Winning the sizes nobody optimized: +97% revenue from non-queen SKUs

In bedding the shopper almost always types the size — “king sheets”, “split king cooling” — which makes every size a separate search market with its own competitors and its own price expectations. The brand's variation family had one set of copy written for the queen, and Amazon indexed the whole family against it.

The result was a strong queen and five weak children. Advertising papered over it by pushing the family's main image, so shoppers arriving on a king search landed on a page that talked about a queen, and the conversion rate showed it.

We treated each size as its own listing problem inside one family.

• Per-size keyword sets — research run separately for each size, including the split-king and twin-XL terms the brand had never indexed for.
• Size-specific bullets and backend terms on every child, with the family's shared claims kept consistent.
• Campaigns per size rather than one family campaign, so bids follow each size's own competition and margin.
• Image stack per size where it matters — the split-king and twin-XL children got their own first frame showing the configuration, which is the question those shoppers actually have.

Revenue from the non-queen sizes nearly doubled — $34k to $67k a month, +97% — lifting total revenue 45% without a single new product.

Indexed size-modified keywords went from 180 to 740, and with them the brand now appears in searches it was previously invisible to. ACoS improved seven points because traffic lands on the size the shopper asked for instead of on a page arguing for a different one.