A keyword audit on a typical Google Ads account turns up the same problems, account after account. Some are structural. Some are match-type errors that looked fine when set up but silently drain budget. A few are outdated practices that Google quietly deprecated years ago.
This guide covers the seven most common ones — what each mistake costs you, and the precise fix.
1. Broad match without a negative keyword list
Broad match is the most permissive match type: Google will show your ad for queries it considers related to your keyword, including synonyms, implied meanings, and conceptually adjacent searches. That's useful — when controlled.
Without a negative keyword list, broad match becomes a budget leak. A campaign targeting project management software on broad match will routinely show for free task apps, Microsoft Project tutorial, and queries nowhere near your offer.
The fix: Before adding any broad match keyword to a live campaign, build a negative keyword list at the account level (shared library → negative keyword lists). At minimum, negate your competitor brand names, non-commercial informational queries ("what is," "how does," "definition of"), and any vertical terms you can't serve. Review your Search Terms report weekly for the first month to catch missed queries early.
2. Keyword cannibalization between ad groups
Cannibalization happens when two or more ad groups bid on the same or overlapping keywords. Google chooses between them in each auction — but the Quality Score signals (CTR history, landing page relevance, expected click-through rate) accumulate separately for each group instead of pooling in one. The result: neither group builds the strong historical signals that lower your CPC, and you're effectively competing against yourself.
The most common version is an exact match keyword appearing in two ad groups with slightly different themes. Another is a phrase match keyword in one group that overlaps the intent of a broad match keyword in another.
The fix: Audit for overlap. The free tool on this site detects both direct duplicates (same keyword, two groups) and subset overlaps (a phrase in one group matching an exact in another) automatically. Once identified: consolidate duplicates into a single canonical ad group, and add negative exact match keywords to the groups that should not win that term.