The parent listing was winning.
By gabriel-caceros-528077bb · September 4, 2026 · Curated by George's Blog
The parent listing was winning.
The wrong child was cashing the check.
A hair-care brand ran most Sponsored Brands and auto traffic to the parent.
Amazon routed shoppers to the child with the cheapest sticker price.
That child had:
→ 19% contribution margin
→ 4.1% CVR
→ The color nobody searched for
The hero child had:
→ 47% contribution margin
→ 9.8% CVR
→ 61% of organic reviews
Ad sales mix:
→ Low-margin child: 58% of ad-attributed units
→ Hero child: 22%
→ Two mid children: the rest
The parent was not a strategy.
It was a routing problem with a nice storefront.
We broke the cannibalization:
→ Exact terms for the hero shade went to the child ASIN, not the parent
→ Auto and broad stopped feeding the cheap child
→ Sponsored Brands collections featured the hero plus one legitimate bundle
→ The cheap child kept a defensive campaign only on its own color terms
→ Weekly mix report by child, not by parent
Revenue dipped for 8 days.
Then contribution rose $22K over 45 days.
Ad-attributed units were lower.
Dollars kept were not.
The cheap child had converted last on a coupon weekend.
Amazon remembered that and kept routing parent traffic there for 11 days after the coupon died.
That is the trap.
Parent ads inherit the last child that got lucky.
If you advertise at parent level, pull the child mix before you congratulate the campaign.
Do it by units and by contribution, not by sessions.
Amazon will send traffic to whatever converted last, not to whatever you can afford.
If 50% or more of ad units sit on a child that cannot carry the margin, split the traffic this week.
Do not wait for the parent ACOS to look sick.
Parent is a listing structure.
It is not a media plan.
#amazon #amazonppc #amazonads #ecommerce #dtc