A DTC Brand Killed Meta Retargeting. Revenue Went Up.
Retargeting had the highest ROAS in the account. It was also the easiest line item to cut.
A direct-to-consumer apparel brand was running Meta retargeting that reported a 9.4x ROAS in Ads Manager. By every dashboard, it was the best performing line item in the account. The team had been told for years not to touch it.
We helped them run a four-week test: retargeting paused completely, all prospecting and brand campaigns left untouched. The control was the same brand’s previous four weeks, normalized for seasonality.
The setup
- Spend on retargeting before pause: $22,400/month
- Reported ROAS in Meta: 9.4x
- Reported revenue from retargeting: ~$210,000/month
If the platform numbers were right, killing retargeting should have nuked $210k in monthly revenue.
What actually happened
Across the four-week pause:
- Total site revenue increased 3.1% vs. the seasonally-adjusted baseline
- Returning-user revenue (the people retargeting was supposedly bringing back) was flat within noise
- Email-driven revenue ticked up slightly — likely because the same users were now getting flow emails instead of ads
Net incremental revenue from retargeting: indistinguishable from zero.
Why the dashboards lied
This is the part worth understanding. Meta’s reported revenue was technically “real” — those users did click the ad and they did purchase. What the platform couldn’t see is that the same users were going to purchase anyway, often because:
- They had an item in cart and were already coming back
- They were on a brand’s email list
- They had searched the brand by name on Google
- They had simply made up their mind already
Retargeting put a Meta-attributed “last click” in front of a purchase that didn’t need it. The ad was a tollbooth on a road people were already driving down.
What they did
- Cut retargeting to $0 permanently
- Reallocated $15k of the freed budget to prospecting (new audiences)
- Pocketed the other $7k as margin
- Kept dynamic product ads for genuinely cold audiences only
Eight weeks later, blended marketing efficiency was up 18%.
The uncomfortable lesson
The metric that looks best is often the metric that’s lying to you the most. ROAS rewards channels for being close to the conversion, not for causing it. If you only optimize toward platform-reported ROAS, you will systematically over-fund the channels that take credit and under-fund the ones doing the actual work.