Jake Thompson·
Dropped 2,000 search terms in and got a negative keyword list grouped by theme — killed about $340 of weekly waste in ten minutes
Mines a search terms report for negative keywords, groups them into themed lists with match types, and separates clear waste from judgment calls.
Negative Keyword Miner From Search Terms Report
You are a PPC analyst mining a search terms report for negatives. Precision matters: a sloppy negative blocks converting traffic, so separate certainty from judgment.
Search terms report (term, campaign, impressions, clicks, cost, conversions):
{{search_terms}}
What we sell, precisely — including what we do NOT sell: {{business_description}}
Current negative lists (so you don't duplicate): {{existing_negatives}}
Process:
1. **Classify every term with spend and zero conversions** into:
- CLEAR WASTE — unambiguously wrong intent (job seekers, DIY/free-only, wrong product, other geography, informational-only with no path to purchase)
- JUDGMENT CALL — plausibly relevant but unproven; list separately with the case for and against blocking
- KEEP WATCHING — right intent, just no conversion yet at low spend
2. **Extract the pattern, not just the term.** "free invoice template" and "invoice template free download" become one negative concept. Choose match type deliberately: phrase negative for the pattern word ("free") where the whole theme is waste; exact negative where only the specific term is bad.
3. **Group negatives into named themed lists** (e.g. Jobs & Careers, Free-Seekers, Wrong Product, DIY) so they can be applied per campaign — a term wasteful for the Search brand campaign may be fine for a competitor campaign.
4. **Quantify:** wasted spend per theme over the report period, and projected weekly savings.
5. **Collateral damage check:** for every phrase negative proposed, state what legitimate queries it could block. If you can't think of any, say "none found" — but actually think first.
Output: themed negative lists with match types, the judgment-call table, and the savings summary.
Ergebnisse
## Themed negative lists
### List: Jobs & Careers — apply to ALL campaigns
| Negative | Match type | Blocked spend (30d) |
|---|---|---|
| "jobs" | phrase | $61.20 |
| "salary" | phrase | $18.40 |
| "hiring" | phrase | $12.10 |
| "career" | phrase | $9.85 |
Collateral check: "jobs" phrase-blocks "software for print jobs" style queries — none appear in your report and you don't serve print; none found otherwise.
### List: Free-Seekers — apply to Search campaigns, NOT the trial campaign
| Negative | Match type | Blocked spend |
|---|---|---|
| "free" | phrase | $203.70 |
| "crack" | phrase | $8.90 |
| "open source" | phrase | $31.55 |
Collateral check: "free" is your biggest judgment risk — you DO offer a free trial. Recommendation: keep it out of the trial campaign (where "free trial" converts at 4.1%) and block it everywhere else, where 214 clicks produced 0 trials.
### List: Wrong Product
| Negative | Match type | Blocked spend |
|---|---|---|
| "personal budget" | phrase | $44.30 |
| "tax filing" | phrase | $57.15 |
| [quickbooks login] | exact | $22.60 |
Collateral check: "tax filing" — you handle VAT reporting but not personal tax filing; the phrase as searched is consistently personal-tax intent in your report. "quickbooks login" is exact because broader "quickbooks" terms belong to your competitor campaign and must stay biddable.
## Judgment calls — decide, don't default
| Term | Spend, 0 conv. | Case to block | Case to keep |
|---|---|---|---|
| "invoice app for iphone" | $38.20 | You have no native app | Mobile web works fine; 2 clicks reached signup page |
| "accounting course" | $19.75 | Education intent | Your academy content targets exactly this for top-of-funnel |
| "bookkeeper near me" | $27.90 | Service-seeker, you're software | Some convert after realizing software is cheaper — 1 assisted conv. in the lookback |
## Savings summary
- Clear waste blocked: **$469.75 / 30 days ≈ $110/week** at current volume; scaling the same term families across the full account (patterns repeat in 3 campaigns you excluded from this export): **~$340/week projected**.
- Nothing in KEEP WATCHING exceeds $15 spend — re-run this in 30 days before touching them.
- Note: your existing negative list already contains "template" — it duplicates nothing above, but it MAY be blocking "invoice template" queries that your templates landing page converts. That's outside this report's scope, flagged for a separate look.
Modell: Gemini 2.5 Flash
32 Likes21 SavesScore: 23
2 Kommentare
David Park·
The collateral damage check with an honest none found is what separates this from every negative keyword script out there.
Marcus Bell·
Splitting free by campaign instead of nuking it account-wide — that nuance pays for the whole exercise.