Sponsored Products Rebuild on Margin-Based Bidding

Bids come from a break-even cost per click calculated per ASIN, not from an ACOS target handed down by finance. A product carrying EUR 6.20 of contribution after the referral fee, FBA fulfilment, storage and a trailing ninety-day return rate, converting at 9 percent on its detail page, can absorb about EUR 0.55 a click before it stops paying for itself. That figure, one per ASIN, is the spine of the rebuild. Everything else in the account is an argument about how close to it a given search term deserves to sit. How the account gets structured: Three layers, not thirty. A discovery layer on automatic and broad targeting with a weekly budget that does not flex. An exact-match layer holding only terms that have earned a decision. A product-targeting layer aimed at ASINs I have looked at one by one. Negatives move upward on a schedule so discovery stops paying a second time for what exact already owns. Campaign and ad group names follow a delimiter convention your analysts can parse without a lookup table, because naming is the one part of an ad account nobody can fix later. I do not build single-keyword ad groups. Amazon's bidding needs conversion volume in one place before it settles; four hundred ad groups taking two clicks a week each never leave the learning phase, and the reports they produce cannot be read by anybody. Placements, budgets and dayparting: Placement multipliers are set from each campaign's own placement report, not applied as a house rule. Top of search usually converts better and costs more; whether the extra cost is worth paying depends on the ceiling above, and across a single account the answer commonly differs by a factor of three. Dayparting goes in only where the hourly conversion data supports it. For most catalogues it does not, and I will say so rather than build a schedule that looks industrious. What is not in scope: No listing copy, no image work, no Sponsored Brands, no Display, no DSP. No third-party bidding software: if a tool moves your bids on rules it will not show you, you cannot explain your own account to your own board. No percentage-of-spend fee, ever. I am paid the same whether your spend rises or falls, which is the only arrangement under which my advice to cut spend is worth hearing. No inventory forecasting, no reimbursement chasing, no case work with Seller Support. Who should not buy this: If your catalogue is under fifteen ASINs and spends less than EUR 8,000 a month across all marketplaces, the ceiling on what a restructure can return sits below my fee. Write to me anyway and I will name the two changes to make yourselves. If your ASINs go out of stock more than a few days a quarter, fix that first: a rebuilt account on broken inventory only loses money more precisely. What you receive: A bid ceiling model, one row per ASIN, with the fee, discount and return assumptions visible and editable by your finance team rather than buried in my method. Bulksheets applied through the Amazon Ads API, with the prior state exported first so any change can be reversed the same day. A written memo naming the decisions taken, the ones I declined to take, and the three things that would change my mind. Two working sessions with whoever owns the P&L, not with whoever owns the dashboard.