Meta Ads Abandoned Cart Retargeting: 3 Audiences After 10.10

Your dashboard this morning probably shows two things at once: strong 10.10 revenue, and far more add to cart events than purchases. Those shoppers are the warmest audience you’ll have before 11.11, and abandoned cart retargeting on Meta Ads is how you reach them. The window to do it well closes in about a week.

Most teams handle it with one ad set called “all visitors 30 days” and frequency left unchecked. Reported ROAS looks great while spend burns on people who already bought.

Below: three custom audiences, a budget split you can defend with arithmetic, and guardrails that switch things off before they leak. If you need the wider campaign architecture first, start with our Meta Ads guide for businesses.

---

Quick Summary

• Three audiences, three levels of intent: add to cart 7 days, product viewers 14 days, video viewers 30 days. Each with layered exclusions so nobody gets chased twice.
• Work out each audience’s ceiling first: Budget = Reach × Target frequency × CPM ÷ 1,000. The percentages fall out of that.
• The cart audience almost always has the highest ROAS and the lowest ceiling. Putting 50–60% of budget there mostly buys frequency.
• Build automated rules on frequency and cost per purchase before launch.
• Reported retargeting ROAS almost always overstates what the ads caused. Measure incrementality before you draw any conclusion for 11.11.

---

Why Post-Sale Retargeting Carries Its Own Risk

In the accounts we run, a large share of carts made in the final two days of a sale are price-driven. Some of that intent disappears once prices return to normal.

The non-discount copy further down separates people who wanted the product from people who wanted the discount. The second problem is mechanical: Meta keeps website visitors for up to 180 days,¹ but only the last seven days of carts matter here, and on a sale-sized budget that audience saturates fast.

A governance note

Quietly extending “10.10 pricing” while the ad says “last day” is misleading. Article 10 of Indonesia’s Consumer Protection Law (No. 8 of 1999) prohibits false or misleading statements about prices and discounts.² Breaching Article 10 carries up to 5 years’ imprisonment or a fine of up to Rp2 billion (Article 62(1)), and for a premium brand the reputational damage arrives sooner than any sanction. Set the end date in the ad and in Ads Manager on the same day.

---

Meta Ads Abandoned Cart Retargeting: The 3 Audiences and Their Exclusions

All three are built in Ads Manager under Audiences > Create audience > Custom audience. Audiences A and B use the Website source (events from the Pixel and Conversions API);³ audience C uses the Video source.⁴ Confirm that AddToCart, ViewContent and Purchase are firing correctly and that Pixel and CAPI events are deduplicated. How to audit event quality is covered in our marketing analytics and attribution guide.

Audience	Source and rule	Retention	Exclude
A. Cart	Website, AddToCart event	7 days	Purchase 30 days, customer list of last-90-day buyers (re-upload weekly)
B. Product viewers	Website, ViewContent event	14 days	AddToCart 7 days, Purchase 30 days, customer list
C. Video viewers	Video, people who watched at least 75% of your 10.10 campaign videos (use ThruPlay instead if the videos are under 15 seconds)	30 days	All website visitors 30 days, Purchase 30 days, customer list

The naming convention we use, so the account is still auditable six months from now:

RT_A_ATC_7D_1010
RT_B_VC_14D_1010
RT_C_VID75_30D_1010
EXCL_PURCHASE_30D
EXCL_CUSTLIST_90D

For audience C, select only the videos you ran for 10.10.

On the customer list: Meta hashes the file before matching,⁵ but that doesn’t remove your obligations as data controller. Indonesia’s Personal Data Protection Law (No. 27 of 2022) requires a lawful basis for processing and transparency toward the data subject.⁶ Make sure your privacy policy and consent flow mention advertising use, and document it before legal asks.

Ad set settings that decide the outcome

Create one campaign with the Sales objective. At campaign level, after choosing the Sales objective, switch the Advantage+ toggle off and leave Advantage campaign budget off. Otherwise ad set budgets and original audience options won’t appear.

Then build three ad sets, one per audience. Standard cart abandoner retargeting setups use one ad set for everyone; splitting by intent is what lets you cap each audience. In each ad set:

1. Set budget at the ad set level so each audience’s ceiling is locked. Campaign budget drifts toward the widest audience.
2. In the Audience section, switch to original audience options and untick Advantage custom audience. With Advantage+ audience on, custom audiences become suggestions; only audience controls such as location, minimum age and exclusions are binding.⁷
3. Add the exclusion audiences from the table above in the Exclude field.
4. Compare the estimated audience size in B and C with and without the exclusions; it should drop. Show audience overlap under Audiences only sizes the raw overlap, it doesn’t prove the exclusions work.

---

The Budget Split Formula: Frequency Ceilings First

Every audience has a spend ceiling set by its size; above it, extra budget only buys repeat impressions.

