Search Ads: Query Auction, Match Governance, and RSA

Query

Search ads are query-triggered text, independent of which account

Search ads management is the auction format: a typed query, an RSA, match rules, and negatives. It runs inside Google or Microsoft. It is not Quality Score, not Performance Max architecture, and not a landing-test protocol.

Buyers say “search ads” when they mean control of the query. That control is this URL. The Google login and Quality Score stay on Google Ads. The Microsoft login and UET stay on Microsoft Ads.

If a paragraph still works after swapping this H1 for “PPC consulting,” it belongs on PPC. Coverage philosophy already decided whether brand and non-brand search are in the program. This page implements the auction, not the map.

We will not invent a CPC floor, a conversion-count quota, or a timeline SKU for the query auction.

Intent

Buckets are copy and match, not hub coverage essays

Brand, competitor, category, and problem queries get different RSA claims and match tightness. That is format work. It is not a reprint of which families the hub allowed.

Competitor terms are a policy conversation on this search-ads URL — allowed, refused, or tightly pinned — not legal advice and not a default template.

Category and problem queries still need a conversion the hub already named before match types widen. This format will not invent “smart bidding always.”

RSA

Pinning versus free rotation is compliance versus learning

Responsive search ads are this format’s creative unit. Pin headlines that legal or brand must not shuffle. Leave the rest free so the auction can learn. That choice is search ads. It is not a Quality Score lever.

Ad strength warnings are RSA hygiene here. They are not a CRO protocol. First-screen experiments stay on landing-page CRO. This page only requires that the pinned claim exists on the destination.

We will not reprint Performance Max asset groups. PMax may serve search-like inventory; architecture stays on Performance Max.

Negatives

Shared lists versus campaign lists — then search-term mining

Shared negatives protect every search-ads campaign that should never buy a class of query. Campaign negatives protect one cluster. Mixing those layers without a rule is refused.

Search-term mining adds exact or phrase where the query earned it, and negatives where it did not. That loop is this auction. Account-wide device defaults, if they exist, stay a Google Ads policy sentence — query-level device waste is mined here.

Dynamic Search Ads remain search ads: query-triggered text from a crawl. Product listing ads stay on Shopping ads. Cold banners stay on display ads.

Expand

Broad and DSA only after a conversion definition exists

Broad match and DSA are expansion tools on this format, not a personality. If `/ppc/` has not named a conversion you will stand behind, expansion stays out.

Microsoft search auctions still use this format. Profile layers and UET do not move onto this H1.

Organic handoff for a query the URL should own remains SEO services. Search ads may pause a term when organic already owns it; this page does not become the SEO hub.

Field notes

Query-auction notes that cannot live on Quality Score, PMax, or CRO

Search-term reports are the query auction’s object; they are not Quality Score columns.

RSA pinning on search ads is a compliance-versus-learning choice, not a Quality Score tactic.

Match-type expansion on search ads waits for a hub-named conversion; this format will not SKU “smart bidding.”

Shared negatives on search ads protect the whole query graph; campaign negatives protect one intent bucket.

Search ads on Microsoft remain this format; the Microsoft login and UET stay on Microsoft Ads.

Brand-query RSA still belongs to search ads as format; brand campaign containers inside Google sit on Google Ads.

DSA is still search ads — crawl-triggered text — not a PLA card and not a PMax asset group.

Query-level device waste is search-ads mining, not an account-wide default treatise.

Close-variant noise on search ads is match governance here, not a GDN placement report.

Search ads will not own Performance Max asset groups even when PMax serves search-like inventory.

The destination must show the RSA claim; A/B protocol for that URL is landing-page CRO, not this format.

We will not print a CPC, ROAS, or conversion-count floor for the query auction.

Competitor RSA on search ads is pinned or refused as policy, not sold as a default pack.

Sitelink extensions on search ads must match the query promise; they are not a CRO test matrix.

