E-commerce Email Automation

Abandoned Cart Sequences

Abandoned cart emails recover revenue that would otherwise be lost. The standard sequence: send the first email 1 hour after abandonment, a second email 24 hours later (often including a small incentive), and a third 72 hours later if no purchase. Adjust timing based on your product's purchase decision timeline.

Abandoned cart email content should show the specific products left in the cart (dynamic product blocks), make returning to purchase frictionless (pre-filled cart link), and address the most common reasons for abandonment — often price, shipping cost, or needing more time to think.

Post-Purchase Flows

The post-purchase period is when customers are most satisfied with their decision and most receptive to further engagement. A well-designed post-purchase flow: confirms the order (transactional), sets expectations for delivery, provides product usage guidance, requests a review at the right moment, and introduces relevant complementary products.

Review request timing matters significantly. Send the request after the customer has had time to actually use the product — not immediately on delivery. For consumables, this might be two weeks; for furniture, it might be a month.

Win-Back Campaigns

Customers who have not purchased in a defined period are lapsing. Win-back campaigns attempt to re-engage them before they become permanently inactive. The sequence typically starts with a re-engagement email (no incentive), moves to a soft incentive if there is no response, and concludes with a strong incentive or a final goodbye email.

The final goodbye email ("We're removing you from our list") has counter-intuitively high engagement — some customers click to stay on the list without purchasing. It also cleans your list of genuinely uninterested subscribers, which improves deliverability metrics.

Sequencing Flows So They Do Not Collide

Each flow works in isolation and the trouble starts when a customer qualifies for several at once. A buyer who abandons a cart, then purchases, then browses again can receive an abandonment reminder, a purchase confirmation, a review request and a browse-abandonment message inside a day.

The fix is a priority order and a global frequency cap applied above the individual flows, not inside them.

A workable priority, highest first: transactional messages always send — order confirmations and shipping updates are expected and exempt from suppression. Post-purchase outranks acquisition, because a message about something already bought is more relevant than one about something not yet bought. Abandonment flows should exit immediately on purchase, which sounds obvious and is the most common defect in a live setup. Win-back and broadcast come last and should be suppressed for anyone active in another flow.

Test the collision cases deliberately: abandon then buy, buy then abandon, buy twice. Each should produce a sequence a person would recognise as sensible.

Consent, Identification and the Unsubscribe That Has to Work

Automation multiplies whatever your consent practice is. A flawed list-building method sends one bad message manually and thousands automatically.

Three obligations carry across jurisdictions even where the detail differs. Consent has to be given for marketing specifically — a checkout that collects an email to deliver an order has not obtained marketing consent by implication, and a pre-ticked box is not consent in regimes that require it to be freely given. Identification means the message says clearly who is sending it, with a genuine postal contact. Unsubscribe must work, must be easy, and must take effect promptly across every flow — not only the one the recipient was reading.

That last point is the one automation breaks. An unsubscribe honoured by the broadcast tool but not by the transactional-adjacent flows produces exactly the complaint that damages a sending domain. Test the unsubscribe from inside each automated flow, not from a newsletter, and confirm the suppression propagates everywhere before the flow goes live.

Sources

What each claim on this page rests on. Entries are typed so you can see which are primary.