Shared negative keyword lists: account-wide or per campaign?

I'm not sure one big shared list is the right default. Terms like "free" or "jobs" are waste almost everywhere, so those belong at account level. The problem is that a shared list can quietly block traffic you want in another campaign.

Take a dental clinic running a brand campaign and an "emergency dentist" campaign. An account-wide negative on "emergency" would kill the second one, and nobody notices until the leads dry up.

So I'd keep a short global list for true waste and put the intent-specific negatives at campaign level. Do you run one big list and audit it, or split them like this?

Agree on the split, but I'd make the rule stricter: a term goes on the global list only if it's waste in every campaign you'll ever run. "Jobs" passes that test. "Emergency" fails it, because it's the whole point of one campaign.

There's a second reason to keep the global list short. Shared lists get forgotten. A year later nobody remembers why a term is there, and it's quietly blocking something new. A campaign-level negative breaks somewhere you're already looking, so it usually gets caught sooner.

It depends on how much your campaigns overlap in queries. If brand and emergency never see the same searches, splitting at the campaign level is clean. But if both pull in "dentist near me," the same junk click can get paid for twice, and a campaign-level negative won't stop that. For the overlapping terms, a shared list makes sense. Add a short note next to each entry saying which campaigns it's meant to protect, since that note is what prevents the forgotten-rule problem you described.

My rule of thumb is that a term only goes on the shared list when every campaign that touches it agrees it's waste. Everything else stays at campaign level. After you add anything, watch the protected campaign's impressions for a week, because a blocked intent shows up as a quiet drop, not an error.

Match type changes the blast radius a lot. A broad negative on "emergency" blocks any query with that word in it, while a phrase negative only catches that exact wording in sequence, so "tooth emergency" would still get through. For anything close to an intent you want to protect, start with phrase or exact match.

Iris, a phrase negative on "emergency" does block "tooth emergency." Phrase match needs the words in order, and a one-word phrase is always in order, so any query containing "emergency" gets blocked. The only way to let "tooth emergency" through is a longer phrase that doesn't match it.

Exact match is narrower still. On a single word it mostly blocks just the bare query and very little else, so it protects almost nothing.

For a dental clinic that wants to cut "emergency room" and "emergency vet" but keep "emergency dentist," phrase negatives on "emergency room" and "emergency vet" do the job. A bare "emergency" phrase negative would kill the campaign you're trying to protect.

Fair enough on the mechanics, a phrase negative on "emergency" does catch "tooth emergency." I had that wrong.

Whether a short phrase is safe comes down to one thing: how often the word shows up in queries you actually want. "Emergency room" and "emergency vet" are almost always waste, so negating those is safe. "Cost" is different. "Dental implant cost" could be the best converting query for a clinic that sells implants, so a phrase negative on "cost" would kill it. For words like that, pull the search terms containing the word and sort by conversions before you add anything. If the converting queries outnumber the junk, go with longer phrases. If the junk is nearly all there is, the single-word negative is fine.