PPC management can operate search-ads campaigns; cadence philosophy stays on PPC management.

Hub coverage still decides whether the query auction is in the program at all — this page does not add Search because a slide said full funnel.

Depth

RSA query auction as the paid-search format object

The RSA query auction is the object of this page: a typed query, a responsive search ad, match rules, and negatives, on Google or Microsoft. It is not Quality Score columns and not an asset group. Control of the query is the job; the login that hosts the campaign is a sibling account page.

Intent buckets — brand, competitor, category, problem — exist so the RSA query auction can pair copy and match type with what was typed. That is not the hub’s coverage philosophy. PPC already decided whether those intents are bought. This URL writes headlines and match governance for the auction that remains.

Pinning RSA headlines is a compliance move in the RSA query auction; free rotation is a learning move. Mixing them without a rule produces ads you cannot explain. We pin what legal and brand require, then let the rest rotate. We do not treat pinning as a Performance Max creative setting.

Shared negatives versus campaign negatives are how the RSA query auction stays honest across ad groups. A shared list catches junk that every query class should refuse. Campaign negatives protect one intent from another. Search-term mining feeds those lists. Display placements are not negatives and do not belong here.

Broad match and DSA enter the RSA query auction only after a conversion definition already exists. Without that definition, expansion is unaccountable. We will not claim “smart bidding always.” Google Ads owns whether Search lives in the Google account; this page owns when the auction is allowed to widen.

Search ads on Microsoft are still this format. The RSA query auction does not become a UET essay when the campaign sits on Bing. UET is the Microsoft account object. Query, RSA, match, and negatives remain the format object even when the SERP chrome is Bing’s.

Competitor terms in the RSA query auction are a policy conversation: trademarks in copy, landing claims, and whether the query is even allowed. They are not a default template and not legal advice as a product. If competitor queries stay, match and RSA must say what you will actually deliver.

Ad group theming exists so the RSA query auction can keep ad relevance without stealing the Quality Score treatise. Tight themes make headlines easier to write. Loose themes dump unrelated queries onto one RSA. We theme for the auction, then send score diagnostics back to the Google account page.

Search-term reports are the ledger of the RSA query auction. They are not Shopping search-term approximations and not PMax’s limited channel view. We mine terms to add as keywords, as negatives, or as proof the query class was wrong. Mining is this format’s loop, not a weekly retainer slogan.

Extensions and sitelinks in search are part of the RSA query auction’s SERP real estate, not a landing-page experiment. They must match the query class. Tests of the page after the click live on the CRO child. Here we only refuse extensions that advertise a promise the RSA cannot keep.

Brand queries in the RSA query auction still need match control so you are not paying for misspellings you already own organically, and so competitors cannot sit on your name unchallenged if the hub included brand. That is implementation of the auction, not a new coverage decision.

The format refuses Shopping product listing ads, cold display, and Performance Max as substitutes for the RSA query auction. Those inventories do not let you pin a headline to a typed query the same way. If the buyer wants query control, they are on the correct URL.

Landing URLs in the RSA query auction must resolve to a page that could satisfy the query. That is alignment, not a CRO test matrix. If the URL 404s, the auction is buying a broken destination. Technical load issues hand off; message-match experiments hand off; the query-to-URL pair stays here as a format rule.

Match-type migration — exact, phrase, broad — is historical debris in many RSA query auction accounts. We do not keep duplicate match types that fight each other. We also do not flatten everything to broad as a fashion. Each remaining keyword has a job in the auction or it is removed.

Callouts about “AI-written RSAs” do not replace query governance. The RSA query auction still needs human negatives and intent buckets. Generated headlines that ignore the query class are waste. We will not invent a productivity metric for how many assets a model produced in a week.

The RSA query auction has no published CPC, conversion floor, or calendar SKU on this URL. The deliverable is query-triggered text you can explain: which RSA, which match, which negatives, on whichever search network the program already opened as a coverage door for Search format.