Most negative-keyword advice is reactive: launch the campaign, wait for the search terms report to fill up, then negate the junk. That workflow is correct but incomplete, because it makes you pay for the first wave of irrelevant clicks before you are allowed to block them. The queries that show up in week one — the job-seekers, the freebie-hunters, the DIY researchers — are predictable. Paying to discover them in your own report is paying for information you already had. A pre-launch negative list is how you keep that money.
Building the list proactively saves you from paying for the obvious misses while you wait for data to accumulate, and it keeps month one's report readable instead of drowning the real signal in noise you could have predicted. This post lays out a tiered pre-launch list — the universal blockers, the campaign-specific layer, and the handoff to reactive review — plus the one trap that turns a defensive list into a self-inflicted delivery problem.
Why reactive-only management loses month one
A purely reactive approach has a built-in lag: the search terms report only tells you about queries that have already spent your money. By the time a wasteful term has accumulated enough clicks to be visibly bad in the report, you have paid for those clicks. For predictable junk — the query patterns that waste money in nearly every account — that payment is avoidable. You do not need data to know a lead-generation campaign should not serve on "free" or "jobs"; you know it before launch.
The lag is worse in 2026 because match types cast wider intent nets than they used to, so a new campaign pulls in more query variety faster. A broad or phrase keyword with no negative floor can spend its first-day budget on adjacent intent before you have even opened the report. The reactive workflow then spends your second week cleaning up what the first week bought. A pre-launch list flips the order: block the known-bad before it spends, so the report you review on day seven is mostly queries worth a real decision. It turns the first wasted-spend audit from a rescue into a refinement.
Tier 1: the universal seed list
Start with the intent mismatches that waste money across almost every advertiser. The core set is well established: free, cheap, discount, jobs, careers, salary, internship, DIY, tutorial, how to make, sample, template, and course. These flag people who are not buyers — job-seekers, freebie-hunters, and do-it-yourselfers — and blocking them costs you nothing in real demand. If you sell rather than inform, add research-destination terms like reddit, quora, and wikipedia, which usually signal someone looking for discussion rather than a purchase.
Keep this tier as a shared negative list applied at the account or shared-library level so every new campaign inherits it automatically on launch. That is the whole point of structuring negatives across account, campaign, and shared lists: the universal blockers should never be re-typed per campaign, they should be a floor every campaign stands on from its first impression. Maintain it as a living document — when you spot a new universally-junk pattern in any account, it graduates into the seed list for all future launches.
Tier 2: the campaign-specific layer
The universal list handles junk that is wrong for everyone; tier two handles what is wrong for this specific campaign. Before launch, write down what the campaign should and should not attract, then translate the "should not" into negatives. If you sell premium software, block the free and open-source alternatives by name. If you serve one region, block the location terms for regions you cannot service. If your offer is one product in a category, block the adjacent products in that category you do not sell — the searches that look relevant to Google's intent-matching but are irrelevant to your actual offer.
This is also where you block competitor and adjacent-intent traffic you have decided not to bid on. The discipline is to base tier two on what you know about your offer, not on guesses about performance — you are fencing intent, not predicting winners. Match type matters here: use phrase and exact for most of these, and reserve broad-match negatives for single words you are certain about, because a broad negative blocks any query containing all its words in any order and can quietly suppress good traffic. When a campaign-specific pattern is broad enough to warrant it, group it deliberately — the approach the pattern-grouping versus single-negatives post works through.
The over-filtering trap
The failure mode of an enthusiastic pre-launch list is under-delivery. Adding too many exclusions before you have data can starve a new campaign, cutting out search terms that might have converted with a lower bid, in a different season, or simply after more time to learn. Pre-launch negatives should block intent you are certain is wrong, never queries you merely suspect might underperform. Uncertainty is not a reason to negate — it is a reason to wait for the report. The list you launch with should be confident, not exhaustive.
This is the pre-launch version of the same problem the over-negating post covers: negatives that strangle the traffic you are paying to attract. It is more dangerous before launch because you have no baseline to notice the damage against — a campaign that never delivered looks like a targeting or bid problem, not an over-blocked one. Guard against it by keeping tier two tied strictly to known-irrelevant intent, and by leaving the borderline calls for after you have real search-term data to judge them on.
From launch list to living list
The pre-launch list is tier one and tier two; the report supplies tier three. Once the campaign has spent enough to produce a meaningful search terms report, the workflow becomes the reactive one — but now it is refinement rather than triage, because the predictable junk was never let in. Review the report on your normal cadence, negate the wasteful queries the pre-launch list did not anticipate, and promote the converting surprises into managed keywords. The pre-launch list did its job by making that first review short.
Close the loop by feeding what you learn back up the tiers. A wasteful pattern that shows up in this campaign's report and would apply to others belongs in a shared list; one that is genuinely universal belongs in the tier-one seed for every future launch. Over time the seed list compounds — each account you run makes the next launch cleaner — and the search terms report shifts from a cleanup tool to a discovery tool, because the noise it used to be full of is now blocked before it ever costs you a click.