Abandoned cart emails and remarketing ads at a glance
The abandoned cart flows tutorial builds the email side: identify the shopper, trigger when an item is added and no order follows within a set wait, send a short sequence that reminds, reassures and asks once more, exit everyone the moment an order arrives, and measure against a group that is eligible but receives nothing. The flow can only reach shoppers the store knows, logged-in customers and guests who typed an address before leaving, and only those with a recorded basis for marketing. The ecommerce email and SMS flows guide places it among the other automated flows a store runs.
Remarketing ads reach shoppers through an ad platform instead. On Meta, a website Custom AudienceAn ad audience built from the advertiser's own sources: customer lists, site visitors or engagers. is built from the events the pixel and Conversions API send, and Meta's help page on Custom Audiences describes audiences built from website activity, customer lists, catalogue interactions and engagement. The custom audiences guide covers how those audiences are matched and expire, and the catalog ads guide covers the retargeting ladder that shows each person the items they viewed or added to the basket. A shopper does not need to be identified to the store to enter such an audience; they need to send events Meta can tie to an account.
So the two reach overlapping but different groups. Email reaches the identified and permissioned; ads reach the matched and tracked. A store that runs only one of them leaves part of its abandoners unprompted, and a store that runs both without coordination prompts the same shopper twice.
Compare table
Abandoned cart emails vs Remarketing ads
| Row label | Abandoned cart emails | Remarketing ads |
|---|---|---|
| Who can be reached | Shoppers the store has identified, by account or a typed address | Visitors whose events reach the ad platform and can be matched to an account |
| Permission needed | A recorded basis to send marketing email; an address typed at checkout is identity, not consent | Tracking consent where consent rules apply; visitors who decline do not enter website audiences |
| Cost per message | Almost nothing beyond the sending platform | Paid for every impression or click in the auction |
| Timing control | Exact: the store sets the wait, the gaps and the sending window | Loose: the platform delivers within the window when the person is on its apps |
| Content | The exact cart, objections answered, an incentive only where policy allows | Products from the catalogue the person viewed or added, in a template |
| Frequency control | A fixed sequence with a cap and exits checked before each send | A frequency cap and audience window, enforced by the platform |
| What is measured | Opens, clicks and orders from entrants, read against a held-out group | Impressions, clicks and platform-attributed purchases |
| Attribution risk | Takes credit for shoppers who were coming back anyway | The same, plus credit for any purchase after a view within the window |
| Setup effort | Shop events, identity, consent records, templates and exits | Pixel and Conversions API events matched to the catalogue, audiences and a campaign |
| What stops it | An order, an unsubscribe, an emptied cart, the end of the sequence | An order event that reaches the platform, or the person ageing out of the window |
The first two rows decide who belongs in which channel. The third and fourth decide which channel should go first: email is precise and nearly free, so it is the default for every shopper it may reach. Ads earn their place with the shoppers email cannot reach, and with any extra orders a holdout shows they add on top of email.
When to choose abandoned cart emails
Choose the email flow first, for every abandoner the store can identify and has a basis to contact. It reaches the person directly, shows the exact items with a button that restores the basket, answers the obstacles that usually stop a purchase, and stops the moment an order arrives. Each further message costs next to nothing, so the flow's economics depend only on whether it changes behaviour, which a holdout shows.
Its limit is the permission it needs. The tutorial treats cart reminders as marketing email in most readings, so the flow needs a consent basis, the full footer and an unsubscribe link. The list building and consent guide explains, from the UK regulator's direct marketing guidance, when consent is needed, when existing customers can be contacted about similar products under the soft opt-in, and why every message must offer a way out; other markets have their own rules, and this is general practice, not legal advice. Capturing the address early in checkout, with a clear choice about marketing, is what makes the flow larger.
When to choose remarketing ads
Choose remarketing ads for abandoners email cannot reach: guests who left before typing an address, shoppers who gave an address without a basis for marketing, and returning visitors the store cannot recognise. They are also the route for people who browsed and never added anything, at a lighter weight, and for a store with no email programme at all.
