A negative keyword tells Google not to show your ad for that search. Add nothing and you pay for every search Google's broad-match algorithm decides is "related". For a tradie advertising on "emergency plumber", that includes "plumber salary", "plumber jobs", "plumber TAFE course". You pay $8 a click on every one.
We add 40 negatives to every new account before the first ad runs. Below is the list, grouped by cluster.
The "free" cluster — 12 terms
People searching with "free" don't want to pay you. They want a freebie, a quote, a comparison, or nothing at all.
- The block : free, gratis, no cost, complimentary, freebie, trial, sample, demo, freeware, freelance, freelancer, voluntary.
The "DIY" cluster — 10 terms
Searches with DIY intent want a YouTube tutorial, not a tradesman. We can't compete with free.
- The block : DIY, do it yourself, how to, tutorial, guide, instructions, YouTube, step by step, learn, course.
The "jobs and salary" cluster — 10 terms
The number-one budget burner on tradie accounts. Someone searching "electrician apprenticeship Brisbane" is not a customer.
- The block : jobs, careers, hiring, salary, wage, pay, employment, apprenticeship, TAFE, trade school.
The "informational" cluster — 8 terms
People in research mode aren't ready to ring you. Block them at the click. Most of these searchers are months away from buying, if they ever do. Pay for the ones at the bottom of the funnel instead.
- The block : what is, definition, meaning, history, statistics, wikipedia, reddit, forum.
Add these as phrase-match negatives — exact-match misses too many variants. Phrase-match catches "what is the cost of" and "what's the definition of" without blocking your legitimate keyword "cost of windscreen repair".
The weekly maintenance rhythm
The 40 terms above are the static base. The dynamic work is the search-term report. Every Monday, 15 minutes — pull the search-term report for the last seven days, scan everything with three or more clicks and zero conversions, decide whether to add it as a negative. The same logic runs across every paid media account we manage.
This single ritual is worth more than most "optimisation" work agencies bill for. On Novus Glass, the first three months of weekly search-term reviews stripped 180 negatives into the account. Cost per conversion dropped 52% inside 90 days. The list, not the bidding strategy, did most of that.
What we won't do — and why it matters here
We won't load up a generic 500-term negative list from some forum. Every list we deploy is reviewed against the client's vertical. A boilerplate list will block searches you want.
For example, "second-hand" is on most generic lists as a negative. For a glass repair business, "second-hand windscreen" might be a legitimate enquiry. Context matters every time. This is the same logic behind how we run paid media without taking a cut of spend — when we don't earn more by spending more, the conversation stays on what works.
The takeaway
Forty negatives at start, 15 minutes a week thereafter. That's the maintenance. If your current agency doesn't have a documented negative list and a weekly cadence to grow it, ask why. The answer will tell you what kind of relationship you've got.