.png)
A negative keyword system for Google Ads agencies is a structured, scalable process for identifying, organizing, and maintaining negative keywords across every client account in your MCC, eliminating wasted spend without strangling reach. Building one properly means moving beyond ad hoc lists into a repeatable framework with shared lists, vertical templates, scheduled audits, and quarterly pruning.
This guide walks you through how to build a negative keyword management system for Google Ads agencies that handles five clients or fifty without breaking. By the end, you will have a master library, an MCC shared list structure, a review cadence that does not eat your week, industry-specific templates, and a pruning process that keeps your negatives from doing more harm than good.
Prerequisites: You will need MCC (manager account) access to all client Google Ads accounts, search term report access across those accounts, a spreadsheet or database for your master library, and at least one media buyer who understands match types at a granular level.
Make sure every client account is linked to your MCC with full administrative access. Shared negative keyword lists only work when accounts sit under the same manager account. If you have orphaned accounts or clients who gave you standard access instead of admin, fix that first.
You should also have at least 30 days of search term data per account. If you are onboarding a new client with a fresh account, you will not have enough signal to build meaningful negatives yet. Start with your universal list (Step 1 below) and let the data accumulate before layering in client-specific exclusions.
Finally, export current negative keywords from every account into a single spreadsheet. You need to see what already exists before you build a system on top of it.
Most agencies build negative keywords the same way they did when they managed one or two accounts: a media buyer spots a bad search term, adds it as a negative, moves on. That is not a system. It is a reaction. At three accounts it works. At fifteen accounts it means your team is rediscovering the same irrelevant queries, making the same additions, and doing it on different days in different accounts with zero consistency. Shared negative keyword lists in Google Ads MCC exist precisely to solve this, but most agencies either do not use them or use them poorly.
A single account leaking 8% of spend to irrelevant queries is a manageable problem. Ten accounts each leaking 5-12% is a portfolio-level crisis that no one notices because it is spread across individual dashboards. If your agency manages $500K in monthly client spend and the average waste rate from poor negative keyword coverage is 7%, that is $35K per month in spend you are responsible for that produces nothing. Clients notice eventually.
The first type is universal negatives: terms that are irrelevant regardless of client, vertical, or campaign type. "Free," "jobs," "salary," "how to become a," "reddit." These belong in a shared list applied across your entire MCC. The second type is contextual negatives: terms that are irrelevant for one client or vertical but perfectly valid for another. "Wholesale" is a negative for a DTC brand but a primary keyword for a B2B distributor. Mixing these two types into a single list is the root cause of most over-negating disasters.
Your master library is the single source of truth for every negative keyword your agency uses. Organize it by intent category, not by client or campaign. This structure makes it reusable across accounts and verticals without constant reclassification.
These are queries from people looking to learn, not buy. Examples: "what is," "how does," "tutorial," "guide," "examples," "definition," "vs," "review" (context-dependent). Not every informational term deserves to be negated, but terms like "how to build your own [product]" or "DIY [service]" consistently attract non-buyers in commercial campaigns. Add these as phrase match negatives so you catch variations without blocking legitimate commercial queries that happen to contain one of these words.
"Jobs," "careers," "salary," "internship," "certification," "course," "degree," "training." This category is nearly universal. Job seekers and students trigger commercial keywords constantly, especially in B2B and professional services verticals. Build this list once and apply it everywhere.
This is where you need to be careful. Competitor brand names might be intentional targets in some campaigns and negatives in others. Keep competitor exclusions in a separate section of your library, clearly flagged as "apply per client strategy." Adjacent product terms, like "software" for a consulting firm or "rental" for a company that only sells, go here too.
Terms related to industries, products, or contexts that no client in your portfolio would want association with. Also include geographic negatives for businesses that only serve specific markets. A plumbing company in Dallas does not want clicks from "plumber Houston."
