Retargeting is where most South African e-commerce stores leave money on the table. You spend on cold traffic, get the click, and then let 97% of those visitors walk without ever seeing your brand again. Here's how to build a Meta Ads retargeting setup in 2026 that actually recovers that traffic and turns it into revenue.
- Install Meta Pixel plus Conversions API together, not one or the other, or you lose 15-30% of purchase events to iOS tracking loss.
- Split retargeting into three temperature tiers: viewed product, added to cart, past purchasers. Treat each with a different bid and creative.
- Cap frequency at 3-4 per week per audience or cost per result climbs past your ROAS floor within 10 days.
- Dynamic product ads pulling from your catalog feed outperform static retargeting creative on click-through rate in most South African e-commerce accounts.
Why this matters
Most South African stores set up a pixel once, build one "website visitors" audience, and never touch it again. That's a broken structure feeding Meta's algorithm mixed signals: someone who bounced off your homepage in two seconds gets treated the same as someone who abandoned a R2,400 cart. Smart Bidding can't optimise for revenue if every visitor looks identical to it.
The fix isn't more budget. It's structure: clean event data, segmented audiences by intent, and frequency limits that stop you burning through the same 3,000 warm shoppers with the same ad for six weeks straight.
What you'll need
- Meta Business Suite access with admin rights to your ad account and pixel
- Your product catalog set up as a feed (Shopify, WooCommerce, or a manual CSV upload works)
- At least 100 purchase events in the last 30 days for Meta to build a usable Purchase custom audience (below that, lean on Add to Cart and View Content events instead)
- A POPIA-compliant cookie consent banner already live on your site, since pixel tracking of South African visitors requires it
- 20-30 minutes to configure events, and another hour to build ad sets once the pixel has been live for a few days
If you're running Meta Ads for an e-commerce store already but retargeting has never had its own campaign, that's the gap this guide closes.
The steps
1. Install the Meta Pixel and Conversions API together
This is non-negotiable in 2026. Apple's iOS tracking restrictions and browser-level ad blockers mean the browser pixel alone misses a meaningful chunk of purchase events, often 15-30% depending on your traffic mix. Conversions API sends the same events server-side, so Meta gets a second, more reliable copy of the data.
In Events Manager, set up both under the same pixel ID and use Meta's Event Match Quality score to check the overlap. Anything below 6.0 out of 10 means your customer data (email, phone, address) isn't being hashed and matched properly, and you should fix that before building any audience on top of it.
Common mistake: installing the pixel via a generic plugin and never verifying which events actually fire. Test every event (ViewContent, AddToCart, Purchase) with Meta's Pixel Helper browser extension before you spend a single Rand on retargeting.
2. Build audiences by intent, not just "all visitors"
A single "website visitors, last 30 days" audience wastes budget on people who left in two seconds and never scrolled. Split into at least three tiers:
- Viewed product, no cart (last 14 days): lowest intent, lowest bid
- Added to cart, no purchase (last 7 days): highest-value retargeting audience, deserves the biggest share of budget
- Past purchasers (last 90-180 days): for cross-sell and repeat purchase, not the same creative as the other two tiers
Each tier gets its own ad set. Mixing them into one audience means Meta's bidding averages out the cost per result across very different buyer intents, and you end up overpaying for the cold end and underserving the hot end.
3. Set up dynamic product ads from your catalog feed
Static retargeting creative (one image, one message) stalls fast because it doesn't match what the person actually looked at. Dynamic Product Ads (DPA) pull the exact product a visitor viewed or added to cart straight from your feed and rebuild the creative per person.
Upload your catalog in Commerce Manager, map it to your pixel, and build a DPA ad set targeting your Add to Cart audience. In most South African e-commerce accounts this outperforms static retargeting on click-through rate, because the ad literally shows the R899 dress or the R1,200 blender the person was already looking at.
Common mistake: letting the catalog feed go stale. If a product is out of stock or the price changed and the feed isn't syncing daily, you're retargeting people with ads for items they can't buy.
4. Structure campaigns by audience temperature
Run three separate campaigns rather than three ad sets under one campaign objective. This gives you cleaner budget control and campaign-level reporting on which temperature actually drives revenue.
- Campaign 1: Add to Cart audience, DPA creative, higher daily budget
- Campaign 2: Viewed Product audience, DPA creative, lower budget
- Campaign 3: Past Purchasers, static creative with a cross-sell or restock message
This structure also protects you if one audience underperforms. You can pause Campaign 3 without touching your highest-converting cart abandoners.
5. Set frequency caps before you launch, not after
Retargeting pools are small (often under 5,000 people), so the same audience sees your ad repeatedly within days. Cap frequency at 3-4 impressions per week per person in Ads Manager's delivery settings. Past that, cost per result climbs fast, usually within 7-10 days, as people start actively scrolling past or hiding the ad.
Exclude anyone who already purchased from your Add to Cart and Viewed Product audiences. Nothing kills trust faster than showing someone an ad for the exact jacket they bought last week.
6. Layer in creative rotation and a real offer test
Retargeting creative fatigues faster than prospecting creative because the audience is smaller and sees it more often. Rotate in a new angle every 10-14 days: a review-led ad, a founder message, an urgency angle (limited stock, seasonal deadline).
Test one variable at a time. If you're testing a discount code against a free-shipping message, run them as separate ad sets within the same audience so you can actually attribute the lift to the offer, not to audience differences.
