Google Ads search terms report hygiene & negative keyword systems for SMBs

Direct answer — how should SMBs use the search terms report?

Treat search terms as a weekly operating system, not a monthly curiosity. Export or filter terms with spend, clicks, or conversions; classify each material row as keep, refine (theme/RSA/lander), or negate; apply negatives at the right level (account for universal junk, campaign for theme purity, shared lists for multi-campaign consistency); assign a human owner. Do this before you “fix” CPA with a harsher Target CPA or another restructure.

This insight owns query hygiene and negative keyword systems for SMB Search (and the Search-adjacent habits that still matter beside Performance Max). Account tree and intent layers: intent-based account structure. Bidding after the tree can learn: Smart Bidding playbook. Leakage estimate: Ad Spend Waste Estimator. Measurement lies that make junk look profitable: bad conversion data.

Working rule: negatives protect themes; they do not replace themes. If your ad groups mix emergency, DIY, and competitor intent, no negative list will stay clean.

What “good” looks like after thirty days: weekly review done in under an hour for a typical SMB account; universal junk blocked once; theme campaigns stop buying jobs, free, DIY, salary, and wrong-geo tourism; broad match (if used) has an owner; leadership can see waste falling without a mysterious volume cliff from over-negation.

If you only remember three moves: weekly triage by spend, layered negatives with an owner, fix soft primaries before you trust conversion-heavy search terms.

What the search terms report is (and is not)

The search terms report shows actual queries (and close variants Google surfaces) that triggered your ads and accrued impressions, clicks, cost, and conversions. It is the ground truth for “what did we buy?” — keywords are the bids you placed; search terms are the auctions you won.

It is not:

• A complete transcript of every query Google considered (privacy thresholds and sampling still hide thin rows).
• Proof that a converting term is a good lead — soft primaries make DIY and spam look like gold (primary vs secondary).
• A substitute for CRM quality. “Converted” form spam is still spam.
• The same thing as Performance Max search themes / limited query insights. PMax has thinner query visibility; do not pretend Search hygiene alone cleans PMax brand theft — that is a sibling problem (PMax attribution, PMax playbook).

Entities to keep straight: search terms, keywords, match types (exact, phrase, broad), negative keywords (exact vs phrase negatives), shared negative lists, account / campaign / ad group negatives, close variants, Smart Bidding, Quality Score (secondary), Search Terms insights in Ads UI, and waste from location “interest” targeting (geo setting — not a negative, but it shows up as junk queries from other cities).

Out of scope as deep dives: full SKAG vs theme structure (structure guide); broad match + Smart Bidding; RSA for service SMBs; PMax-only brand exclusion mechanics as the whole article.

Why search terms hygiene fails on SMB accounts

Pattern we see on audits:

1. Nobody owns it. Agency “optimizes bids” while junk queries run for months.
2. Review is vanity. Someone glances at converting terms only — never the high-spend zero-conversion rows.
3. Over-negation panic. After one bad week, they block half the category language and volume collapses.
4. Wrong level. Ad group negative mazes instead of campaign themes; or account lists so aggressive brand and money themes starve together.
5. Broad without adult supervision. Broad + dirty form primary + no negatives = spam factory.
6. Structure debt. One campaign holds brand, emergency, and research — search terms become unreadable soup (structure guide).
7. Path blindness. Calls and WhatsApp close jobs while Ads celebrates form junk on irrelevant queries (multi-path, Lead Path Coverage).

Hygiene is ops discipline. It is not a one-time “add 200 negatives” SOW line.

Example triage — local trades non-brand (illustrative)

Imagine a UK plumber non-brand campaign spent £1,200 last week. Material search terms include:

• emergency plumber near me — money; RSA/lander already match; keep.
• how to unblock toilet diy — junk/education; phrase-negate diy and how to at campaign level if not already on shared DIY list.
• plumber jobs birmingham — employment; account/shared employment list.
• boiler repair cost — adjacent commercial investigation; keep if you sell boiler work; else campaign-negate or move to a boiler theme with its own lander.
• cheap plumber free quote — tire-kick; phrase-negate free quote only if your offer is never free quotes — many trades keep “free quote” as money language, so decide from CRM, not vibes.
• A converting term toilet repair that sales marks as DIY tire-kick — do not only negate; check whether form primary is too soft and whether the lander attracts DIY intent.

