Query-State Economics: why your Amazon budget should follow the state of each search query

The average that lies to you

An Amazon campaign reports one ACOS. Inside it, a dozen search queries behave a dozen different ways. One converts at a cost that would justify twice the budget. Three convert at a cost that would justify none. The rest sit between. The campaign average blends all of them into a single number, and then a bid rule acts on that number as if it described anything real. Budget goes up on the strength of the winners and the losers ride along. Budget comes down because of the losers and the winners starve. Either way the decision was made on a fiction.

Query-State Economics replaces the average with a state. Every search query that touches an account is placed in one of five states based on where the brand stands on that query and what the query returns. The decision follows the state. Budget, bids, content, and inventory each have a rule per state, and the rule changes the week the state changes.

The five states

• Dominating. The brand holds the top organic position and a large share of purchases on the query. The job is to defend at the lowest possible cost. Paid spend here mostly buys sales that would have happened anyway, so bids come down until share starts to slip, and not before.
• Winning. The brand leads on the query but the lead is not settled. Press. Budget is protected, bids are held or raised, and content is checked against the two closest competitors so the lead becomes a position.
• Competing. The brand is one of several credible answers. This is where most accounts waste the most money, because the honest question is whether the query is worth the fight at all. Contribution margin decides. If the query cannot pay for itself at a realistic share, it is cut, and the budget moves to a query that can.
• Opportunity. Demand exists and nobody has claimed it, or the brand is absent from a query it should own. Fund it, build content for it, and watch the conversion rate before the rank, because a query that converts will rank.
• Bleeding. The query costs more than it returns and has for weeks. Cut it. Not optimized, cut. A bleeding query is a decision that was never made, and every week it runs is the same decision deferred.

How you find the state

Three reports are enough to start. Search Query Performance shows the brand's share of impressions, clicks, cart adds, and purchases per query, which is the clearest picture of standing Amazon offers. The advertising search term report shows what each query costs and returns. Organic rank tracking shows whether the paid position is building an organic one or replacing it. Put the three side by side per query and the state is usually obvious. Where it is not, the rule is to hold the current state for one more week rather than guess.

The cadence matters more than the tooling. States are read weekly, on the same day, from the same reports. A query moves state when the evidence moves, not when someone has a feeling about it.

What changes when budget follows the state

Spend stops going where the account is already strong and starts going where the account can become strong. Bleeding queries stop draining the budget quietly. Content work gets a target list, because opportunity queries tell the listing what it needs to say. Inventory planning gets a warning, because a winning query with a stockout coming is a rank about to be lost. The account starts to behave like a portfolio of positions instead of a pile of campaigns.

Where to start

Pick one ASIN. Pull its queries from the three reports, place each one in a state, and look at where the spend actually goes. Most accounts find that a small number of bleeding queries were funding themselves out of the winners. If you want an operator to do that read with you, start with your MTB Score.