Weekly ceiling per audience
= (Audience size × % reachable) × Target 7-day frequency × CPM ÷ 1,000

Split % = Audience ceiling ÷ Sum of all three ceilings

Recalculate ceilings every 2–3 days; audience A shrinks daily after 10.10.

Reachable percentage and CPM must come from your own account. For CPM, use your retargeting CPM from the 30 days before the promo; peak-sale CPMs aren’t representative.

The higher the intent, the more frequency is still reasonable:

Audience	Target frequency / 7 days	Stop threshold (with CPA condition)
A. Cart	4–6	8
B. Product viewers	2–3	4.5
C. Video viewers	1.5–2.5	3.5

The stop rule also requires cost per purchase above your ceiling; see Guardrails. These are starting points from accounts we manage. Meta doesn’t publish an ideal frequency for conversion campaigns.

If your budget doesn’t match the total ceiling

• Below the total ceiling: fill A to its target frequency first, then B, then C. Spend in intent order; a pro-rata split starves A of frequency.
• Above the total ceiling: don’t push the surplus into these ad sets. Move it to prospecting or hold it for 11.11.

In our accounts (not a published benchmark), this usually lands at 20–30% A, 35–45% B and 30–40% C.

---

Worked Example: A Premium Leather Bag Brand in Jakarta

A local leather bag brand with an AOV of Rp850,000 and 55% gross margin. Break-even ROAS is 1 ÷ 0.55 = 1.82x. Audience sizes after exclusions, as of 11 October:

	A. Cart	B. Product viewers	C. Video
Audience size	18,000	95,000	220,000
% reachable	70%	60%	50%
Reach	12,600	57,000	110,000
Target frequency	5	3	2
CPM (account assumption)	Rp80,000	Rp55,000	Rp40,000
Weekly ceiling	Rp5,040,000	Rp9,405,000	Rp8,800,000
Daily budget	Rp720,000	Rp1,344,000	Rp1,257,000
Split	21.7%	40.5%	37.9%

Total: roughly Rp23.2 million a week, with the cart audience getting about a fifth. If the team insists on 50% (Rp11.6 million) for audience A, frequency climbs to about 11.5 a week at the same CPM. That’s more than 1.6 impressions a day to the same person, every day for a week.

Projected results, using this brand’s previous retargeting conversion rates:

	A	B	C	Total
CVR per reached person	0.6%	0.12%	0.04%
Orders	76	68	44	188
Revenue	Rp64.6m	Rp57.8m	Rp37.4m	Rp159.8m
Reported ROAS	12.8x	6.1x	4.3x	6.9x

None of this is corrected for incrementality. Many of A’s 76 orders would probably have happened without any ad. Suppose a holdout test shows only 35% of A’s orders are truly incremental: incremental revenue of Rp22.6 million, an iROAS of 4.5x. Still comfortably above the 1.82x break-even, but a very different story from 12.8x.

Note that the free shipping in A’s copy cuts margin. Assuming average shipping of Rp35,000 per order, A’s contribution margin drops to Rp432,500 (break-even ROAS 1.97x), still well below the 4.5x iROAS.

The 35% is illustrative. Get your own number with Conversion Lift under Experiments if your account has access, or a manual holdout. Audience A is Pixel-based, so Meta won’t let you export it or split it at random; the data has to come from your store:

1. Export abandoned-checkout contacts (email/phone) for 4–10 October, the same 7 days covered by RT_A, from your store platform (Shopify, WooCommerce or your own checkout).
2. Split them at random in a spreadsheet into 85–90% test and 10–15% holdout.
3. Upload only the holdout group as a customer list and exclude it from RT_A (the test group is reached through RT_A as normal).
4. After 7 days, compare purchase rates between the two groups from your order system, not Meta reporting.

This only covers carts where you captured contact details. Comparing attribution columns only shows how much credit depends on view-through; it doesn’t measure incrementality.

To read ROAS, CAC and payback consistently across channels, use the framework in our CAC, LTV and ROAS metrics guide.

---

Guardrails: Automated Rules and Attribution

Guardrails go in before the campaign goes live. In Ads Manager, open Rules > Create rule and set up three rules per ad set.⁸ Rules can run as often as every 30 minutes.

RULE 1  RT_A_freq_stop
Apply to   : Ad set RT_A_ATC_7D_1010
Condition  : Frequency > 8
         AND Cost per purchase > Rp150,000
Time range : Last 7 days
Action     : Turn off ad set + notify account owner

RULE 2  RT_A_freq_warn
Apply to   : Ad set RT_A_ATC_7D_1010
Condition  : Frequency > 6
Time range : Last 7 days
Action     : Send notification only

RULE 3  RT_A_freq_nosale
Apply to   : Ad set RT_A_ATC_7D_1010
Condition  : Frequency > 8
         AND Purchases < 1
         AND Amount spent > Rp1,000,000
Time range : Last 7 days
Action     : Turn off ad set + notify account owner

Rule 3 closes a gap in rule 1: with zero purchases there is no cost per purchase, so rule 1 never fires.

