What the search terms report shows
Keywords are what you asked to bid on. Search terms are what people actually typed. Once broad match and close variants are involved the two drift apart quickly, and the report is the only place you can see how far. Each row gives the real query, the keyword it was matched to, and the clicks, cost and conversions it produced.
It is not complete. Google withholds queries that too few people searched, for privacy reasons, so part of your spend sits in a remainder you cannot inspect term by term. That gap is normal, but it means the report tells you where money went, not where all of it went — the totals in the campaign will always be larger than the totals in the list.
Why the search terms report matters
This is where waste becomes visible. Job seekers on a recruitment-sounding keyword, students researching for an assignment, people looking for a free version, searches for a competitor’s brand, enquiries from a country you do not serve — none of these look wrong in a keyword list, and all of them look obvious the moment you read the queries they triggered.
It is also the best keyword research you own, because it is your own market rather than a tool’s estimate. Queries that convert deserve their own ad groups with matching copy. Queries phrased as questions are usually telling you what a page or an article is missing. On a modest daily budget, and most budgets in Nepal are modest, a handful of irrelevant terms can quietly absorb a meaningful part of the day’s spend before anything useful happens.
Where it goes wrong
The usual mistake is reading the converting rows first. Those are pleasant but settled. The rows that need a decision are the ones with cost and no conversions, so sort by spend and work down. The second mistake is blocking too bluntly: a negative keyword added on broad match can cut off far more than the phrase you were looking at, and negatives have match types exactly as keywords do.
The third is timing. Reviewing the report once during setup and never again lets a campaign drift for months, because match behaviour changes as the system gathers data. The fourth is judging a term on a single click — one expensive click proves nothing, and blocking on that basis removes queries that would have worked.
How to act on it
Give it a fixed slot in the week rather than checking it when something looks wrong. Sort by cost, then put every row into one of three buckets: block it, promote it to its own keyword, or leave it alone because it is doing its job. Most rows belong in the third bucket, and resisting the urge to act on all of them is part of the skill.
Keep negatives in shared lists so the same waste is not blocked account by account, and check that a new negative does not conflict with a keyword you are paying for elsewhere. This weekly loop is the least glamorous and most reliable part of ongoing Google Ads management.
