient opens or clicks and later buys, including the many who would have returned unprompted. A holdout group is the only way to separate recovered revenue from revenue that was already on its way.
How it is done
The work, in order
What changes
- Three levels of intent receive three sequences instead of one message for everybody
- Messages land inside the window where the decision is actually being made
- Incentives follow a written policy rather than being added by reflex
- The revenue claimed for the flow can be checked against a holdout group
Solve identification first
The flow can only reach people the store recognises. Sign-up placement, email capture at the earliest checkout step, and identity that persists across devices set the ceiling on everything else. Doubling the number of recognised visitors is usually worth more than rewriting any message.
Split the three triggers apart
Build browsing, cart and checkout abandonment as separate flows with their own entry rules and exclusions. A browser who looked once should not receive the sequence written for someone who typed their address and stopped at payment.
Take the timing from real orders
Look at how long buyers who did complete took between their last cart activity and the order, then place the first message inside the window where most of them decided. For many stores that is far sooner than the default the platform ships with.
Give each message a different job
The first is a reminder showing the item. The second answers the practical objections of delivery time, returns and payment options. The third offers help or a comparable alternative. Whether any of them carries an incentive is a business decision made once and applied consistently.
Coordinate channels and suppress properly
Exclude anyone who has purchased, who is already in another sequence, or who received the text version, and cap how often the same person can re-enter within a period. Frequent abandoners exist and they should not receive the sequence weekly.
Keep a holdout and use it
Hold back a random share of eligible people, compare their purchase rate against those who received the flow, and use that gap when reporting what the flow earns. Re-run it after any material change, since a rewrite can move the number in either direction.
Platform notes
Shopify
Email entered at the first checkout step is captured before payment, which is what makes checkout abandonment recoverable, and that data is exposed to connected email tools along with the cart contents and the restore link.
WooCommerce
Cart contents are not tied to an identified visitor by default, so recovering carts depends on a capture step or a plugin that records the address earlier, and the restore link needs testing against the store's session handling.
Does adding more messages to the sequence keep earning?
Returns fall off quickly, and each extra message costs some unsubscribes and some goodwill. Add one, measure with the holdout, and keep it only if the recovered orders outweigh the list damage. The right length differs by price point, since expensive considered purchases justify a longer conversation.
Should the first message include an incentive?
Usually not. The first message reaches people who were interrupted, and many of them return without any discount, so leading with one gives margin away for nothing. If an incentive is used at all, put it late in the sequence and keep the policy stable so shoppers cannot learn to wait for it.
What should happen with repeat abandoners?
Cap re-entry so the same person cannot receive the sequence every week, and treat frequent abandonment as a signal worth investigating. It often points at a shipping cost, a payment method or a stock problem that shows up in the cart, which is a store fix rather than an email one.
Can anything be done for shoppers who never gave an address?
Not by email, which is the honest answer and the reason identification matters. Those visitors can only be reached through paid remarketing, if consent allows it, or through an on-site prompt when they return. Growing the recognised share of traffic is the lever that changes this.
Related work and answers
Find the leak.
A free Growth Analysis ranks what your store should fix first, by revenue at stake.