GOOGLE SHOPPING NEGATIVE KEYWORDS: THE REAL TARGETING LEVER

Every advertiser who moves from Search to Shopping asks the same thing: where do I put my keywords? The answer reframes the whole channel. You don't pick the queries you show for, Google matches your feed to searches, which means the only direct control you own over query traffic is subtractive. That's why I treat google shopping negative keywords not as hygiene but as the primary targeting mechanism. In a channel with no positive keyword field, the list of searches you exclude is the closest thing you have to steering.

This is the article I give clients who think negatives are a minor cleanup task. In my experience managing e-commerce accounts, a disciplined exclusion list moves ROAS more than almost any other single lever, precisely because it's the one place where you, not the algorithm, decide. I'll cover how to add negative keywords to a shopping campaign in both Standard Shopping and Performance Max, the starter list I paste before I have any data, and the match-type and review routine that keeps the whole thing sharp.

Think of it as inverted targeting. In Search you build toward the queries you want. In Shopping you carve away the queries you don't, and what's left is your real audience. Once that clicks, negatives stop feeling like a chore and start feeling like the steering wheel they actually are.

Why negatives are the real targeting lever in Shopping

Google is explicit that Shopping ads use your Merchant Center product data, not keywords, to decide how and where to show. That single design choice is the reason negatives carry so much weight: if you can't add positive keywords to aim your ads, the excluded set becomes your only precise, query-level control. Everything else, titles, product types, categories, shapes eligibility indirectly. Negatives act directly on the query, which is why I lean on them harder in Shopping than I ever did in Search.

The mechanic worth internalising is that negatives are subtractive. They don't choose the pool of searches your products match into, they trim it. Google first decides your feed is relevant to a search; a negative keyword then vetoes that match. This is exactly why they can coexist with a keyword-less targeting model, and why the absence of a positive field, covered in depth in why Google Shopping keywords work differently, doesn't leave you powerless. You still get to say no, and saying no is most of the job.

In practice, the accounts where I've turned Shopping around fastest were rarely fixed by a feed rebuild. They were fixed by reading three months of ignored search terms and cutting the dead weight, mismatched intent, wrong products, tyre-kickers. On several accounts that single pass reclaimed 20-30% of spend that had been quietly leaking into queries the store was never going to convert. That's the lever, and most advertisers barely touch it.

How to add negative keywords to a shopping campaign

Mechanically, to add negative keywords shopping campaign work splits by campaign type. In Standard Shopping you have a true campaign and ad-group negative field. Open the campaign, go to the Keywords section, select the negative keywords tab, click the plus button and add terms one per line, then optionally save them to a shared negative list you can reuse across campaigns. Google's own walkthrough on how to add negative keywords to campaigns is the reference I'd keep open the first few times.

Performance Max is more restricted but no longer the black box it was. You can add campaign-level negatives directly in the PMax negative keyword tab, and Google has steadily raised the ceiling here over the last year. On top of that, account-level negatives automatically apply across all Search and Shopping inventory, including PMax, up to 1,000 terms. That account-level list is where I put universal junk, free, jobs, adult intent, so it never triggers any campaign regardless of type.

My rule of thumb on placement: account level for terms that should never fire anywhere, shared lists for exclusions that repeat across a product line, and campaign or ad-group level for the specific, one-off negatives a single campaign throws up. Keeping those tiers clean is what stops the classic disaster, someone blocking a bestselling brand three years ago in the wrong list and nobody noticing the revenue quietly vanish. If you're still weighing whether to touch the positive side at all, the piece on what you actually add to a Shopping campaign explains why negatives end up doing the heavy lifting.

My starter negative keyword list for Shopping

Here's the list I actually paste into a fresh Shopping account before I have a single search term of my own. It's not a finished list, it's a set of hypotheses that hold on almost every e-commerce account, ready to be confirmed and expanded once real data lands. I group my google shopping negative keywords into intent buckets, because that's how the waste tends to cluster.

Free and cheap intent

Terms like free, gratis, cheap, cheapest, discount code, coupon. These pull bargain hunters and freebie seekers who rarely convert at a healthy margin. On most stores this bucket alone accounts for a meaningful chunk of wasted clicks in the first month.

DIY and information intent

Terms like how to, DIY, plans, tutorial, instructions, template, PDF. People researching how to make or fix something aren't buyers yet. Unless you sell the guide itself, this intent almost never belongs in a Shopping campaign.

Second-hand and repair intent

Terms like used, second hand, refurbished, repair, spare parts, replacement. If you only sell new full-price product, these queries are a mismatch that Google will happily serve you into if you let it.

Research, jobs and reviews

Terms like review, reviews, complaints, vs, job, jobs, salary, wiki, meaning. This is pure non-transactional intent, plus jobseekers who found your brand for the wrong reason. Cheap to exclude, expensive to leave running.

Then there's the niche-specific layer: competitor and wrong-brand terms, adjacent products you don't stock, and any locale or spelling variant irrelevant to your catalogue. That part you build from your own account, and it overlaps with the broader Google Shopping fundamentals that determine which searches you were eligible for in the first place. The four buckets above are universal; the niche layer is yours to discover.

Match types: don't accept the exact-match default

This is where a lot of starter lists quietly underperform. By default Google adds new negatives as exact match, which only blocks that precise query with no extra words. Add free as an exact-match negative and you'll still show for "free shipping running shoes". For broad junk intent I use phrase match instead, so the whole family of searches containing the term is excluded, which is what you actually want from a word like free or DIY.

Phrase match is my default for most Shopping negatives because it catches variations without the collateral damage of broad. I reserve exact-match negatives for the narrow case where a term is fine in one context but wasteful in another, a model number that's valuable on its own but junk when paired with "manual", say. Getting this wrong in the over-blocking direction is the more dangerous failure, so I review every negative's match type before I hit save rather than trusting the default.

One more caution specific to Shopping: because you can't add positive keywords to compensate, an over-aggressive negative has no counterweight. In Search a broad negative that clips a good query can be offset by a tighter positive keyword. In Shopping there's nothing to offset it, the query is simply gone. That asymmetry is why I'm deliberate here and why match-type discipline matters more than in a Search account.

The weekly search terms report routine

The starter list gets you to launch; the search terms report keeps you profitable. My routine is a short weekly pass: open the Shopping search terms report, sort by cost or clicks descending, and scan for queries burning budget with no conversions. Anything clearly off-intent becomes a negative on the spot, at whatever tier makes sense, account, list or campaign. This single habit does more for Shopping efficiency than most of the dashboard-watching people spend their time on.

The report is also the honest read on how your feed is being matched, which is why I treat it as the trigger for every exclusion rather than guessing. It slots directly into how I approach optimising a Shopping campaign end to end: negatives are the fastest, highest-leverage move in that loop, and the one I check first when performance drifts. Search behaviour changes, new irrelevant queries appear, and a list left untouched slowly loses its edge.

If I had to compress the whole thing: in Shopping you don't add keywords, you subtract them, and that subtraction is your targeting. Launch on a solid starter list, run a weekly search terms pass, choose match types deliberately, and keep your account, shared and campaign tiers clean. Do that and negatives stop being cleanup and become the steering they were always meant to be, the one lever a keyword-less channel actually hands you.