Negative keyword strategy for industrial B2B (with library)

Most B2B accounts have around 100 negative keywords. The accounts I take over from agencies usually have a standard "B2B negative list" they found online: "free," "job," "resume," "salary," maybe a few others.

That's not a negative keyword strategy. That's a starter kit that stops 5% of the garbage.

A real industrial B2B negative library has 5,000-15,000 terms. It's one of the most defensible assets in the account. And building it is the difference between a $297 CPL and a $890 CPL.

Why this matters more in industrial than anywhere else

Consumer overlap in industrial B2B is massive.

You're selling a $15,000 dot-peen marking machine. Google thinks "marking machine" is close to "tattoo machine," "embroidery marking machine," and "how to mark metal at home as a hobby." Your ad shows to all of them.

You're selling OSHA-compliant arc flash labels. Google expands your match to "arc flash game," "flash label maker app," "printable arc flash warning signs free download."

Every irrelevant click is money gone. And in niche verticals with $10-25 CPCs, a single afternoon of garbage traffic can burn thousands.

The five negative categories

1. Consumer/DIY modifiers Terms that signal personal, hobby, or consumer intent:

• "how to," "DIY," "at home," "yourself," "tutorial," "step by step"
• "cheap," "affordable," "discount," "free," "Amazon," "eBay," "Etsy"
• "small," "mini," "personal," "portable" (when your product is industrial-scale only)
• Specific consumer brand names that overlap with your category

2. Academic and informational Terms that signal research, not purchase:

• "definition," "meaning," "what is," "types of," "history of"
• "Wikipedia," "research paper," "thesis," "study," "textbook"
• "for students," "assignment," "homework," "learn about"
• "example of," "diagram," "explained," "overview"

3. Employment and career

• "job," "jobs," "career," "careers," "hiring," "employment"
• "salary," "pay," "wages," "compensation," "benefits"
• "training," "certification course," "how to become"
• "internship," "apprenticeship," "entry level"

4. Competitor names (if not running conquest) Every direct and indirect competitor. Every variation and misspelling. Updated quarterly.

5. Product category mismatches Terms that are genuinely similar to your product but represent a different category entirely:

• If you sell industrial labels: "label maker" (consumer), "label template" (design), "label printing service" (different business model)
• If you sell marking systems: "marking pen" (consumer), "grading marks" (education), "benchmark marks" (financial)

How to build it

Start with your search term report. Go back 12 months. Pull every query that received a click. Sort by zero-conversion queries with any spend. Read them.

Every irrelevant query teaches you a new category. "Laser marking tutorial" → add "tutorial" and "tutorials" and "how to" as negatives. "Arc flash safety game" → add "game," "games," "gaming" as negatives.

Do this weekly for the first three months on a new account. By month three, you'll have caught most of the obvious categories. Then shift to monthly maintenance.

Also use Google Keyword Planner in reverse: type in your core keywords and look at the adjacent terms it suggests. Many of those adjacent terms are exactly what you don't want.

Where to apply them

Three levels:

MCC-level shared negative lists. The universal garbage that applies across every brand and campaign. Consumer modifiers, employment terms, academic terms. Every account inherits these automatically.

Account-level shared lists. Brand-specific terms that don't belong in any campaign for that brand.

Campaign-level negatives. Query-specific terms that appear in the search term report for that specific campaign.

The hierarchy means you write the universal terms once and they protect everything. Campaign-level catches the edge cases that make it through.

The library is never finished. It's a living document. Build it, maintain it, and protect it when agencies try to "streamline" by reducing the list size. The list is the moat.