For agencies managing accounts across many verticals, keeping this library organized and current is one of the most time-intensive parts of Google Ads management at scale. This is exactly the type of execution bottleneck that the groas engine is designed to handle, surfacing negative keyword opportunities from patterns across massive volumes of ad spend data that no individual media buyer could manually identify.
Shared negative keyword lists are the operational backbone of agency negative keyword strategy in Google Ads. They let you maintain a list once and apply it across every account under your MCC, so updates propagate instantly instead of requiring manual changes in each account.
In your MCC, navigate to Tools > Shared Library > Negative Keyword Lists. Create a list, add your keywords, then apply it to specific accounts. Any change you make to the list, adding or removing a term, takes effect across all linked accounts automatically. Each MCC supports up to 20 shared negative keyword lists, and each list can hold up to 5,000 keywords.
Do not dump everything into one massive list. Use a tiered structure:
If a term would be irrelevant for 90%+ of your clients, it belongs in the universal list. If you can think of even one client where that term is a valid keyword, it does not belong in the universal list. When in doubt, put it in a vertical list instead. Over-applying universal negatives is the fastest way to accidentally block profitable traffic across your portfolio.
For a deeper look at structuring your MCC operations beyond just negative keywords, the complete guide to MCC management at scale covers the operational framework.
A scalable weekly audit does not mean reviewing every search term in every account every week. It means a structured rotation with priority triggers.
Here is the workflow:
The entire process should take 30-60 minutes per week for a portfolio of 10-20 accounts once the system is built. If it takes longer, your list structure is not doing its job.
Not every account deserves equal audit time. Prioritize by: highest absolute spend (more spend means more potential waste), accounts with broad match or Performance Max campaigns (these generate the widest variety of search terms), accounts that recently launched new campaigns or keywords, and accounts where CPA or ROAS has shifted negatively in the past 7 days.
Google Ads scripts can automate the search term pull and flag high-spend, zero-conversion queries automatically. The most useful script pattern: pull all search terms from the past 7 days, filter for terms with spend > $X and conversions = 0, deduplicate, and email you the list. Several open-source scripts exist for this. If you are running the groas engine across your client accounts, this detection happens continuously rather than on a weekly schedule, because the engine processes search term data in real time across every account it touches.
Vertical templates are the highest-leverage asset in your negative keyword system. They let you onboard a new client in a known industry with 80% of the right negatives already in place on day one, instead of waiting weeks for enough data to surface problems.
Common ecommerce negatives: "free," "DIY," "homemade," "used" (if selling new only), "repair," "fix," "manual," "instructions," "recall," "coupon code" (if you do not want coupon seekers), "wholesale" (if DTC only), "bulk" (context-dependent). Also negate competitor brand names unless the client's strategy includes conquest campaigns. For Shopping campaigns specifically, remember that negative keywords work differently: you cannot add negatives at the product group level, only at the campaign or ad group level. This matters when structuring Shopping campaigns for performance.
SaaS accounts attract an enormous volume of informational and free-tool traffic. Standard negatives: "free," "open source," "free trial alternative," "crack," "download," "GitHub," "template," "spreadsheet" (if selling software that replaces spreadsheets, this one is nuanced). Also negate educational terms: "course," "certification," "tutorial," "learn." For SaaS lead gen, you also want to block bottom-of-funnel terms for products you do not sell. A CRM company does not need clicks on "project management software."
Lead gen campaigns (legal, financial services, home services, B2B services) attract especially high volumes of irrelevant traffic because the query space overlaps heavily with informational searches. Standard negatives: "salary," "jobs," "how to become," "near me" (if not a local business), "free consultation" (if you charge for consultations), "pro bono," "[service] for friends and family." Geographic negatives are critical here. If a law firm only serves California, negate every other state.
Build each vertical template as a living document. Every time you onboard a new client in that vertical and discover a negative keyword that should have been in the template, add it. The template gets better with every client.