Common mistake: running the same creative for two months because "it's still getting clicks". Click-through rate can hold while conversion rate quietly drops as the audience gets fatigued on the message. Check cost per purchase weekly, not just CTR.
7. Set a ROAS floor and stick to it
Decide your minimum acceptable return on ad spend before you launch, based on your margin, not on what looks impressive in a screenshot. A 3:1 ROAS on a product with 25% margin barely breaks even once you account for returns and payment processing fees.
If a retargeting ad set drops below your floor for five consecutive days, pause it and diagnose before scaling anything else. Don't average a bad ad set's performance against a good one and call the campaign healthy.
8. Monitor and prune weekly
Retargeting audiences shrink and refresh constantly as customers exit the 7-14 day windows. Check audience size weekly. If your Add to Cart audience drops under 1,000 people, Meta's delivery gets erratic and you're better off widening the window to 14 days than trying to force spend into a shrinking pool.
Troubleshooting
Pixel isn't firing on Purchase events. Check for a checkout redirect or a third-party payment gateway (PayFast, Ozow, Yoco) that takes shoppers off-domain before the confirmation page loads. The pixel needs to fire on the page the customer actually lands on after payment, not before.
Audience too small to serve ads. Meta generally needs at least 1,000 people in a custom audience for stable delivery. If your Add to Cart pool is under that, extend the lookback window from 7 to 14 or 30 days rather than force-feeding a tiny audience.
Frequency climbing past 5 within a week. Your audience is too small for your budget. Either lower daily spend on that ad set or widen the audience window before cost per result spikes.
Catalog ads showing wrong prices or out-of-stock items. Your feed sync is delayed. Set your platform (Shopify, WooCommerce) to push catalog updates at least daily, ideally hourly during sales periods.
Overlapping audiences competing against each other. If your Viewed Product and Add to Cart audiences overlap heavily, you're bidding against yourself in auction. Use Audience Overlap in Ads Manager to check, and exclude the higher-intent audience from the lower-intent campaign.
iOS visitors missing from reported conversions. This is expected post-iOS14. Conversions API closes some of the gap but not all of it. Use Meta's own attribution window reporting alongside your store's actual sales data rather than trusting Ads Manager numbers in isolation.
Tools and resources
- Meta Events Manager, for pixel setup, event testing, and match quality scoring
- Meta Commerce Manager, for catalog feed upload and DPA setup
- Meta Pixel Helper (browser extension), to verify events fire correctly before spending
- Your store platform's native Meta integration (Shopify's Meta channel, for example) for automatic catalog syncing
- If your fashion or apparel catalog needs vertical-specific creative guidance, see Meta Ads for fashion brands in South Africa
- Beauty and skincare stores with high repeat-purchase cycles should look at Meta Ads for beauty brands in South Africa for restock-timing angles
What to do next
Once retargeting is structured and your ROAS floor is holding, the next lever is usually ad copy, not audience. Weak retargeting copy relies on the discount alone instead of addressing why someone didn't buy the first time. If your Google Ads product listings need the same treatment, how to write Google Ads copy for e-commerce covers the same intent-matching logic applied to search.
If you're building this out for the first time and want a second set of eyes on pixel setup and campaign structure before you scale spend, that's exactly the audit work an agency should be doing, not guesswork.
FAQ
How long does it take for a Meta Pixel to collect enough data for retargeting?
A pixel needs roughly 7 days of active traffic and at least 100 events (View Content or Add to Cart) before Meta can build a stable custom audience. Purchase-based audiences need 100 purchase events in a 30-day window to work reliably in 2026.
What’s the best retargeting window for cart abandoners?
A 7-day window works best for most South African e-commerce stores because purchase intent drops sharply after a week. Extend to 14 days only if your Add to Cart audience is under 1,000 people and delivery is unstable.
Is Conversions API necessary if I already have the pixel installed?
Yes. The browser pixel alone misses 15-30% of purchase events due to iOS tracking restrictions and ad blockers. Conversions API sends the same events server-side and closes most of that gap.
How much should I spend on Meta retargeting versus prospecting?
A common starting split is 20-30% of total Meta Ads budget on retargeting and the rest on prospecting, though the right ratio depends on your average order value and how many warm visitors you generate monthly. Stores with high cart abandonment often shift closer to 35%.
Do dynamic product ads work for small catalogs?
Yes, but the benefit is smaller. DPA shines when you have 50+ SKUs and visitors view different products. With under 20 products, a well-made static retargeting ad often performs just as well.
How often should retargeting creative be refreshed?
Refresh every 10-14 days for small retargeting audiences under 5,000 people. Larger audiences can run the same creative for 3-4 weeks before frequency fatigue sets in.
Can I retarget people who never visited my site?
Not through pixel-based retargeting, since that requires a prior site visit. For cold audiences with no site history, use lookalike audiences built from your Purchase or Add to Cart custom audiences instead.
What ROAS should e-commerce retargeting achieve?
Retargeting campaigns typically deliver a higher ROAS than prospecting because the audience already knows the brand, often 2-3 times the account average. The exact floor depends on your margin, so calculate it from cost of goods before setting a target.
One last thing
The single biggest lever most South African stores miss isn't audience size or creative, it's the exclusion list. If you're not excluding recent purchasers from your Add to Cart and Viewed Product campaigns, you're paying to show ads to people who already gave you their money. Check your exclusions this week before you touch anything else.
Related guides
- Meta Ads for e-commerce stores in South Africa
- SEO agency for e-commerce stores in South Africa
- Best Meta Ads agencies in South Africa
