How to Set Up Google Ads Negative Keywords for Startups

Set up startup negative keywords by defining the current offer mismatch, reviewing Search terms with enough activation or account-quality evidence, checking positive overlap, choosing the narrowest match type and scope, and monitoring both waste and useful learning after saving. Keep early exclusions reversible; a generic list can remove valuable market signals before the startup understands its audience.

Start with Google’s negative keyword guide, Search terms workflow, and list-management guidance.

1. Define the current mismatch

Write why the query should not create a trial, activation, demo, opportunity, or order for this startup’s current offer. It may represent employment, student research, free or open-source intent, an unsupported use case, an incompatible company size, an existing-user support need, or a market the startup does not serve.

Record one query to block, one similar query to preserve, and one edge case. “Free,” “pricing,” “API,” “template,” “support,” and “jobs” can each be valuable or irrelevant as the offer changes. A negative list is a hypothesis about a product boundary, not a substitute for learning.

2. Review Search terms and available maturity

Open Search terms for the exact campaign and date range. Review cost, clicks, trials, activation, key product action, qualified demos, accepted leads, opportunities, account fit, and response time. Use the normal product or sales lag before declaring a query poor.

Early data can be dominated by one customer, launch, founder follow-up gap, outage, or pricing change. Prioritize repeated mismatches with material spend or unmistakable intent. One non-converting query in a small sample is weak evidence.

Record query, reason, report window, owner, proposed type, scope, and application date. A negative added today cannot change historical traffic. Annotate product, pricing, onboarding, landing-page, support, audience, and CRM changes.

3. Check the product path and positive overlap

Review positive keywords, ads, landing pages, trial or demo flow, product promise, integrations, documentation, content, account lists, and sales routing. Write one query to block and one similar query to preserve.

“Free” may be wrong for a paid-only tool and central to a freemium funnel. “Student” may be irrelevant to enterprise software and valuable to education. “Support” may be an existing-user mismatch or the product itself. “API” and “pricing” often describe legitimate research.

If a market or company-size constraint is real, campaign, audience, or location structure may be more precise. If activation is weak, fix onboarding and measurement before widening exclusions.

4. Choose negative match type

For Search campaigns, choose negative broad, phrase, or exact from the exclusion boundary. Google documents:

Negative broad match: all words in the negative keyword must occur in the search, in any order.
Negative phrase match: the complete terms must occur in the same order, with possible extra words before or after.
Negative exact match: the same terms must occur in the same order without extra words.

Use broad when every word in the category should be excluded, phrase when the ordered concept is the mismatch, and exact when one known query is confirmed. Negative keywords do not automatically match close variants, synonyms, or every word form. Add variants only when they express the same proven mismatch.

Do not choose broad because it blocks more. The larger blast radius can remove useful startup learning.

5. Select the narrowest scope

Use an ad-group negative for one use case or product theme, a campaign negative when all ad groups share the same audience rule, and a shared list only when all applied campaigns have the same reviewed policy. Reserve account-level rules for truly global exclusions.

An employment rule may be right for a product-led startup and wrong for a recruiting startup. A free-software rule may be wrong for a free-trial motion. A company-size rule can remove an early segment worth testing. Check positive and sibling campaign overlap.

6. Add a learning-oriented list

Confirm account, campaign or ad group, keyword text, match type, and scope. For a shared list, record purpose, owner, review date, source Search terms, blocked and preserved examples, applied campaigns, and rollback.

Google’s negative keyword management guide describes adding negatives and applying lists. Read the saved row back. A list can exist without association, or it can block a segment another campaign is meant to learn about.

Revisit terms after a plan, market, audience, landing page, product, or qualification change. Early product boundaries are not permanent.

7. Check network behavior

Confirm whether the campaign is Search, Display, Video, Shopping, or another type. Search negative broad, phrase, and exact behavior is documented by Google. Display and Video negatives behave differently and are less precise; placement, content, audience, product, or other controls may be better.

For account-level negatives, verify the current campaign and inventory scope. Do not describe an exclusion as universal while the startup is still learning which audience and inventory it can serve.

8. Verify before and after saving

Reopen the negative-keyword view. Confirm text, match type, status, campaign or ad-group scope, list association, and account-level source. Check Change history and retain the previous known-good configuration.

If an excluded query appears, check date range, application date, campaign, network, match type, word order, close variants, list association, and another negative source. Google’s troubleshooting guide supports this sequence.

If a high-fit query stops showing, inspect negative overlap before changing the positive keyword. Remove or narrow the negative, verify the trial or sales path, and record the rollback reason.

9. Monitor learning and mature quality

Compare new Search terms, spend, trials, activation, qualified demos, accepted leads, opportunities, orders, and pipeline after the normal lag. Check that the mismatch declined while useful product discovery remained available.

Set guardrails using activation, accepted pipeline, revenue, or learning coverage. Lower sign-up or click volume is not success if the startup also loses valuable learning. If results are inconclusive, stop widening the list.

Common startup setup mistakes

Adding a negative from a few noisy searches.
Treating “free,” “student,” “jobs,” “support,” “API,” or “pricing” as universal mismatches.
Optimizing clicks or sign-ups before activation or account fit is understood.
Applying a product, audience, or market rule at account level.
Ignoring positive keywords, trial flow, landing pages, content, and routing.
Assuming negative keywords match close variants or every word form.
Copying Search behavior to Display, Video, or other inventory.
Using negatives to hide onboarding, tracking, or founder follow-up problems.
Removing traffic the startup still needs for market learning.

Use a narrow, reversible exclusion that removes proven waste without claiming the startup already understands every audience.

Startup negative-keyword checklist

Confirm current offer, unwanted intent, blocked and preserved examples, Search terms range, sample size, mature outcome and lag, positive overlap, match type, campaign or list scope, network behavior, owner, review date, application date, learning metric, and rollback. Use the startup negative-keyword framework and negative keyword checklist for companion reviews.

Start with one repeated mismatch and the narrowest scope. Recheck activation, qualified demos, opportunities, or orders after the normal lag, and revisit the list whenever the startup changes its offer or market.