A software company can report a healthy click-through rate, an acceptable cost per lead and a full demo calendar while still buying the wrong demand. That is the commercial problem negative keywords for software ads are meant to solve. They do not merely tidy up a Google Ads account. Used properly, they prevent budget being assigned to searches that were never likely to become qualified pipeline.
For B2B SaaS, the distinction matters. A search for a free template, a job, a course or a consumer tool can look relevant at keyword level while being commercially useless. If those clicks convert on a broad demo form, they can also distort automated bidding and persuade the account to find more of the same.
Why irrelevant searches become a CAC problem
Google Ads works from the signals it receives. If the primary conversion is a form completion, the platform cannot inherently know whether that person has the right company size, use case, budget or buying authority. It knows only that a conversion occurred.
This creates a common failure pattern. Broad or phrase-match keywords generate inexpensive traffic. A permissive landing page converts some of it. Cost per lead falls. Sales rejects a high proportion of the leads, but that information never reaches Google Ads in a usable form. Spend then expands into adjacent, low-intent queries because the account has been rewarded for volume rather than commercial quality.
Negative keywords interrupt that loop at the search-query stage. They are a control mechanism, not a substitute for good targeting, a relevant landing page or CRM-based measurement. Their job is to remove demand that is predictably misaligned before it consumes budget and conversion capacity.
Negative keywords for software ads: what to exclude
The strongest negatives usually come from a company’s own search-term data and sales evidence, rather than a generic list copied from another account. Context decides whether a query is wasteful.
Take “software training”. A self-serve product with a large education market may want that traffic. An enterprise platform selling annual contracts may find it produces users looking for certification or employment, not buyers. Likewise, “open source” can be a valuable competitor or a clear mismatch depending on the product and commercial model.
Start by classifying irrelevant queries into four practical groups:
- Free or low-commitment intent: free, download, template, PDF, spreadsheet, crack, torrent, open-source and trial-related terms where the offer cannot serve that expectation.
- Research, education and employment intent: course, tutorial, definition, examples, jobs, salary, interview questions, certification and university.
- Wrong product category or use case: consumer apps, adjacent tools, hardware, services your business does not provide, or a feature that is not part of the product.
- Wrong customer profile or geography: markets, languages, company types or local-service searches that fall outside the sales model.
These are starting points, not automatic exclusions. “Pricing” is often a high-intent term, even when conversion rate is lower than branded search. “Alternatives”, “comparison” and competitor names can signal an active evaluation. Blocking them to improve cost per lead may remove some of the most valuable pipeline opportunities.
The test is not whether a query is broad, informational or expensive. The test is whether a plausible buyer could move from that query to your commercial offer.
Build the list from evidence, not assumptions
Review the Google Ads Search Terms report on a regular cadence. Weekly review can make sense for high-spend campaigns or a newly launched account. Fortnightly or monthly may be enough where query volume is stable. The right frequency depends on spend, match types and how quickly search terms accumulate.
Do not assess a search term by clicks alone. A query with ten clicks and no conversions is not automatically irrelevant. It may simply lack enough data. Conversely, a single lead marked as unqualified in the CRM can expose a repeatable source of wasted demand.
For each meaningful query, ask three questions. Is the searcher looking for the product or outcome offered? Does the query fit the target customer and sales model? What happened to similar leads after sales qualification?
This produces better decisions than blanket exclusions. For example, “project management software for students” may deserve a negative if the product sells to mid-market operations teams. “Project management software for small business” may require a test instead. If small businesses occasionally become profitable customers, the better response could be a separate campaign, different message and tighter lead qualification rather than an immediate block.
Search-term reports are useful but incomplete. Google may not show every query because of privacy thresholds and low volume. That makes CRM feedback, form responses, call notes and landing-page behaviour especially valuable. If sales repeatedly reports leads seeking a different product category, look for the language they use, then review whether that language appears in keywords, ad copy or page copy as well.
Choose the right level and match type
Negative keywords can be applied at account, campaign or ad-group level, as well as through shared lists. The placement should reflect how universal the exclusion is.