Repeat for B (warn at 3, stop at 4.5) and C (warn at 2.5, stop at 3.5), with a cost threshold derived from your own economics. For the leather bag brand: break-even CPA = Rp850,000 × 0.55 = Rp467,500. We set the stop at Rp150,000 because A’s projected CPA is about Rp66,000, and anything above roughly 2.3x that on a saturated audience is waste. By the same logic, B stops at about Rp280,000 (projected CPA Rp138,000) and C at about Rp400,000 (projected CPA Rp200,000), both still under the Rp467,500 break-even CPA. The double condition matters: high frequency that still produces cheap purchases doesn’t need an automatic kill.

On attribution: the current default for website conversion campaigns is 7-day click-through, 1-day engage-through and 1-day view-through, and since the March 2026 change only link clicks count as click-through.⁹ In reporting, use Compare attribution settings and look at each audience on a 7-day click column alongside the default. If audience C’s ROAS drops sharply on click-only, most of its credit comes from view-through, which is a strong reason to run a holdout on that audience.

---

Ad Copy per Audience

The first rule: don’t write as if you’ve been watching someone. Lines like “We saw you left this bag behind” read as surveillance on a premium brand. Write about the bag: the leather, the stitching, the warranty. Leave the shopper’s browsing history out of it.

The second rule: don’t rerun the 10.10 discount. For a premium brand, the post-sale incentive should be a service: shipping, warranty, exchanges, installments.

If you run a catalog, test audience A as an Advantage+ catalog ad so the exact bag in the cart appears. Keep the primary text below, and let the catalog fill in product name and price. The trade-off: catalog retargeting defines its own audience inside the ad set, so your A exclusions and naming need to be rebuilt there.

Audience A, cart 7 days (goal: remove the last obstacle)

Primary text:
Free shipping anywhere in Indonesia until Tuesday, 13 October,
23:59 WIB. Every bag ships with a 2-year stitching warranty.

Headline: Your bag ships free until Tuesday
CTA: Shop now

Audience A after 13 October (evergreen)

Primary text:
Every bag is hand-stitched and comes with a 2-year stitching
warranty. If a seam fails, we repair it free, return shipping
included.

Headline: 2-year stitching warranty, free repairs
CTA: Shop now

Audience B, product viewers 14 days (goal: answer the doubt)

Primary text:
Full-grain cowhide from a tannery in Garut, hand-stitched
in our workshop in South Jakarta. Changed your mind? Swap it
for another model within 14 days, no charge.

Headline: Free model exchange for 14 days
CTA: Learn more

Audience C, video viewers 30 days (goal: proof of quality)

Primary text:
Every bag takes 11 hours to make by hand. This one has done
two years of daily commutes.

Headline: See a bag after 2 years of daily use
CTA: Learn more

Honor the date in A’s copy: if the shipping offer ends on 13 October, turn that ad off at 23:59 that day.

If your landing page is slow or checkout is long, check the friction first with the checklist in our landing page CRO guide.

---

When This Setup Doesn’t Fit

If audience A is under a few thousand people, a standalone ad set will struggle to deliver; merge A and B into one “ATC + VC 14 days” ad set. If more than half your sales close on WhatsApp, a Pixel-based purchaser exclusion will leak, so send purchases from your order system through the Conversions API or upload a daily buyer list as a customer list. If your 10.10 was small, one retargeting ad set with a purchaser exclusion is enough.

---

Schedule: 11–24 October

Date	Action
11 Oct	Build audiences and exclusions, set rules, launch 3 ad sets
13 Oct, 23:59	Turn off the shipping copy in audience A, swap in the evergreen warranty copy
14–15 Oct	Read frequency and the 7-day click column; recalculate ceilings, move A’s unused ceiling to B if A is saturated
17–18 Oct	Audience A no longer contains 10.10 carts: remove copy mentioning 10.10, rename to RT_A_ATC_7D_EVERGREEN, recalculate the ceiling
18–24 Oct	Freeze winning creatives, record iROAS per audience as the 11.11 baseline

---

How Eranya Digital Can Help

In the accounts we audit, the problem is usually tracking (Purchase counted twice, WhatsApp buyers missing from the exclusion) and reporting that doesn’t separate reported ROAS from incremental ROAS. If you want a short audit of your retargeting and tracking setup before 11.11, book a strategy session with our team.

