Expensive pattern
- Same query across three campaigns
- Too few keywords — or only vague ones
- Negatives too aggressive or missing the real junk
- RSAs / asset groups half empty; homepage for every ad
- Intake: “these aren’t buyers”
Account structure
Search Structure That Stops Expensive Leads
Bad structure doesn’t just look messy — it buys the wrong people. Here’s the operator meat under the diagram: coverage, negatives, RSAs, landing pages, and targeting.
- BrandOwn it. Don’t let junk bleed into brand CPCs.
- OfferOne intent theme — not fifty geo clones.
- NegativesKill DIY, jobs, and freebie queries first.
Clean boxes on a slide deck don’t save an account. Thin keyword coverage, generic landing pages, half-built RSAs, and sloppy negatives still buy the wrong people — just with prettier campaign names.
Not enough keywords — operators under-cover real demand, then wonder why volume sits with competitors. Expand from search terms that already convert; don’t invent fluff.
Generic keywords — bare terms like “lawyer” or “software” without intent modifiers pull browsers and tire-kickers. Pair modifiers to how buyers actually search.
Generic landing pages — one homepage for every ad group kills message match. Each theme needs a page that finishes the ad’s sentence.
Bad negatives — too aggressive and you starve volume; too loose (or the wrong list) and you keep buying DIY, jobs, and freebie queries. Update from live search terms.
Half-filled RSAs / asset groups — thin assets = thin relevance. Fill headlines and descriptions with honest benefits, offers, and proof pins. See Google’s responsive search ads overview.
Poor targeting — loose geos and “everyone” audiences teach Smart Bidding to chase cheap, wrong clicks. Tighten to where sales actually closes.
Expensive vs protected
Messy structure doesn’t just look bad — it buys the wrong people.
Healthy pattern
- Query → campaign mapping is obvious
- Enough intent keywords to cover real demand
- Negatives updated from real junk search terms
- Full RSA assets + landing pages that match the ad
- Sales feedback steers bids and copy
Want a faster gut check on ad-to-page leaks? Take the Ads Eye Challenge.
Build around intent (and lead quality)
Campaigns should map to commercial intent themes: core service, brand, competitor conquest only if economics work, and geo splits when bids and messaging truly differ. Pretty folders that don’t match how people search still leak money.
Real operator mistakes that burn budget
Most “structure problems” are coverage and match problems wearing a neat campaign name. These are the failures I see in live accounts — not theory:
- Not enough keywords — skimpy coverage leaves high-intent demand on the table while competitors scoop it
- Generic keywords with no intent filter — “lawyer,” “software,” “tax help” without modifiers that separate buyers from browsers
- Generic landing pages — same homepage for every ad group; message match dies on arrival and Quality Score follows
- Bad negatives — too many (starving volume) or the wrong ones (still buying DIY, jobs, freebie, and info queries)
- RSAs / asset groups half-filled — thin headlines and descriptions = thin relevance and weaker ad strength
- Poor targeting — loose geos, wrong audiences, or “everyone” when the offer is niche; Smart Bidding can’t fix a map that includes non-buyers
Mistakes operators keep repeating in 2026
Beyond the classic six, these show up in almost every audit once match types loosened and automation got louder:
- Top search terms never promoted to keywords — the query converts, but it sits in the wrong ad group (or loses to PMax) because you never claimed it
- Broad match without target CPA/ROAS — Max Conversions + broad will chase volume; targets + shared negatives keep it honest
- Stale shared negative lists — decade-old blocks conflict with new keywords and quietly starve winners
- Brand traffic mixed into non-brand — cheap branded clicks inflate “non-brand” CPA and hide how bad prospecting really is
- RSA assets that rhyme instead of cover angles — fifteen near-duplicate headlines ≠ diversity; need offer, proof, urgency, and CTA themes
- Auto-apply recommendations left on — Google can add broad keywords or change bidding while you’re not looking
How to fix it without a six-week redesign
Start with search terms: expand coverage where real buyers show up, cut or negative the junk, and fill every RSA pin you can honestly support — with distinct themes, not copy-paste variants. Then align each ad group to one converting page — not a generic homepage.
Targeting next: tighten geos and audiences to where sales actually closes. Isolate brand. Turn auto-apply off and review negatives when you add new keywords. Structure is a container; these levers are what fill it with the right people.
Where Performance Max fits
PMax can complement Search. It should not silently cannibalize brand or paper over broken measurement. Dirty conversions + PMax = expensive noise with better creative reach. Claim your top converting search terms in Search first so PMax isn’t the default auction winner for queries you already know how to convert.
Official reference
Fill RSA assets honestly — Google’s overview of responsive search ads is the baseline. Then send each theme to a converting Google Ads landing page, not a generic homepage.
Need a Search rebuild that cuts junk?
Send the vertical, markets, and approximate monthly Search spend — I’ll tell you if structure, tracking, or the offer is the real bottleneck.