Fifteen minutes of classification beats two hours of random exact negatives. Repeat weekly; the family list stabilizes.

The weekly triage system (45–60 minutes)

Step 1 — Scope the right campaigns

Prioritize non-brand Search with meaningful spend. Brand usually needs lighter review (misspellings and competitor leakage). Pause or tiny research campaigns can wait if cash is tight — unless they are bleeding.

Step 2 — Filter for materiality

In the last 7–14 days (or 30 if volume is thin), sort by cost. Open rows with cost above your “care threshold” (example: anything over ~1–2% of weekly non-brand spend, or every term with 3+ clicks). Also open converting terms — confirm they are qualified in CRM, not just marked converted.

Step 3 — Classify each material term

| Label | Meaning | Action | |---|---|---| | Money | Real buyer language for this theme | Keep; ensure RSA/lander match | | Adjacent | Related but wrong theme or weak intent | Move theme, tighten match, or campaign negative | | Junk | Jobs, free, DIY, porn, wrong service, wrong geo tourism | Negate at correct level | | Spy / competitor | Competitor brand or comparison | Campaign policy: allow only in competitor campaign with unique lander | | Research | How-to / definition with no commercial intent | Negate from money themes; optional tiny research budget |

Step 4 — Apply negatives in batches

Do not toggle one exact negative per row for an hour. Cluster families: e.g. free, diy, how to, salary, course, training, jobs, hiring. Prefer phrase negatives for families; exact for single dangerous tokens that would overblock if phrased broadly.

Step 5 — Log decisions

A simple sheet: date, campaign, term family, negative added, level (account/campaign/shared), owner. Without a log, the next strategist re-discovers the same junk and re-blocks brand by accident.

Step 6 — Spot structure feedback

If the same junk family appears every week inside a “service” theme, the theme is too wide or match is too loose — negate and fix the tree. Search terms are a diagnostic for structure, not only a negation inbox.

Negative keyword layers (the system)

Account-level (or shared list applied account-wide)

Universal junk that should never spend for this business: employment intent, porn/adult, warez, obvious DIY where you never sell DIY, irrelevant languages you cannot serve, scam patterns you have already proven.

Keep this list boring and stable. Review monthly, not daily.

Campaign-level

Theme purity. Emergency campaign negates diy, how to replace, parts only if those belong elsewhere or nowhere. B2B SaaS demo campaign negates salary, internship, jobs. Local trades negate wrong trades (electrician inside a plumber-only theme).

This is where most SMB value lives.

Ad group-level

Rare. If you need heavy ad group negatives, your theme is probably wrong — split the ad group instead of building a fractal maze (structure guide).

Shared negative lists

Best for multi-campaign accounts: one “Universal junk” list, one “Employment,” one “DIY/how-to,” applied to the right campaign set. Still assign an owner. Lists without ownership rot into sacred cows nobody dares edit.

Brand campaign specifics

Negate competitor brands if close variants leak. Negate non-brand service junk that somehow matches. Do not copy the entire non-brand negative stack blindly onto brand — you can block legitimate branded + service queries.

Exact vs phrase negatives (practical rules)

Phrase negatives block queries that contain the phrase as a contiguous concept (with Google’s matching nuances). Use for families: free quote may be too aggressive for some lead-gen offers; free download, how to, diy, salary are usually safe killers for service money themes.

Exact negatives block that specific term (plus close-variant behavior — do not treat exact as a mathematical guarantee). Use when a single query is toxic but related language is money (example: negate a specific competitor misspelling without nuking a category).

Do not paste a 500-row exact list of every bad URL-looking query you saw once. Cluster. Exact list bloat is how accounts become unmaintainable.

Conflict awareness: a negative at account level can silently kneecap a campaign you later want for research. Prefer campaign-level for anything that might be intentionally allowed in Layer 3 research budgets.

Broad match, Smart Bidding, and negatives

Broad match expands reach. Smart Bidding decides which auctions look efficient against your conversion definition. If that definition is soft, broad finds soft junk efficiently (bad conversion data).

