How can negative keywords block traffic I actually wanted?
Negative keywords do not behave like positive ones. A negative broad term blocks any query containing all of its words in any order, negative phrase blocks that exact word sequence, and negatives do not expand to close variants — so plurals and misspellings need their own entries. Most accidental blocking comes from single-word negative broad terms that appear inside perfectly good queries.
Negatives are stricter and dumber than positives
Two asymmetries cause almost all the trouble. First, negatives do not match close variants: adding a negative for one spelling leaves the misspelling and the plural running. Second, a negative broad keyword blocks any search that contains all of its terms, regardless of order or what else is in the query.
That second rule is why a single-word negative is a blunt instrument. It blocks every query the word ever appears in, including the ones that would have booked.
The words that most often overblock
- free. Intended to block freebie hunters, also blocks free estimate and free inspection, which are frequently your best-converting phrasing.
- cheap and price. Blocks price shoppers and also blocks people asking a legitimate replacement price question who are ready to buy.
- jobs, salary, hiring. Usually correct, but they will also block queries where a homeowner mentions a job in the ordinary sense.
- parts and manual. Blocks DIY research, and also blocks people searching for a part they have concluded they cannot install themselves.
- A competitor's name that is also a common word. Blocks far more than the competitor.
The list-level failures
Beyond match types, the structural problems are account-level negative lists applied to campaigns nobody intended, ad-group negatives added years ago to stop internal competition and never revisited, and lists inherited from a previous agency with no documentation of why anything is on them.
These are hard to detect because blocked traffic is invisible. There is no report of searches you did not receive. The absence has to be tested for rather than observed, which is why negative lists accumulate quietly for years.
A test you can actually run
Take twenty queries you know are valuable — your best converting terms, drawn from real bookings rather than from a keyword tool — and check each one against the negative lists applied to the campaign that should serve it. The keyword diagnosis tools will tell you whether a specific query is blocked, and doing this by hand for a short, high-value list catches more than a general audit.
Sourcing that list from actual booked jobs rather than from the platform is the part that makes it useful. Call analysis gives you the phrasing customers really used, and joining it to which of those calls became revenue tells you which phrasings you cannot afford to block.
Maintenance rules that keep it clean
Prefer negative phrase to negative broad for anything ambiguous. Document a reason next to every negative you add. Review lists at least seasonally, since a term that was junk in spring may be demand in autumn. And keep a small, deliberate list rather than an enormous inherited one — long negative lists are usually a symptom of a match type problem rather than a solution to one.
Topics: negative keywords · match types · wasted spend · account audit
Have a version of this question about your own business?
The useful answer usually depends on which systems you run and how they're connected. That's a conversation, not a blog post.