This is the step most agencies skip, and it costs them. Negative keywords do not just block bad traffic. Poorly maintained negatives block good traffic too. A phrase match negative for "free" blocks "free shipping," which is a buying signal. A negative for "cheap" blocks "cheap flights to Cancun" for a travel client where "cheap" is the entire intent. Over time, aggressive negating without pruning narrows your targeting until campaigns cannot find enough qualified traffic to spend their budgets. Impression share drops, CPCs rise as you compete in a smaller auction pool, and performance degrades in ways that look like a bidding problem but are actually a targeting problem.
Every 90 days, run this audit:
Watch for: declining impression share (eligible) with stable budgets, rising CPCs without competitive changes, decreasing search impression share specifically in Search campaigns, campaigns that consistently underspend their daily budgets, and a shrinking pool of search terms appearing in your reports. If you see three or more of these simultaneously, your negatives are probably too tight.
When dealing with accounts that already have too many keywords compounding the problem, the interaction between keyword count and negative keyword coverage gets especially complex.
The real test of a negative keyword system is whether it holds up when you go from 10 clients to 30. The key principles: your master library is the single source of truth and it lives outside of Google Ads (in a spreadsheet, Notion, or database). Shared lists in MCC are the deployment mechanism, not the library itself. Every new media buyer on your team gets trained on the system, not just handed a login. And the quarterly prune is non-negotiable, even when things feel like they are running well.
Document your list structure so anyone on your team can understand which lists apply to which accounts and why. When a media buyer leaves, the system should survive intact.
Using only broad match negatives. Broad match negatives block any query containing all the negative keyword's terms in any order, which catches far more than you intend. Default to phrase match or exact match negatives unless you have a specific reason to go broad.
Copying negative keyword lists from blog posts without review. Generic negative keyword lists circulate constantly online. They are starting points, not solutions. Every list needs to be reviewed against your specific clients' businesses before being applied.
Applying shared lists to new accounts without checking for conflicts. When you onboard a new client, review every shared list that will apply to their account. A vertical list built for DTC ecommerce clients might block queries that are profitable for a B2B ecommerce client in the same vertical list.
Negating converting search terms. It happens more than it should. A media buyer sees a term that looks irrelevant, adds it as a negative, and does not check whether it was actually converting. Always check conversion data before adding a negative.
Never pruning. Your negative keyword lists should get shorter over time in some areas as you identify over-negating, not only longer. If your lists only ever grow, you are not auditing.
Ignoring Performance Max. You can now add account-level negative keywords that apply to Performance Max campaigns. If you are not doing this, your PMax campaigns are matching to queries your Search campaigns have already excluded.
Everything described in this guide, the library, the shared lists, the weekly audits, the vertical templates, the quarterly pruning, is real work that scales linearly with your client count. Every new account means more search terms to review, more negatives to manage, and more potential for something to slip through.
Agencies using groas get direct access to a proprietary engine trained on over $500 billion in profitable ad spend. That engine identifies wasted spend patterns, surfaces negative keyword opportunities, and processes search term data continuously across every connected account, not on a weekly schedule. You connect unlimited client accounts under one subscription, keep your brand and margin, and run everything yourself with the engine doing the heavy lifting underneath.
The system described in this guide is sound. It works. But it depends entirely on your media buyers having enough hours in the week to execute it consistently across every account. The groas engine does not run out of hours.
Start your 7-day free trial and see what the engine finds in your accounts that your team has been missing. If you want to understand how groas compares to other tools agencies use to scale execution, the comparison between Optmyzr and autonomous management breaks down the differences, and the guide to white-labeling Google Ads execution without adding headcount covers the broader agency scaling question.
A scalable negative keyword system for Google Ads agencies requires five layers: a master library organized by intent type, shared MCC lists deployed in tiers, a weekly audit workflow with clear prioritization, vertical-specific templates that make onboarding fast, and quarterly pruning that prevents over-negating. Build each layer deliberately, document everything, and treat the system as infrastructure rather than a one-time project. The agencies that do this well protect client budgets, retain accounts longer, and scale without proportionally scaling headcount. The agencies that do not bleed client spend silently across dozens of accounts until someone notices. Start your 7-day free trial with groas and let the engine handle the execution your team cannot get to.
