Negative Keyword Sweep

Variables

Replace each placeholder before you run the prompt.

• {{industry_context}}Industry, target persona, what you do NOT sell.

  Industrial fluid power. We sell hydraulic pumps for OEMs and large maintenance contracts. We do NOT sell: hobby/auto/home/jewelry/medical/aerospace/defense.

You are building a negative keyword list from scratch for a new B2B account. The account has not run yet, so you need to anticipate contamination patterns proactively.

INDUSTRY CONTEXT:
{{industry_context}}

Build a comprehensive negative keyword list organized into these buckets:

1. Wrong vertical (we don't sell this)
2. Wrong customer type (B2C, hobbyist, DIY, school, etc.)
3. Job seekers (jobs, careers, salary)
4. Informational only (what is, how to, definition, wikipedia)
5. Geographic exclusions (countries / regions you don't serve)
6. Pricing intent we don't serve (cheap, free, sample, trial)
7. Adjacent products that look similar but aren't ours
8. Pure brand-protection (competitor brands you do not want to bid on, even broadly)

For each negative, suggest match type (negative exact / phrase / broad). Aim for 100–200 negatives total. Quality over quantity — every negative should have a clear bucket.

Do NOT include obvious universal negatives like "free porn", "free download" — assume those are at the account level already.

Expected output shape

A bucketed negative-keyword list of 100–200 entries with match-type per term.

Why we wrote it

New B2B accounts launch with empty negative lists and learn the hard way. This prompt builds the safety net before launch instead of after.

How to use

1. Open Claude or ChatGPT. The recommended model for this prompt is claude-sonnet-4-6 — opus when the prompt requires deep reasoning, sonnet for the rest.
2. Replace every {{variable}} with content specific to your account. The examples above are starting points, not templates to ship as-is.
3. Paste the prompt and run.
4. Read the output against the expected shape above. If the model produced a structurally different response, re-prompt rather than accept the drift.