⏱ 5-minute read
Most brands treat retargeting as “everyone who visited.” That sounds logical on the surface, but you end up spending real budget on people who bounced once, won’t return, and will never pay you back.
The reason this persists is that the campaign numbers often look fine. Consistent ROAS, consistent volume. The people who were already going to convert are carrying the people who never will in your pool.
The underlying problem is identity. Cookie-based retargeting pools are full of fragments: one-time bouncers, in-app browser sessions that get wiped, the same person counted four times across devices.
The fix isn't spending more or less. It's knowing who's actually in-market and building your retargeting around that group as a dedicated closing lane.
Most of Your Retargeting Pool is Noise
If you look at a typical retargeting audience, you'll find something like this: the majority are ghost visitors: one session, low page depth, no real research behavior.
A small minority are true in-market researchers: multiple sessions, product page depth, comparison behavior, checkout intent signals.
The problem is that most stacks can't reliably isolate that second group, because identity is fragmented across sessions and devices. You can't separate tourists from researchers, so you spend like they're equal. They're not.
5 Principles For Retargeting
Retargeting is a closing strategy, not a catch-all net. "Everyone who visited" isn't an intent audience. It's a list.
Identity resolution makes intent measurable. Without it, you can't separate tourists from researchers and your budget allocation is structurally wrong.
In-Market Customers deserve their own lane. Dedicated budget, tighter frequency cap, no discounts. Their job is to close.
BFCM customers belong in a seasonal lane. Year-round, they suppress efficiency and train your platforms on the wrong signals.
Retargeting is a portfolio. Each segment has a job, a message, and a budget rule. Mix them and you compensate with discounts.
What A Clean Retargeting Portfolio Looks Like
Split retargeting into two lanes: In-Market and everyone else. Define In-Market with behavioral signals, 2+ sessions in a short window, deep PDP engagement, comparison behavior, checkout start. Give this group its own ad set, its own budget, and no promotional offers.
Then run the rest as a portfolio:
In-Market → close with proof and reviews, no discount
VIP → early access, replenishment cues, protect margin
Lapsing → timing-based nudges, selective offers
BFCM → activate in sale windows only
Worst customers → exclude
The goal isn't to retarget less. It's to stop retargeting people you can't realistically close, and shift that spend toward audiences that can actually pay back.
Putting This to Work (Easily)
Stop retargeting visits. Start retargeting real intent.
The RX Audience Builder makes this operational. The segments above: In-Market, VIP, Lapsing, are available as ready-to-use audience templates, built on RX Identity Resolution and synced automatically across your ad platforms within 24 hours.
As customers enter or exit a segment, your audiences update. No CSV exports, no manual refreshes. And because RX resolves identity across devices and sessions, match rates run ~30% higher than Klaviyo, meaning more of your segment actually becomes a targetable audience.
The goal isn't to retarget less. It's to stop retargeting people you can't realistically close and concentrate spend where it actually pays back. Reply INMARKET and I'll show you what your In-Market segment looks like in your data today.
