Jake Thompson·
Monthly search-term audit prompt. Dumped 90 days of terms in, got back a negative list and a bid map — found 18% of spend going to terms that never converted once.
Clusters raw search terms, quantifies wasted spend, and outputs a ready-to-paste negative keyword list with match types.
Google Ads Search Term Waste Audit
You are a Google Ads specialist auditing search terms for wasted spend. I will paste a search terms report export.
Account context: {{account_context}} (what we sell, target customer, geographic focus)
Conversion definition: {{conversion_definition}}
Data (CSV columns: search term, campaign, ad group, impressions, clicks, cost, conversions, conv. value):
{{search_terms_csv}}
Analyze in this order:
1. **Cluster the terms** into intent groups: buying intent for our offer, research intent, competitor names, job seekers, DIY/free seekers, wrong product entirely, and ambiguous. Show cost and conversions per cluster.
2. **Waste quantification:** total spend on clusters with zero conversions across the full period, and spend on terms with clicks above {{click_threshold}} and zero conversions individually. Report both as absolute currency and percentage of total spend.
3. **Negative keyword list:** produce a paste-ready list grouped by theme. For each entry choose the match type deliberately — phrase negatives for patterns ("free", "jobs", "salary"), exact negatives for specific ambiguous terms we should exclude narrowly. Flag any negative that could block converting traffic based on the data (show the conflicting term).
4. **Rescue candidates:** terms with conversions that are NOT yet keywords — recommend adding as exact match with a starting bid inferred from their current CPC and conversion rate.
5. **Structural findings:** if one ad group attracts wildly mixed intent, say so and propose the split.
Rules: never propose a negative for a term with a conversion. Show your math on the waste numbers. If the data window is under 30 days or under 1000 clicks, warn that conclusions are directional only.
Ergebnisse
## Search Term Audit: HomeShield (pest control, Denver metro) — 90 days, 14,212 clicks
### 1. Intent clusters
| Cluster | Cost | Conv. | Examples |
|---------|------|-------|----------|
| Buying intent | $8,940 | 187 | "pest control near me", "exterminator denver cost" |
| Research | $2,210 | 11 | "what do termite droppings look like" |
| DIY/free | $1,830 | 0 | "diy ant killer", "home remedies for wasps" |
| Job seekers | $640 | 0 | "pest control jobs denver", "exterminator salary" |
| Competitor names | $780 | 6 | "[brand] pest control reviews" |
| Wrong product | $410 | 0 | "pest control for cannabis grow" |
| Ambiguous | $890 | 4 | "bug spray" |
### 2. Waste quantification
- Zero-conversion clusters (DIY + jobs + wrong product): $1,830 + $640 + $410 = **$2,880 = 18.4% of total spend** ($15,700).
- Individual terms >15 clicks, 0 conversions outside those clusters: 23 terms, additional $1,120 (7.1%).
- Total addressable waste: **~$4,000/quarter (25.5%)**, of which $2,880 is safe to cut immediately.
### 3. Negative list (paste-ready)
**Theme: DIY/free (phrase):** "diy", "home remedy", "home remedies", "homemade", "free"
**Theme: employment (phrase):** "jobs", "salary", "hiring", "careers", "training"
**Theme: wrong vertical (exact):** [pest control for cannabis grow], [pest control license colorado]
**Theme: ambiguous (exact):** [bug spray] — CAUTION: "bug spray for house" converted 2x at $31 CPA. Exact negative on [bug spray] alone preserves the longer converting variant; do NOT use phrase match here.
### 4. Rescue candidates (add as exact keywords)
| Term | Conv. | CPC now | Suggested start bid |
|------|-------|---------|---------------------|
| "same day exterminator denver" | 9 | $6.10 | $7.50 (conv. rate 2x ad group avg) |
| "mice removal cost" | 5 | $4.20 | $4.80 |
### 5. Structural finding
Ad group "General Pest" receives buying, research, AND DIY intent (61% of its spend is non-buying). Split into: Emergency/Same-day, Cost/Quote, and Species-specific (mice, wasps, termites). Route research terms to a separate low-bid group with an educational ad + remarketing tag instead of competing on service CPCs.
Data window: 90 days, 14k clicks — statistically solid for the cluster-level cuts; individual-term negatives under 30 clicks remain directional.
Modell: Grok 3
7 Likes3 SavesScore: 6
1 Kommentar
Hannah Meier·
The caution flag on [bug spray] is the detail that separates this from every bulk negative script that quietly blocks converting long-tails.