Structural rule from the structure guide still holds: broad lives only where measurement and negatives are mature. This insight is the negatives half of that bargain.

Operating rules when broad is on:

• Weekly search terms are mandatory, not optional.
• Shared + campaign negatives must exist before scale.
• If junk rate stays high after two weeks of hygiene, pause broad on that theme — do not “bid harder.”
• Deep broad + Smart Bidding strategy: broad match + Smart Bidding. Do not treat this URL as that full essay.

Junk families SMBs should expect

Customize per vertical, but these recur:

Employment: jobs, hiring, salary, career, vacancy, cv, resume, apprenticeship.

DIY / education: how to, diy, tutorial, course, certification, training (when you sell the service, not the course).

Free / tire-kick: free, cheap as a sole intent, “under $X” fantasy for premium work, student discount spam if you do not offer it.

Wrong trade / wrong SKU: adjacent services you do not sell; parts-only when you sell install+labor.

Wrong geo tourism: city names you do not serve (also fix location options — presence vs interest).

Legal / complaint adjacent: “lawsuit,” “scam” + your category when you are not a lawyer — usually negate unless reputation defense is intentional.

Adult / scam / malware adjacency: obvious; account list.

Competitor: policy choice — separate campaign with unique lander, or negate from money themes.

Vertical notes:

• Home services / trades: DIY and “parts only” dominate waste.
• Automotive workshops: DIY, used parts, wholesale, wrong package tier language.
• Clinics: symptom research vs treatment intent — be careful not to overblock symptoms that precede booking; prefer lander match and theme split.
• SMB SaaS: jobs, login (sometimes), free forever tire-kick, competitor brand wars.

Think like revenue keywords: volume of junk is not a reason to keep buying it.

What you should not negate

Over-negation is how “hygiene” becomes a quiet lead outage.

Do not negate:

• Core service language because one bad lead used those words.
• Near-me / local modifiers wholesale when you are a local business.
• Brand misspellings on the brand campaign.
• Every research term forever if you intentionally run a small research layer — segregate it instead.
• High-intent commercial modifiers (emergency, same day, near me, book, price) on money themes without evidence they are toxic for you.
• Entire cities via negatives when the real fix is location targeting or a geo campaign split.

Test: if removing a negative would clearly buy a customer you want, you overblocked.

After a large negation batch, watch impression share and search terms for seven days before another purge. Stacked purges plus a new Target CPA is how accounts enter learning phase limbo with no diagnosable cause.

Search terms ↔ structure ↔ bidding (one system)

| Symptom | Likely cause | Fix order | |---|---|---| | Same junk every week | Theme too wide or broad unsupervised | Split theme; add campaign negatives; review match | | Converting junk | Soft primary | Fix conversions first | | CPA spiked after negation day | Overblock or learning reset | Roll back aggressive list; freeze bids | | Brand CPA “perfect,” non-brand dead | Blended reporting / shared budget | Split brand; read non-brand terms alone | | Good terms, bad leads | Lander/offer/path coverage | LP + call/WhatsApp + CRM — not more negatives |

Estimate leakage patterns with the Ad Spend Waste Estimator, then confirm in live search terms — estimators do not replace the report.

Score structure debt with Google Ads Health Score. Hygiene cannot save a tree that mixes brand and cold prospecting under one Target CPA (Max Conv vs tCPA).

Ops playbook — first 14 days on a dirty account

Days 1–2: Export 30-day search terms for top spend non-brand campaigns. Tag junk families. Screenshot worst offenders for stakeholders.

Days 2–3: Confirm conversion hygiene and path coverage. If forms are primary and phones close jobs, pause the negation hero story until measurement is trustworthy.

Days 3–5: Build shared lists (universal, employment, DIY). Apply to non-brand money campaigns. Do not nuke brand yet.

Days 5–7: Campaign-level theme negatives. Fix obvious structure soup if search terms prove mixed intent.

Days 7–10: Second triage pass. Remove accidental overblocks. Document the log.

Days 10–14: Only then discuss Target CPA tightening or broad expansion. Hygiene first, aggression second.