---

References

---

<script type="application/ld+json">
[
  {
    "@context": "https://schema.org",
    "@type": "Article",
    "headline": "Meta Ads Abandoned Cart Retargeting: 3 Audiences After 10.10",
    "description": "Meta Ads abandoned cart retargeting after 10.10: 3 custom audiences with exclusions, a frequency-based budget split, automated rules and ad copy per audience.",
    "image": "https://eranya.digital/images/blog/kenapa-pelanggan-kabur.webp",
    "author": {"@type": "Organization", "name": "Eranya Digital", "url": "https://eranya.digital"},
    "publisher": {"@type": "Organization", "name": "PT Eranya Digital Nusantara", "logo": {"@type": "ImageObject", "url": "https://eranya.digital/images/logo.png"}},
    "datePublished": "2026-10-11",
    "dateModified": "2026-10-11",
    "mainEntityOfPage": {"@type": "WebPage", "@id": "https://eranya.digital/blog/meta-ads-cart-abandoner-retargeting/"},
    "inLanguage": "en-US",
    "keywords": ["meta ads abandoned cart retargeting", "meta ads cart abandoner retargeting", "facebook abandoned cart ads", "add to cart custom audience meta", "retargeting budget split", "exclude purchasers meta ads"],
    "articleSection": "Strategi"
  },
  {
    "@context": "https://schema.org",
    "@type": "FAQPage",
    "mainEntity": [
      {"@type": "Question", "name": "How long should an abandoned cart retargeting audience be on Meta Ads?", "acceptedAnswer": {"@type": "Answer", "text": "For abandoned cart retargeting, seven days is a sensible starting point, because purchase intent decays quickly. Product viewers can run at 14 days and video viewers at 30. Meta allows website custom audience retention up to 180 days and engagement custom audiences up to 365, but windows that long suit nurturing. If your 7-day audience is too small to deliver steadily, stretch it to 10–14 days."}},
      {"@type": "Question", "name": "Should I use Advantage+ audience or manual targeting for retargeting?", "acceptedAnswer": {"@type": "Answer", "text": "For retargeting you want to control per audience, a manual Sales campaign with ad set budgets is easier to audit. At campaign level, after choosing the Sales objective, switch the Advantage+ toggle off and leave Advantage campaign budget off; otherwise ad set budgets and original audience options won't appear. Then switch to original audience options and untick Advantage custom audience. Under Advantage+ audience, custom audiences are only suggestions; only audience controls such as location, minimum age and excluded custom audiences are binding."}},
      {"@type": "Question", "name": "What ad frequency is too high for Meta retargeting?", "acceptedAnswer": {"@type": "Answer", "text": "Meta doesn't publish an official number, so set your own ceiling per audience over a 7-day window. Starting points: 4–6 for add to cart, 2–3 for product viewers, 1.5–2.5 for video viewers. Set an automated rule that turns off the ad set or notifies you when frequency passes the ceiling and cost per purchase is above target."}},
      {"@type": "Question", "name": "How do I exclude recent purchasers from Meta retargeting ads?", "acceptedAnswer": {"@type": "Answer", "text": "Build a website custom audience from the Purchase event (for example, 30 days) and a customer list of recent buyers, then add both to the Exclude field in every retargeting ad set. Layer exclusions too: the product viewer audience excludes add to cart, and the video audience excludes all website visitors from the last 30 days, so one person isn't chased by three ad sets."}}
    ]
  }
]
</script>

Footnotes

1. Meta Business Help Center. About Website Custom Audiences. facebook.com/business/help/610516375684216. Website custom audience retention of up to 180 days. ↩

2. Law of the Republic of Indonesia No. 8 of 1999 on Consumer Protection, Article 10. peraturan.bpk.go.id/Details/45288/uu-no-8-tahun-1999 ↩

3. Meta Business Help Center. Create a Custom Audience from Website Events. facebook.com/business/help/666509013483225 ↩

4. Meta Business Help Center. Create a video engagement custom audience. facebook.com/business/help/1099865760056389 ↩

5. Meta Business Help Center. Create a customer list Custom Audience. facebook.com/business/help/170456843145568 ↩

6. Law of the Republic of Indonesia No. 27 of 2022 on Personal Data Protection. peraturan.bpk.go.id/Details/229798/uu-no-27-tahun-2022 ↩

7. Meta for Business. Advantage+ Audience. facebook.com/business/ads/meta-advantage-plus/audience ↩

8. Meta Business Help Center. About Automated Rules in Meta Ads Manager and Pause Ad Set If Your Frequency Is Greater Than a Certain Number. facebook.com/business/help/1694779440789213, facebook.com/business/help/265593457478252 ↩

9. Meta Business Help Center. About attribution settings. facebook.com/business/help/458681590974355; Meta for Business, Simplifying Ad Measurement for a Social-First World, facebook.com/business/news/click-attribution (Meta’s announcement of the March 2026 change). Secondary source: Jon Loomer Digital, How Meta Ads Attribution Works in 2026, jonloomer.com/meta-ads-attribution-2026, covering the change from engaged-view to engage-through and the narrower click-through definition. ↩