You turned on Google Ads with a list of the queries you want to win. Two weeks later, the budget is gone, the account reports a respectable number of clicks, and the two inquiries that came through wanted something you do not sell. Nothing broke. The account did exactly what you asked it to do, and you asked for only half of what you meant.
The keyword list you build is not the list Google Ads buys.
Your list is a request. Match types decide how loosely that request is read, and the looser the setting, the more of the auction you have agreed to enter without saying so.
Adjacent is a wide word. It covers people looking for a job at a company like yours, students writing about your category, buyers hunting for the free version of what you charge for, and anyone comparing two competitors who have never heard of you. None of those people are wrong to be searching. They are just not yours.
For a company with a large budget, that traffic is rounding. For a startup, it can be most of the test.
Negative keywords decide what a small budget can teach you.
The first paid campaign at seed stage is not there to produce revenue. It is there to answer one question: do people who type this problem into a search box respond to how we describe the answer? That question needs a sample, and a sample costs money you’ve already spent.
If half your spend goes to queries that were never going to convert, you do not get half an answer. You get an unreadable one. At the end of the month, you cannot tell whether the channel failed, whether the page failed, or whether you bought a few hundred visits from people who were never buying anything. That is the real cost of skipping the exclusion list, and it is paid in weeks, not in money.
A startup’s search terms report is positioning research.
The report showing which queries actually triggered your ads is the closest thing a young company gets to hearing its buyers think out loud. These are the words real people use for the problem, chosen before anyone taught them your product name or your category label.
Read it for the exclusions, then read it again for the language. Read it that way, and you will usually find a phrase you would never have written yourself, and that phrase belongs on the homepage. If the words in that report don’t match the words on your site, the ads aren’t the problem. That is a positioning problem wearing a paid search costume.
Build three exclusion lists in the first week.
The first is the never list: jobs, careers, salary, internship, free, crack, torrent, tutorial, course, and the word “meaning”. Few of the people typing those are buying anything this quarter, and most accounts attract them anyway.
The second is the not-us list, and it takes thought. Name the adjacent categories you are constantly mistaken for, the market segment you deliberately do not serve, and the version of your product you decided not to build. Excluding those queries is the same decision as your positioning, written where it costs money to be vague.
The third is competitor brand names. Decide deliberately whether you bid on them; most seed-stage startups should not, because you pay a premium to interrupt someone who already has a preference. Whichever way you decide, decide it once and write it down, rather than letting the match types decide for you.
Exclusion is a weekly job, not a setup task.
New queries appear every week because people keep inventing new ways to describe their problem. An account left alone for a quarter drifts back toward the internet average.
Fifteen minutes on a Friday covers it: open the search terms report, exclude what does not belong, and copy anything interesting into whatever file holds your messaging notes. That is the whole practice. It is unglamorous, it is rarely the thing an agency puts in a deck, and it separates a paid account that teaches you something from one that quietly spends.
If you are still deciding whether paid search deserves any of this month’s money, the budget question comes first, and the answer is often no. When we run paid search for a startup, we write the exclusion list before the keyword list, because the exclusions are the argument about who the company is for and the keywords are only the shopping.
Tell us what you’re building.