Order Value Held, $46.4K and 13.84% Added

The challenge

Twelve thousand orders a month at $27.91, leaking in two places and neither had an owner. Third party sellers were taking the Buy Box on the core ASINs often enough to matter, which meant paid traffic to those pages was handing orders to a reseller at our expense. Underneath that, broad campaigns were bidding on queries the exact campaigns already owned, so the best converting terms were bought twice. The second problem could not be fixed first: cleaning up the advertising while a reseller held the box would only have sent cheaper traffic to someone else's order. The obvious move, a deeper coupon, would have paid for both leaks rather than closing either.

What we changed

• Started with Buy Box ownership rather than the advertising, because paid traffic to a page where a reseller holds the box buys that reseller an order. Split the losses by cause: price, availability, seller rating.
• Removed the unauthorised offers through Brand Registry and test purchases, accepting that enforcement takes weeks and that the month would look slow until the box came back.
• Floored repricing on contribution per unit rather than the lowest competing offer, which means losing the box on some ASINs by choice instead of winning it at a price that does not pay.
• Only once box share was stable, ring fenced the proven converters in exact match and negated them out of every broad campaign, ending an internal auction that blended ACoS had been hiding.
• Replaced the account wide coupon percentage with depth set per ASIN against contribution, so promotional visibility continued without any listing carrying a discount deep enough to reset its price history.

The outcome

• $46,405.90 of additional sales, a rise of 13.84%, taken without a single discount deeper than the month before.
• Order value finished at $27.93 against $27.91. Unchanged, which was the whole brief on an account clearing twelve thousand orders.
• Orders landed at 13,668 and units 14,321, both advancing 13.7%, moving in step because units per order stayed at 1.05.
• March traded three days more than February. Measured per day the gain is 2.82%, not 13.84%, and that is the honest read of the month.
• Orders per day moved from 429.2 to 440.9, so the account picked up roughly twelve orders a day rather than changing its run rate.