The Negative Keyword List Every Store Should Start With
I read the pages ranking for “negative keywords google ads” before writing this. They all explain the three negative match types, they all tell you to open the search terms report, and not one of them hands you a list. That is the whole gap. The search terms report is a record of money you have already spent, and a new campaign has no search terms report at all on the day it goes live — which is the day you most need one.
So here is the list I load into a new ecommerce account before the first click, grouped by what each group is actually protecting you from, plus the parts the ranking pages leave out: where the list goes, what changes in Shopping and Performance Max, and how you tell afterwards whether any of it worked.
Why a starter list exists at all
A search terms report is retrospective. You find the query, you exclude it, and you have already paid for the lesson. That is fine for the queries nobody could have predicted. It is a bad way to learn the ones every store learns.
Almost every direct‑to‑consumer store loses money to the same handful of query shapes: people who want the information rather than the product, people who want it free, people who want to fix the one they own, people who want a job, and people who are looking for someone else’s store. None of that is specific to your brand. It is worth excluding in advance and then letting the search terms report do the work it is actually good at, which is finding the ones peculiar to you.
Start narrow and loosen, rather than the reverse. A negative you add on day one costs you traffic you never see, and traffic you never see cannot argue back. Every group below is a default to adopt with a reason, not a rule. Read each one against your own catalog before you paste it in.
The list
Add these as phrase match negatives unless a line says otherwise. Phrase is the right default for a starter list: broad negatives are easier to overshoot with, and exact negatives will only ever block the one query you typed.
| Group | Terms | What it stops |
|---|---|---|
| Research and definition | what is, how to, how does, why is, guide, tutorial, meaning, definition, examples, vs, versus, difference between, ideas, tips | People reading about the category. Real volume, real clicks, no purchase intent. The single biggest group for most stores. |
| Free and no‑cost | free, freebie, no cost, giveaway, sample, samples, trial, download, pdf, printable | Self‑explanatory, and expensive because these queries are cheap to enter and never convert. Cut “free” if free shipping is your offer — see the warning below. |
| Employment | job, jobs, career, careers, salary, hiring, resume, internship, vacancy, recruitment | Anyone searching for work at a company in your category. Pure waste, and steady. |
| Ownership and repair | repair, fix, broken, replacement part, parts, manual, instructions, warranty, troubleshooting, how to clean, how to install | People who already own the product. Cut this one carefully if you sell parts, refills or consumables — for some stores this group is the business. |
| Secondhand and rental | used, second hand, refurbished, rent, rental, hire, borrow, thrift, pre owned | Demand for a version of your product you do not sell. |
| Marketplaces and other retailers | amazon, ebay, walmart, etsy, temu, aliexpress, target, costco, alibaba, wholesale | Navigational queries. The shopper has already decided where they are buying and it is not your site. |
| Complaints and returns | complaint, complaints, scam, lawsuit, recall, return policy, refund, cancel order, customer service phone number | Existing customers and people researching a problem. Support traffic you are paying search prices for. |
| Wrong audience | wholesale, bulk, distributor, supplier, manufacturer, dropship, dropshipping, reseller, trade account | B2B and supply‑chain intent hitting a consumer store. Skip this group entirely if you run a trade program. |
| Discount hunters | coupon code, promo code, discount code, voucher, clearance, wholesale price | The most argued‑about group on this list. Exclude it only if you do not run promotions. If you do, these are often your cheapest converting queries. |
Two more groups you have to write yourself. Homonyms: the words in your catalog that mean something else. If you sell mounts, you will pay for horses. If you sell bands, you will pay for music. Search your top twenty product nouns in a private window and you will have these in ten minutes. And adjacent products you do not stock: the sizes, materials, colors, fitments and formats a shopper will specify that you do not sell. Those two groups are the difference between a generic starter file and one that fits your store.
Negative match types do not work like positive ones
This is the part that catches people who assume negatives mirror the positive side of the account. They do not, and Google’s own documentation is explicit about it.
- Negative keywords do not expand to close variants. A positive keyword matches synonyms, related searches and variations. A negative matches what you typed and nothing adjacent to it.
- Plurals and synonyms have to be added separately. Excluding “job” does not exclude “jobs”. Excluding “repair” does not exclude “fixing”. This is why starter lists you find elsewhere quietly leak: they list one form of each word.
- Casing and misspellings are handled for you. Google accounts for both, so you do not need to add “Free”, “FREE” and “freee” as separate entries.
- Negative broad needs every word present. A negative broad of “free shipping” blocks queries containing both words in any order, not queries containing only one of them. That makes it more surgical than people expect, and it is the reason a negative broad of a single common word is so easy to overshoot with.
So when you paste the table above into a list, expand each line into its singular, its plural and the one or two synonyms your customers actually use. It doubles the length of the list, and it is the difference between a list that works and one that looks like it works.
Account, campaign or ad group
Three levels, and the choice matters more than the list does.
| Level | Use it for | Watch out for |
|---|---|---|
| Account‑level list | Everything in the table above. These are terms you never want, in any campaign, forever. Account‑level lists apply automatically across eligible Search and Shopping inventory, which is exactly what a starter list needs. | Set it once and it is invisible six months later. Keep a copy somewhere you will find it when a campaign mysteriously has no volume. |
| Campaign level | Separating intent between campaigns. Excluding your brand terms from a non‑brand campaign so the two do not compete for the same query is the common case. | Applied as a list rather than loose keywords, so you can maintain it in one place across several campaigns. |
| Ad group level | Steering traffic inside a campaign. If you have an ad group for a category and one for a specific product, negatives are how you stop the broader one from eating the narrower one’s queries. | Easy to create conflicts you cannot see. A negative here will block a positive keyword elsewhere in the same campaign if they overlap. |
One conflict rule is worth stating plainly, because it produces a campaign with no impressions and no explanation: a negative keyword always wins against a positive one. If the term you want to bid on is sitting in a negative list from two years ago, the campaign will not serve, and nothing in the interface shouts about it.