Ongoing: weekly 45–60 minutes; monthly account-list prune; quarterly theme review with the structure guide.

Common failure patterns

• Monthly “optimization” that never opens search terms.
• Exact-negative museums with 2,000 rows and no clusters.
• Account negatives that block a future research or competitor campaign by accident.
• Negating converting spam instead of demoting the soft primary.
• Adding broad match the same week as a 400-negative dump and a new tCPA.
• Treating PMax “brand is fine” dashboards as proof Search query waste is fixed.
• No owner named in the SOW or internal RACI.
• Reporting “we added negatives” without showing junk spend before/after.
• Blocking cheap / cost / price on service accounts that win on commercial investigation queries.
• Using ad group negatives to paper over three services in one ad group.

How this fits the Google Ads cluster

• This insight (CP-026): search terms hygiene and negative keyword systems.
• Structure guide (CP-025): where negatives hang on the tree.
• Smart Bidding / Max Conv vs tCPA / learning phase: what happens after queries are clean enough to learn.
• Waste Estimator + Health Score: fast flags; this URL is the weekly OS.
• Shipped siblings: broad + Smart Bidding (CP-027), RSA for service SMBs (CP-028), AI Max for Search (CP-035).

Browse more Google Ads writing in the Google Ads insights category. Service overview: Google Ads.

FAQ

How often should I check the Google Ads search terms report?

Weekly for active non-brand Search spend. Daily only during launches, broad experiments, or active waste incidents. Monthly-only review is how junk becomes “normal.”

What is the difference between keywords and search terms?

Keywords are what you bid on. Search terms are the queries that triggered ads. You optimize themes with keywords; you police waste with search terms and negatives.

Should negatives be account-level or campaign-level?

Both. Universal junk at account/shared level; theme purity at campaign level. Heavy ad group negatives usually mean the theme should split.

Can negatives hurt Smart Bidding?

Yes — if you overblock money language or change huge lists while also changing bids, learning destabilizes. Hygiene helps Smart Bidding when it removes junk without deleting demand.

Do I need different negatives for broad match?

You need more discipline, not a mystical separate religion. Broad without weekly search terms and layered negatives is negligent. Pause broad if junk persists after hygiene.

Should I negate every zero-conversion term?

No. Low volume terms need patience; high-spend zero-conversion junk families need action. Negating every thin row creates bloat and accidental blocks.

How do search terms work with Performance Max?

Search campaigns still need this system. PMax query visibility is limited; use brand exclusions, negatives where available, and Search foundations. Full PMax exclusions OS: Performance Max brand exclusions & negatives.

What if converting search terms are still bad leads?

Fix conversion definitions and CRM loops first. Negating a term that “converts” under a soft primary just moves the spam to the next cheap query.

Who should own negatives in an agency engagement?

Name one owner in the SOW. “The team watches search terms” means nobody does. Client stakeholders should see a monthly junk-spend before/after, not only blended CPA.

Exact or phrase negatives for DIY?

Usually phrase for families like DIY / how to / tutorial. Use exact when a single token is toxic but related phrases are commercial. When unsure, phrase for the family and monitor for seven days.

Can shared negative lists replace campaign negatives?

No. Shared lists handle universals. Campaign negatives protect theme economics. Using only shared lists either under-protects themes or overblocks everything the same way.

How do I prove hygiene worked to a non-marketer owner?

Show junk family spend and spam lead rate before vs after — not a screenshot of a longer negative list. Pair with booked jobs and metrics that matter.

What to do next

1. Open last 14 days of search terms on your top non-brand campaigns.
2. Build or refresh universal + employment + DIY shared lists.
3. Apply campaign-level theme negatives with a written owner.
4. Confirm conversions are trustworthy before trusting “converting” junk.
5. Rebuild messy trees with the account structure guide.
6. Estimate remaining leakage with the Ad Spend Waste Estimator.

Get a free Google Ads audit — we will sample your search terms, map negative layers, and tell you whether waste is a hygiene problem, a structure problem, or a tracking lie.

Search terms are where the auction confesses. Read the confession weekly, negate with a system, and stop letting Smart Bidding optimize toward queries you never intended to buy.