Account-level negatives suit terms that are never relevant, such as jobs or torrent for most B2B SaaS advertisers. Campaign-level negatives are useful when a term is inappropriate for one product line but valid elsewhere. Ad-group negatives help preserve intent separation. A campaign targeting payroll software, for instance, may need to avoid queries meant for expense management, while another campaign may actively target that category.
Negative match types need care. Negative broad match blocks searches containing all the negative terms, even if the order changes. Negative phrase match blocks searches containing the phrase in that order. Negative exact match blocks the precise query. Unlike positive keywords, negative matching does not expand through close variants in the same way, so plural forms, misspellings and related wording may need separate review.
Use broad negatives for genuinely universal concepts. Use phrase or exact negatives when a term can mean different things in different contexts. Overly broad exclusions are a quiet source of lost demand. Adding “free” broadly, for example, may stop a relevant search such as “free payroll software trial” if a trial is part of the offer.
Connect exclusions to measurement quality
Negative keywords reduce bad traffic, but they cannot correct misleading conversion signals on their own. If a low-quality searcher can still submit a form, Google Ads may continue to value that lead unless the measurement model distinguishes it from a qualified demo or opportunity.
A more reliable setup sends downstream milestones back to the advertising platform where practical: qualified lead, sales-accepted lead, opportunity and, where volume and timing allow, closed revenue. The exact event chosen for bidding depends on conversion volume and sales-cycle length. Optimising directly to opportunities is attractive, but if there are too few of them, the platform may not receive enough signal. A qualified-lead event with consistent sales criteria can be a better interim optimisation point.
This is where negative keyword decisions become commercially measurable. Compare query groups not just by cost per lead, but by qualification rate and cost per qualified lead. If sufficient data exists, extend the view to opportunity rate, pipeline and CAC.
A simple calculation makes the issue clearer. If a campaign produces 40 leads at £100 each, its cost per lead is £100. If only four become qualified, cost per qualified lead is £1,000. Removing a set of cheap but consistently rejected queries may raise reported cost per lead while reducing cost per qualified lead. That is not deterioration. It is a correction in what the account is buying.
Avoid turning the negative list into a blunt instrument
There is a temptation to add negatives after every poor-quality lead. Resist it. Sales feedback can be noisy, especially where follow-up is slow, qualification criteria vary or a lead has not been contacted properly. Before excluding a query, look for a pattern across search terms and outcomes.
Also check whether the landing page is creating the mismatch. An ad for enterprise compliance software that sends visitors to a generic page with a broad “Book a demo” offer may attract curiosity clicks and weak submissions. A clearer page can qualify traffic more effectively through product context, target-company cues, integration requirements, implementation expectations or a form that captures useful qualification data.
The goal is not the longest possible negative list. It is a search programme where each campaign has a credible relationship between query, ad, landing page, conversion event and sales outcome.
FAQ
How often should software companies review negative keywords?
Review search terms weekly during launches, major budget changes or periods of high spend. For mature, stable campaigns, a fortnightly or monthly review is usually more proportionate. Review sooner if sales identifies a repeated source of poor-fit leads.
Should B2B SaaS advertisers exclude “free” searches?
Only if free intent is clearly incompatible with the offer. If your product has a free trial, freemium tier or useful evaluation path, excluding “free” may remove legitimate buyers. Check the qualification and opportunity rate of those queries before deciding.
Can negative keywords improve lead quality without reducing lead volume?
Sometimes, but that should not be the expectation. Removing irrelevant clicks may reduce lead volume, particularly where forms are easy to complete. The more useful question is whether qualified demos, opportunities and pipeline improve relative to spend.
What is the biggest negative-keyword mistake?
Treating exclusions as a one-off account clean-up. Search behaviour changes, match types expand reach and new campaigns introduce fresh query patterns. Negative keywords need ongoing review alongside landing-page performance and CRM outcomes.
The useful closing question is not “Which searches can we block?” It is “Which searches are teaching the account to buy demand that sales would never want?” Answer that with search-term evidence and CRM feedback, then make exclusions one part of a tighter pipeline-led search strategy.