Shopping and Performance Max are a different problem
Every page ranking for this keyword writes as though the only campaign type is Search. For an ecommerce account, most of the budget usually is not.
Shopping has no positive keywords, which means negatives are the only keyword control you have over it. That makes them more important there, not less. The account‑level list applies, and it is doing a larger share of the work than it does in Search.
Performance Max is the awkward one, and the honest answer is that the controls have moved more than once. Account‑level negative keyword lists apply to it and remain the mechanism I would rely on. Campaign‑level negative keywords have been rolling out, so check your own account rather than trusting any article, this one included, about what is available in the interface today. For excluding your own brand queries, brand exclusions are the right tool instead of negatives — they are built for that job and they cover the whole campaign rather than one query shape at a time.
The reason this matters for a store: if you run Performance Max alongside a branded Search campaign, the two will compete for your own brand queries and you will pay for traffic you would have got anyway. That is a bigger number than anything in the starter list, and it is a structural fix rather than a keyword one.
The negatives that cost people money
Negative keywords are the cheapest lever in an account, which is why they get pulled too hard. Three mistakes I find repeatedly:
- Negating a word that appears in your own product names. “Free” is the classic: exclude it as a phrase negative and you have also excluded every search containing “free shipping”, “sugar free”, “fragrance free” and “gluten free”. Search your catalog for every negative you are about to add before you add it.
- Negating brand terms in the wrong direction. Excluding your brand from a non‑brand campaign is correct. Excluding it account‑wide takes your branded campaign down with it, and the drop looks like a tracking problem rather than a self‑inflicted one.
- Negating an entire intent that was converting. “Cheap” and “discount” are the usual casualties. If price is your position, those queries are your customers. Check the conversion data for a query group before you exclude it, not after.
A sensible habit: never add more than one group of negatives at a time to a campaign that is already spending. If performance moves, you want to know which change moved it.
How to tell whether it worked
Nobody ranking for this keyword tells you how to measure this, which is convenient, because “negative keywords save you money” is then a claim that never has to survive contact with a report. Understand first that a negative keyword does not reduce your spend. It moves it. The budget still goes out; it goes out somewhere else.
| What to look at | What it tells you | When it is readable |
|---|---|---|
| Count of distinct search terms | Whether the exclusions are actually biting. Should fall, visibly, within days. | Days |
| Share of cost going to converting queries | The real measure. Take cost against queries with at least one conversion as a fraction of total cost, before and after. | Two to four weeks |
| Click‑through rate | Usually rises, because you have removed queries that were never going to click. A weak signal on its own. | One to two weeks |
| Conversion rate | The one people quote. It rises almost mechanically when you remove non‑converting traffic, so it flatters the work. | Two to four weeks |
| Total conversions and revenue | The one that decides whether it worked. If conversion rate went up and conversions went down, you over‑negated. | Three to six weeks |
| Impression share lost to budget | Whether the freed budget went anywhere useful. | Weeks |
Watch the last two together. Conversion rate up and revenue flat means you cut cost, which is worth something. Conversion rate up and revenue down means you cut demand, which is worth less than nothing. That pair is the entire test, and it is the reason to add negatives in groups rather than in one heroic afternoon.
What to do first
If you have an hour, in this order.
| Step | Why it is here | Time |
|---|---|---|
| Export the last 90 days of search terms with cost and conversions | Your baseline. Without it you cannot demonstrate anything afterwards. | Minutes |
| Check existing negative lists for terms you now want to rank for | Finds the campaign that has been quietly throttled. Most valuable single step in an inherited account. | Minutes |
| Build the account‑level list from the table above, in phrase match | Covers Search and Shopping in one place. | Under an hour |
| Expand every line to plurals and synonyms | Negatives do not expand on their own. Skipping this is why starter lists leak. | Minutes |
| Add your homonyms and the products you do not stock | The half of the list that is specific to your catalog, and usually the half that saves the most. | An hour |
| Separate brand from non‑brand at campaign level | Stops your campaigns bidding against each other. | Low |
| Decide how Performance Max handles your brand queries | Brand exclusions, not negatives. Bigger number than the whole starter list. | Low |
| Set a recurring search terms review | The starter list handles the predictable waste once. The rest arrives weekly. | Ongoing |
The starter list is a one‑off that stops being interesting the moment it is loaded, which is why it belongs on day one. The ongoing discipline is the search terms report, and it works better when it is not being used to relearn that people search for jobs.
Negatives are also the cheapest half of a duller problem: a lot of paid traffic is wasted after the click rather than before it, which is a Shopify speed and landing page problem as often as a keyword one. The rest of what I write about Google Ads for ecommerce starts from the same place: the cheapest fix first, whether or not it looks like strategy.
Questions people ask
What are negative keywords in Google Ads?
How many negative keywords should an account have?
Where should negative keywords go — account, campaign or ad group?
Do negative keywords work with Performance Max?
Want this run properly?
I am Greg Asuncion, an ecommerce paid media consultant. I run Google Ads, Meta, Amazon and Microsoft for direct‑to‑consumer brands on Shopify — one person, published pricing, and a free audit after a call. Here is how I run Google Ads for ecommerce brands.