The Search Terms Report: The First Place to Look

There is one report that answers the question every advertiser actually has — what did people type before they clicked? Everything else in the interface describes what you asked for. This one describes what you got.

It is also the report most people glance at, skim for something embarrassing, and close.

Where it is and what it is not

In the current interface it sits under Campaigns > Insights and reports > Search terms. You can also reach a filtered version by selecting keywords and using the search terms view from there.

The distinction that trips people up: a keyword is what you bid on, a search term is what a real person typed. They are only the same thing under exact match, and even then not always — Google matches close variants including misspellings, plurals, reordered words and phrasings it judges to have the same meaning. Phrase match casts wider than most people assume, and broad match is a request rather than an instruction.

So the gap between your keyword list and your search terms list is not a fault. It is the entire point of looking.

Make the report readable before you read it

Straight out of the box the report is a long list sorted unhelpfully. Three changes make it usable.

Add the columns that matter. Cost, Conversions, Cost per conversion and Conversion rate as a minimum. Impressions alone will tell you nothing about money.

Use the match type column. Each row shows how that query was matched to your keyword. A term matched by broad match that cost you money and converted nothing is a different problem from the same term matched exactly — the first is a targeting issue, the second means you deliberately bid on something that does not work.

Set a date range long enough to mean something. A fortnight of data on a small account is noise. Ninety days is usually the shortest window where patterns are real, and you should allow for conversion lag: enquiries from the last week or two will still be arriving, so the most recent days always look worse than they are.

Then sort by cost, descending. Not by impressions, not by clicks.

Reading it in four passes

Pass one: expensive and silent. Terms with real spend and no conversions at all. Do not react to a single click — one click and no enquiry is not evidence of anything. Look for terms that have absorbed the equivalent of several leads’ worth of budget without producing one. Then ask why before you act. Is the intent wrong, or is the intent right and the landing page wrong? Those lead to opposite fixes.

Pass two: converting terms you never targeted. These are the valuable rows and almost nobody looks for them. A query that produced enquiries but does not exist in your keyword list is telling you where demand is. Promoting it to its own keyword — often in its own ad group, with ad copy using its exact wording — gives you control over the bid and a more relevant ad.

Pass three: the language people actually use. Read the terms as prose rather than data. You will find the words your customers use for your service, which are frequently not the words on your website. That is free copy research for ad headlines, service pages and content alike. If a recognisable share of queries are questions, that is a content brief, not a campaign change.

Pass four: intent categories. Step back and group the list. Research queries, job seekers, people looking for a competitor, people looking for a cheaper version, people in the wrong place, people looking for a service you do not offer. Categories are more useful than individual rows, because a category tells you something about your targeting and a single row only tells you about itself.

The data is incomplete, and you should know how

Google does not show every query. Terms that have not been searched by enough people are withheld on privacy grounds and rolled into an unlabelled bucket, along with a share of the clicks and cost they generated. Google has not published the threshold or the proportion this affects, and independent estimates vary far too widely to quote as fact — the honest position is that some of your spend is not itemised, and that the share tends to be larger on broad match than on exact.

Two practical consequences.

First, the numbers in this report will not reconcile exactly with your campaign totals. That is expected, not a tracking bug.

Second, you cannot audit your way to complete coverage. If a large slice of a campaign’s spend is going to queries you cannot see, the fix is structural — tighter match types, tighter ad groups — not more time spent staring at the report.

There is a third limitation worth naming: this report only ever shows searches you appeared for. It cannot show you the demand you are missing entirely. For that you need a different instrument — auction insights, keyword research tools, or the site search on your own website.

How often, and by whom

New campaigns, or any campaign using broad match, deserve a weekly look for the first month or two. Established campaigns with settled match types can go monthly. Beyond that you are reviewing a list that has stopped changing.

Export it when you want to work properly. The report editor will build you a scheduled version that lands in your inbox, which is the difference between a task that happens and a task that is supposed to happen.

One caution about cadence: reacting weekly to small numbers is how accounts get over-managed. Half the value of this report is deciding that a row is too small to act on yet.

What it cannot do

The search terms report diagnoses. It does not treat. It will tell you that a third of your budget went to people researching rather than buying, and it will not tell you whether those people are worth reaching later with something other than a sales page.

It is also silent on the two things that most often decide whether paid search works: whether the page the click lands on deserves the click, and whether anyone answers the enquiry quickly. Neither appears in any column here.

If you want a second opinion on what your own report is showing you, that is a reasonable place to start a conversation about paid search.