Negative Keywords That Do Not Bite, and a Query Worth Promoting

Negative keywords are scoped, not absolute.

• For Sponsored Products, negatives apply at the campaign level. A negative added to the wrong campaign - or to a campaign that is not the one serving the impression - has no effect. The first check is the campaign column of the search term report: if the kids impressions come from a second campaign, the negative in YOGA-Broad is irrelevant to them.
• Match type mechanics. A negative phrase blocks search terms that contain the phrase as a contiguous sequence, so a campaign-level negative phrase 'kids' should block 'yoga mat for kids' in that campaign. A negative exact 'kids' would not, because it only blocks the exact query 'kids'. Confirm which one was actually created.
• Timing. Search term reports include historical rows. If the negative was added two months ago, look at the impression dates before assuming it failed; if impressions continue after the date, it is a placement or scoping issue.
• Also check whether a second ad group or a product-targeting campaign is catching the same queries, and whether a broad term other than 'yoga mat' (for example 'exercise mat') is matching into kids queries.

Only after confirming scope should you add anything, otherwise you accumulate negatives that duplicate each other and mask the real source of the spend.

• 'yoga mat cleaner' (45 clicks, 0 orders): the intent is a different product category. Add a campaign-level negative phrase 'cleaner' (plus 'cleaning') rather than a single negative exact, so variants are covered. If you actually sell a cleaner, do not simply negate it - give it its own campaign with its own budget and a mat-cleaner ad group, because it does not belong in the mat campaign's wallet.
• 'yoga mat for kids' (120 clicks, 2 orders): the decision depends on the catalog. If you sell kids' mats, this is a relevance and structure failure - split it into a dedicated kids campaign with its own bid and budget. If you do not, negate at phrase level in the serving campaign and verify scoping as above.
• 'extra thick yoga mat' (80 clicks, 6 orders): promote it. Create an exact-match ad group with a bid reflecting the observed economics, and keep broad running so you keep discovering adjacent long-tail variants. Negating the phrase in the broad campaign is optional: it forces the query into the controlled campaign, but it also removes volume. Do it only if the exact campaign demonstrably converts better.
• 'yoga mat' (400 clicks, 30 orders, 16% ACOS): leave alone - it is your best-performing query and well inside a 30% target. The broad campaign's 58% blended ACOS is caused by the long tail, not by this term.

First, decide whether the overlap is actually costing you. Two campaigns eligible for the same query mean you may pay a higher clearing price than necessary, and your reporting splits the query's performance across two places. That is a real cost, but it is not automatically larger than the volume you would lose by narrowing coverage.

Evidence here points to keeping both. Broad's 'yoga mat' traffic converts at 16% ACOS, well inside the 30% target. Reflexively adding a negative exact 'yoga mat' to the broad campaign would shrink eligibility for your best query. The 58% blended ACOS is a tail problem.

Practical management:

• Set bids deliberately different across the two campaigns so the one with the better economics wins more of the impressions, and compare their search term reports for the same query.
• Use negative exact in the broad campaign only if you see duplicated spend with flat total volume, or if the exact campaign is clearly more efficient and you want to force the query there.
• Alternatively, consolidate so all match types for one theme sit in one campaign with separate ad groups. This makes negatives, bids, and budget decisions live in one place and removes the chance of adding a negative to the wrong campaign - which is how this account ended up with kids traffic it thought it had blocked.