Ads come with costs email does not have. Every impression is paid for, delivery timing belongs to the platform, and an audience full of people already on their way back reports a return the ads did not cause. The comparison of prospecting and remarketing explains why remarketing should be a capped, frequency-controlled layer judged by holdouts, and the diagnosis page for retargeting that gets all the credit shows how a last-touch report hands it sales other work created. The ROAS calculator is useful for the arithmetic once it is run on incremental revenue, as the comparison of iROAS and ROAS explains.
Decision table
| Situation | Choose | Why |
|---|---|---|
| Identified shopper with a basis for marketing email | Email flow first | Precise, timely and nearly free per message |
| Guest who left before typing an address | Remarketing ads | Email has no one to send to |
| Shopper already in the email flow | Email, with ads suppressed for the first part of the sequence | Two prompts at once spend money on the same decision |
| Small store with no email programme yet | Build the email flow; use ads meanwhile | The flow is the cheaper long-term channel |
| Both running and both reporting strong returns | A holdout on each | Each report credits sales that would have happened |
| Browse without a basket | A light email to known contacts, or a low-weight ad rung | Weaker intent calls for fewer, later prompts |
One recovery plan: suppression, timing and a shared holdout
Running both well means treating them as one plan with three rules.
Suppress across channels. A shopper who is identified and in the email flow does not also need a cart ad an hour after leaving. Exclude them from the cart ad audience for the first part of the sequence, or keep them in only where a test shows the ads add orders. The tutorial's advice that paid remarketing should share the purchase exit and frequency logic applies in both directions: an order has to end both, which means the purchase event must reach the ad platform as reliably as it reaches the email platform.
Sequence the timing. Email goes first because it is cheapest and most precise. Ads follow for the shoppers email cannot reach, and if they are used for identified shoppers at all, after the email sequence has run. Short audience windows for basket abandoners and longer ones for viewers keep the ads in proportion to the intent, as the catalog ads guide's ladder sets out.
Hold out together. Each channel's own report credits it with shoppers who were returning anyway. The honest reading randomly withholds a share of abandoners from email, a share from ads and a share from both, using a Holdout GroupA randomly chosen part of an audience kept from exposure to measure what exposure caused. for each, and compares order rates. That shows the IncrementalityConversions caused by marketing. of each channel and of the two together, and whether the ads add anything on top of the email for the shoppers who receive both. The comparison of email and SMS covers the separate choice of whether a text belongs in the cart sequence.
A yarn shop splits cart recovery between email and ads
Quilvexa Yarns, an invented online yarn shop, records 9,000 abandoned baskets above its floor over 90 days. Of these, 4,800 belong to shoppers it can email; it splits them at random into four cells. The 3,600 who get both the email flow and catalogue remarketing order at 13.0%; the 400 who get email only order at 12.5%; the 400 who get ads only at 9.0%; and the 400 who get neither at 8.0%. Email alone adds about 4.5 points, ads alone about 1.0, and ads on top of email about 0.5, around 18 orders across the 3,600, though cells of 400 leave wide intervals. For the 4,200 abandoners it cannot email, ads reach 3,570 and 630 are held back: 6.2% against 5.0%, about 43 extra orders. Meta reports 410 purchases from £2,700 of remarketing spend, £18,860 of revenue at an average order of £46 and a reported return of about 7.0. The holdouts imply about 61 extra orders, £2,806 of revenue, an incremental return of about 1.04, well below the 2.0 a 50% margin needs to break even. The shop keeps the email flow for everyone it can contact, removes them from cart ads, and keeps a smaller, capped ad budget for the shoppers email cannot reach, with the holdout running again next quarter.
Common mistakes
- Running both channels at full weight on the same shopper. Two prompts for one decision double the cost, not the recovery.
- Emailing every typed address. An address in a checkout field is identity, not a basis for marketing.
- Reading either channel's own report as its contribution. Both credit shoppers who were already coming back; hold out a share of each.
- Forgetting the purchase exit on the ads. If the order event does not reach the platform, buyers keep seeing the basket they bought.
- A discount in the ads that the email policy refuses. One incentive policy, read from the customer record, must govern both channels.
- No frequency cap on cart ads. A small audience of warm shoppers is easily shown the same products far too often.
- Treating the two as rivals. They reach different abandoners; the comparison library lists other pairs that are easy to confuse.
