eBay, Etsy and Walmart ads priced against organic sales

The first question I answer: Whether the sponsored placement bought a sale or bought back a sale you already had. On eBay an item can hold a top organic position and appear as a promoted result on the same page, and Promoted Listings Standard charges the ad rate on the sale either way. Across the seller accounts I have audited, between a fifth and a third of promoted revenue sat on items already ranking inside the first ten organic results for the same query. That percentage is the first number I produce and the reason the rest of the work is worth paying for. eBay: Standard and Advanced are two different products and most sellers run one of them for the wrong reason. Standard is a percentage of the sale charged after the fact, which makes it a tax on items that would have sold anyway, so items with defensible organic rank either sit at a floor rate or come off promotion. Advanced is a cost-per-click auction with keyword control and top-of-search placement, which is where queries you do not rank for belong. Suggested ad rates are a market average, not advice, and I do not follow them. Etsy: There is one control, a daily budget, plus the choice of which listings are eligible. There is no keyword bid, so the work sits upstream: which listings are switched on, whether attributes and category are filled in well enough for Etsy's own matching, and whether the offsite fee that applies above the revenue threshold is being carried in your margin arithmetic. Sellers who treat Etsy Ads as a keyword channel spend months tuning a lever that does not exist. Walmart Connect: Eligibility comes before bidding. Without buy box ownership the sponsored placement does not serve, and listing quality gates what you are allowed to buy at all. I keep automatic and manual campaigns separate so the automatic one functions as query discovery and nothing else, and I read search carousel and item page placements apart from each other, because item page inventory is frequently buying a competitor's traffic at a rate no keyword bid would justify. Item-level reporting comes through the advertising API rather than an interface export. How the decision is made: Every recommendation to keep spending is tested by stopping. A matched cell of items runs unpromoted for 14 to 21 days and I compare total units and total revenue, organic and promoted together. If the total holds, the promotion was paying for traffic you already owned. This is uncomfortable to run in a quarter with a number to hit, and it is the only method I trust. If no part of the catalogue can be held out, we are not a fit and I would rather say that now. Not part of this: Listing copy, photography, category expansion, pricing and inventory. Amazon Ads, deliberately: different ad server, different reporting surface, and you almost certainly have someone on it already. No fee as a share of ad spend or of marketplace revenue. Not for: Sellers under roughly 250 orders a month per marketplace, where a quarter of data will not separate signal from noise. Catalogues where one or two articles carry most of the revenue, because the answer there is merchandising rather than ad rate. And anyone who wants the ad rate raised until the units appear. That does work, briefly, and it is not what you would be paying me for.