The report shows what shoppers typed, not what you targeted
The Sponsored Products Search Term report lists the actual queries that triggered ads, plus campaign, targeting, impressions, clicks, spend, orders, sales, and ACOS. It answers which real searches your budget bought.
Account or campaign ACOS averages a few expensive terms with many cheap ones. The report’s job is to break that average back into term-level actions.
Columns to read first: spend, clicks, orders, then ACOS and source
- Customer Search Term: the shopper query. This is what you negate, promote, or watch.
- Clicks and Spend: whether the sample is large enough. ACOS on a handful of clicks is not a conclusion.
- Orders and Sales: whether the query converted. High-spend zero-order terms go first.
- ACOS: compare it with this SKU’s target and break-even ACOS, not with a rumored industry average.
- Campaign / Targeting: the same query can perform differently by match type. Keep the source attached.
Judgment: three buckets, not row-by-row instinct
Give yourself two thresholds: a minimum click count to judge a term, and a target ACOS for the SKU. Those come from your price and allowable CPC, not from a generic industry article. Higher prices usually need more clicks before you decide.
- Promote to exact: clicks reached the watch threshold, orders are stable, ACOS is below target, and the query is relevant.
- Watch: clicks are still below the threshold, or orders are few but relevance is clear. Do not negate yet.
- Stop: spend is near or past allowable ad cost per unit, clicks are enough, and there is still no order or the query is irrelevant.
Steps: from the file to a first action list
- Pick a date range that matches stock and promotion state. Start with 7 and 14 days so testing and harvest are not mixed.
- Sort by Spend descending. Work the head terms first. Zero-spend long-tail rows are not the first job.
- Mark irrelevant queries for match-type negatives. Wrong accessories, mistaken brand terms, and off-audience queries go first.
- Mark high-spend zero-order terms. Check whether clicks already reached your watch threshold.
- Mark low-ACOS converting terms to move into exact or to raise the parent ad group budget.
- Write the three lists back into campaign structure. Do not mix negate, promote, and watch inside one broad ad group with no record.
Example: three terms in one report
Target ACOS 22%, watch threshold 12 clicks. Term A “glass coffee cup”: 18 clicks, $9 spend, 3 orders, ACOS 15%, relevant → promote to exact. Term B “plastic lid replacement”: 22 clicks, $11 spend, 0 orders, irrelevant → exact negative. Term C “insulated glass tumbler”: 6 clicks, $3 spend, 0 orders, relevant but thin sample → watch, do not negate.
Risk boundaries: where Search Term reports are misread
- Negating before the sample is enough closes relevant terms that had not converted yet, especially on new products.
- Negating the search term but ignoring the target lets broad match keep buying nearby junk queries.
- Low CTR alone is not a negate reason. Offer, price, and main image also suppress CTR.
- Ranges that include a big deal or a stockout distort ACOS. Ask whether the window still represents the current stage.
Build the same three-column list from your file
Sort your Search Term report by spend, apply your click threshold and break-even line, and you get promote, watch, and stop. Upload the file and check the same buckets instead of working row by row from memory.