Why retargeting works (and where it doesn’t)
Retargeting targets people who already know the brand. They’ve visited the site, viewed a product, added to cart, or bought before. The conversion rate is much higher than cold prospecting, usually 5-10×.
But high conversion rate doesn’t mean high incrementality. Many retargeted users would have converted anyway. A cart abandoner who comes back via a retargeting ad might have come back via organic search the next day. The retargeting ad got the credit. The conversion would have happened either way.
This is why retargeting often shows great attributed ROAS and disappointing incremental ROAS. The first number is real. The second is reality.
Common retargeting audiences
- Site visitors, anyone who visited any page within N days (typically 7-30)
- Product viewers, viewed a PDP but didn’t purchase
- Cart abandoners, added to cart but didn’t check out
- Checkout abandoners, started checkout but didn’t complete
- Past customers, for upsell, cross-sell, retention campaigns
- Email subscribers, engaged with email but didn’t recently purchase
- Custom segments, high-value visitors, returning visitors, etc.
How privacy changes hit retargeting
Retargeting was hit hardest by browser privacy changes. The audience definitions depend on persistent identifiers (cookies, device IDs), exactly what ITP and ETP target.
Practical impact:
- Cookie-based audience pools shrunk 30-50% over the past five years
- Cross-device retargeting broke without persistent identity stitching
- iOS 14.5+ App Tracking Transparency further restricted mobile retargeting
- Third-party cookie deprecation in Chrome will further compress the pool
Modern retargeting compensates with first-party identity (hashed emails matched to ad-platform Custom Audiences and Customer Match) and server-side audience syncing.
How to measure retargeting fairly
Three patterns:
- Incrementality testing. Hold out a portion of the retargeting audience. Compare conversion rate against the exposed group. The lift is the true incremental ROAS.
- Separate prospecting and retargeting ROAS reporting. Don’t blend. The two have different economics and need different reads.
- MMM for channel-level allocation. Top-down attribution caps retargeting overspend by tying it to incremental revenue rather than attributed conversions.
Common mistakes
- Optimising spend toward retargeting. Easy because it looks efficient. Bad because incremental ROAS is often disappointing.
- Retargeting without consent. GDPR requires consent for marketing tracking. Retargeting on non-consenting audiences is a violation.
- Letting retargeting cannibalise organic returning traffic. Customers who would have come back organically now arrive via paid retargeting instead. Same revenue, more ad spend.
FAQ about Retargeting
What is retargeting?
Retargeting is paid-media targeting aimed at users who have already engaged with the brand, site visitors, cart abandoners, past customers. Conversion rate is much higher than cold prospecting, often 5-10×.
Is retargeting still effective in 2026?
Yes, but the audience pool has shrunk dramatically as ITP, ETP, and ad blockers stripped first-party cookie persistence. Modern retargeting relies on first-party identifiers (hashed email Custom Audiences, Customer Match) rather than cookie pools.
Why does retargeting have high attributed ROAS but low incremental ROAS?
Because many retargeted users would have converted anyway. They visited the site, intent was already there, the retargeting ad got the credit but did not cause the conversion. Incrementality testing reveals the gap.