The Negative Keyword List We Build in the First Week

In a new account, the cheapest improvement isn't a better advert. It's not paying for searches that have nothing to do with you.

Where to get the list, in practice

From the search terms report, read daily in the first week, not weekly. In an account launched for a client who installed interior doors, the first few days brought searches about door repair, garage doors, and a furniture manufacturer with a similar name. None of these could become an order.

Categories that almost always appear: searches with "free", "cheap", "second-hand", "repairs" when the client sells new products; "courses" and "jobs" when the client sells services; plus names of competitors we don't want to pay for initially.

What we don't exclude immediately

We don't exclude everything that doesn't convert in the first week. A few days of data tell us very little about a low-volume term. We only exclude what is clearly irrelevant to the offer. The rest stays on the watch list.

We made a mistake here with a construction client: we quickly cut a phrase that brought clicks without enquiries and, two months later, the same phrase generated enquiries in a different season. We revamped the list and left the term running, but with a separate budget.

How we organise negatives

We keep one account-level list for universal terms (jobs, courses, free) and separate lists for each campaign group, which include exclusions that depend on the offer. If mixed, you end up in the classic situation: someone excludes a word in January and in May no one understands why a new campaign isn't getting impressions.

Match type matters. A broad match negative can also stop good searches if the term is common.

What we track afterwards

We look at how much of the budget is spent on terms we didn't ask for. When that proportion decreases and the remaining clicks resemble what the client sells, we move on to the next part: the landing page and the offer.

I can't promise a result in figures — it depends on the niche, how crowded the auction is, and what happens after the click. This clean-up merely stops waste; sales are made by everything else.

If you already have campaigns running, we'll look at your search terms report and tell you what we'd exclude in the first week.

The types of searches we cut, grouped

It helps to keep negatives in categories, otherwise the list becomes a pile no one looks at anymore. For a service account, it looks something like this.

The competitor category is the only one without a fixed rule. Sometimes it's worth it, because people compare. Other times we pay for clicks from someone who wanted the other company anyway.

What the list looks like after three months

It gets thinner. We start broad, then delete exclusions that were blocking good searches. It's happened a few times that we cut a term too quickly and realised a month later that enquiries were coming from there. Since then, we look at what the exclusion produced before adding it, not just what the search looks like.

Exclusions are easy to write and easy to forget. We put a monthly review in the calendar, otherwise the list stays as it was in the first week.

What negatives don't solve

They don't fix a weak advert and don't compensate for a landing page that doesn't state what you sell. I've seen accounts with impressive lists of exclusions and a nine-field form on the page where traffic landed. The first problem was the page.

If you want us to look at what you're paying for unnecessarily right now, give us read-only access to your account and we'll send you a list of searches we'd cut first.

How often we revisit the list

Not after ticking off a list, but after a simple question: if a new client on this topic came tomorrow, would they find the answer here that would change their decision? If not, something is missing and it's usually a practical detail, not an entire section. We revisit texts after a few months, when we see what people are asking us on the phone and what wasn't written anywhere.