Retargeting for Shopify Stores: The Meta + Google Funnel That Actually Recovers Sales

Retargeting is the easiest campaign to make look good and the easiest to get wrong. Show ads to people who were about to buy anyway and the ROAS is spectacular; the incremental revenue is close to zero. Done properly — segmented by intent, with different creative per stage and honest exclusions — it is a reliable 10–20% of revenue at 3–6× blended return. This is the structure.

Segment by intent, not by "visited site"

Exclude purchasers (last 7–30 days depending on repurchase cycle) from every segment above, and exclude each higher-intent segment from the lower ones so a checkout abandoner does not see three different ads.

Meta: what to run

• Dynamic product ads (catalogue) for checkout and cart abandoners. Overlay template with price and a subtle "still available" cue. Let Meta pick the product; they saw it.
• Static and video for product viewers and engagers — this is where reviews, UGC and comparisons work.
• Frequency cap: 3–4 impressions per week per person. Above that, CPA rises and brand sentiment falls. Meta's own reporting shows frequency; watch it weekly.
• Placement: all placements is fine for DPA; for video, prioritise Reels and Stories.
• If you run Advantage+ Shopping, keep a separate retargeting campaign at 10–20% of Meta budget so you control the above — see the ASC guide.

Google: what to run

• Performance Max already retargets site visitors with your feed. Add your customer lists and GA4 audiences as signals; do not build a separate Display remarketing campaign that competes with it.
• Search RLSA / audience bid adjustments: +20–40% on cart abandoners searching your brand or category terms.
• YouTube for engagers and product viewers with a 15s product demo — cheap reach, decent assist value.
• Demand Gen for lapsed customers with new-arrival carousels.

Email and SMS are retargeting too

Klaviyo abandoned checkout and browse abandonment flows reach the same people for a fraction of the cost. Coordinate: if the email flow is firing, the paid retargeting for that person can start 24 hours later, not immediately. Klaviyo → Meta audience sync lets you exclude email-recovered carts from paid. The Klaviyo flows guide covers the sequences.

Creative rules per stage

1. Checkout abandoners: plain product, price, "free UK delivery", one review line. No discount in the first 72 hours — many will convert on reassurance alone. A modest offer (free shipping or 10%) after day 3 to non-converters only.
2. Cart and product viewers: social proof heavy. Star ratings, review quotes, "2,000+ five-star reviews", UGC.
3. Engagers: the brand story or the hero product's core benefit. Treat as prospecting with a warmer audience.
4. Lapsed customers: "time to restock?" for consumables; "new this season" for fashion and home.

Budget split and expectations

Retargeting should be 15–25% of paid social budget and mostly automatic within PMax on Google. If retargeting is over 40% of spend, the account is harvesting rather than growing — prospecting is underfunded. Expect 3–6× ROAS on properly excluded retargeting; anything above 10× usually means exclusions are missing and you are paying to reach existing customers.

Measuring whether it works

• Holdout test: exclude 10–20% of a segment (random split by Meta's Conversion Lift or a manual customer-list split) for 4 weeks. Compare conversion rates. If the held-out group converts nearly as well, your retargeting is not incremental.
• MER trend: total revenue ÷ total spend, weekly. Retargeting that "works" in-platform but does not move MER is not working.
• New vs returning customer split from Shopify, weekly.

Get the tracking right before any of this: retargeting audiences depend on events firing correctly. The server-side tracking guide is the prerequisite.

Good retargeting is defined by who it excludes. Every audience you build, ask: who in here was going to buy anyway?

Frequently asked questions

How long should retargeting windows be?

Short for high intent (1–7 days for checkout abandoners), longer for low intent (up to 30 days for engagers). Match windows to your purchase cycle — consumables are shorter, furniture longer.

What frequency cap should I use for retargeting?

Around 3–4 impressions per person per week on Meta. Higher frequency raises cost per acquisition and irritates prospects; the ads become wallpaper.

Should I offer a discount in retargeting ads?

Not immediately. Reassurance (delivery, reviews, returns) converts many abandoners without cost. Introduce a modest offer after 3 days to those who still have not bought, and exclude anyone who already purchased.

Is retargeting still effective after iOS privacy changes?

Yes, but audiences are smaller and attribution is noisier. Conversions API, first-party lists from Klaviyo/Shopify, and holdout testing are now essential to run it well.