Google Ads MCC accounts support up to 20 shared negative keyword lists, with each list holding a maximum of 5,000 keywords. For most agencies, a practical structure includes one universal list, one list per industry vertical you serve, and optional lists segmented by campaign type (Search vs Shopping vs Performance Max). This tiered structure keeps your lists manageable while giving you enough granularity to avoid over-negating. If you find yourself hitting the 20-list cap, consolidate your least-used vertical lists or move rarely-applied negatives into account-level lists instead.
Account-level negative keywords apply across every campaign in a single Google Ads account, including Performance Max campaigns. Campaign-level negatives only apply to the specific campaign where they are added. For agencies, account-level negatives are ideal for universal exclusions like job seeker terms, while campaign-level negatives handle situations where a term is relevant in one campaign but wasteful in another. Use MCC shared lists for cross-account negatives and account-level negatives for client-specific exclusions that should apply everywhere within that account.
The recommended cadence is weekly search term audits with a structured rotation, plus a full pruning audit every 90 days. Weekly audits should take 30 to 60 minutes for a portfolio of 10 to 20 accounts when your system is built properly. Prioritize high-spend accounts, accounts running broad match or Performance Max, and accounts with declining ROAS or rising CPA. Agencies using groas can skip much of this manual work because the engine processes search term data continuously across every connected account, flagging wasted spend patterns as they emerge rather than waiting for a weekly review.
Yes. Google now supports account-level negative keywords that apply to Performance Max campaigns. This is critical for agencies because PMax campaigns match to an extremely wide range of queries by design, and without account-level negatives, they will serve on search terms your Search campaigns have already excluded. Add your universal negatives at the account level to ensure PMax respects the same exclusions as the rest of your campaign structure.
Default to phrase match or exact match negatives. Broad match negatives block any query containing all the negative keyword terms in any order, which catches far more queries than most people expect. For example, a broad match negative for "free trial" would block "best free trial CRM software" and "trial free offer," both of which might be relevant. Use phrase match for most negatives and reserve broad match only for terms that are clearly irrelevant in any context, like "jobs" or "salary."
Watch for these signals: declining impression share (eligible) with stable budgets, rising CPCs without competitive changes, campaigns consistently underspending their daily budgets, and a shrinking pool of search terms in your reports. If three or more of these appear simultaneously, your negatives are likely too aggressive. Run a quarterly pruning audit where you cross-reference all negatives against actual conversion data from the past 90 days. Remove or downgrade any negatives blocking terms that now convert.
Start with your universal shared list (applied automatically through your MCC), then apply your industry vertical template. These two layers should cover roughly 80% of the negatives a new client needs on day one. Then export the previous 30 to 90 days of search term data from the account, identify any client-specific irrelevant queries, and add those as account-level negatives. Agencies running the groas engine can accelerate this further because the engine draws on patterns from over $500 billion in ad spend data to identify negative keyword opportunities that manual review would miss entirely.
Not always. Negative keywords work differently in Shopping campaigns. You cannot add negatives at the product group level, only at the campaign or ad group level. Shopping queries also tend to be more product-specific, so your negative strategy needs different calibration. Keep a separate consideration layer for Shopping campaigns within your vertical templates. Universal negatives like job seeker and informational terms still apply, but product-specific exclusions need careful handling to avoid blocking relevant Shopping traffic.
Agencies using groas get access to a proprietary engine trained on over $500 billion in profitable ad spend. The engine identifies wasted spend patterns and surfaces negative keyword opportunities continuously across every connected client account, not on a weekly schedule. You connect unlimited accounts under one subscription, keep your brand and margin, and run everything yourself while the engine handles the execution underneath. The result is that the system described in this guide, weekly audits, vertical templates, pruning, runs automatically at a scale no human team can match. Start your 7-day free trial to see what the engine finds in your accounts.
