How to Set Up Google Ads Match Types for Agencies

Set up Google Ads match types for each client from the offer and qualified outcome outward. Build a small, coherent keyword and ad-group structure; choose broad, phrase, or exact for a documented reason; add negative guardrails; and give one person ownership of Search terms review. Google supports unformatted broad keywords, quoted phrase keywords, and bracketed exact keywords, but the syntax is only one part of a safe agency setup.

This workflow is educational. Use Google’s keyword matching guide and Add keywords while checking the current interface and behavior in the client account.

Define the client decision before touching keywords

An agency setup should answer a business question, not merely fill an ad group with variants. Start with the client’s offer, audience, service area, buying stage, and qualified outcome. An ecommerce client may care about profitable orders; a local provider may care about booked jobs; a SaaS client may need activated accounts; and a B2B client may need accepted opportunities rather than completed forms.

Write down the event that will guide decisions, how it is counted, and how long it takes to mature. Include the conversion action, value rule, CRM or commerce stage, owner, response-time expectation, and any margin or capacity constraint. If the account is optimizing for a convenient but weak event, the agency cannot reliably decide whether broader discovery is helping the client.

Also record what the client has approved. A match-type change can alter eligible searches, spend distribution, landing-page traffic, and the Search terms review load. The account manager should know whether the client wants discovery, protection of a high-value theme, a clean routing test, or a controlled recovery from wasted spend.

Map the account before choosing syntax

Read the existing campaigns, ad groups, keywords, landing pages, negative lists, locations, audiences, budgets, bidding settings, and conversion actions. Note which parts are live, which are experimental, and which are inherited from a previous agency. Do not assume that a shared negative list or a copied keyword structure means the underlying client offers are the same.

Create an intent map with three columns:

searches that clearly describe the client’s offer and should be eligible;
related searches that may be useful but need query and outcome review; and
searches that are outside the offer, customer, geography, buying stage, or business model.

Then map each important theme to one page and one ad-group promise. Separate acquisition from jobs, support, training, research, comparison, and used or replacement intent when those themes lead to different outcomes. If two client campaigns could answer the same query, define the routing rule before adding keywords. A match type cannot repair an ambiguous campaign structure.

Build a keyword and ad-group plan

Start with a small seed set that expresses the client’s real products or services. Include the language customers use, but keep each ad group coherent enough that the ad and landing page can answer the theme. A long list of loosely related synonyms makes review and reporting harder, especially when the agency manages many accounts.

For every seed, record the client theme, campaign, ad group, landing page, proposed match type, reason for that type, negative categories, expected mature outcome, owner, approval status, and rollback. This sheet is more useful than a bare keyword export because it preserves the decision behind the setup.

Use a naming convention that exposes the account’s intent without pretending that naming itself controls matching. For example, a service campaign might separate non-brand discovery, high-value exact routing, and brand protection; the exact names matter less than consistent ownership and documentation.

Before implementation, look for duplicates and near-duplicates across sibling campaigns. A keyword that seems harmless in one client account can compete with another theme, receive the wrong page, or make a later Search terms decision difficult to explain.

Choose broad, phrase, or exact deliberately

Google’s input syntax is simple: broad match uses the keyword without special formatting, phrase match uses quotation marks, and exact match uses brackets. For a seed such as emergency plumber, the entries would be emergency plumber, "emergency plumber", and [emergency plumber]. The decision behind the syntax needs more care than the punctuation.

Use broad for supported discovery

Broad match can reach searches related to the keyword, including searches that do not contain the same words. Google describes signals such as the landing page, ad-group keywords, a person’s previous searches, and other ad-group settings. That can help a client discover useful language, but it also makes page relevance, conversion quality, and Search terms ownership important.

Choose broad for a theme when the landing page is specific, the ad group has a clear promise, the account has a trustworthy conversion or value signal, and the agency can review resulting queries. If the client has low volume, weak qualification, a narrow service area, or little capacity for waste, start with a more controlled scope and earn the right to broaden it.

Use phrase for a defined concept

Phrase match can preserve an ordered product or service concept while allowing additional words to qualify the search. It can be useful for a client’s named service, product category, or problem where words such as a location, urgency, industry, or model matter.

Phrase is not a promise that only the literal characters will be eligible. Review actual queries and the page they receive. A phrase keyword can still bring a search with the wrong buying stage or business model, so the agency needs negative categories and an owner for the review queue.

Use exact for an understood route

Exact match is useful when a known, high-value intent needs a distinct ad, page, budget, or close inspection. It can match same-intent queries and close variants under current Google Ads behavior, so exact is a control choice, not a literal-only guarantee.

Use exact for specialized products, branded or named offerings, a routing test, or a theme where the client wants transparent review. Do not call exact the “best” type by default: extra control can also reduce qualified discovery and make learning slower. Compare the client’s qualified outcome and not only impression share, click-through rate, or cost per click.

Record the reason, not just the type

In the agency worksheet, write a sentence such as: “Phrase preserves the core service concept while we check location and urgency language,” or “Exact routes this enterprise product to its dedicated page while sales validates fit.” A future account manager should be able to understand the choice without asking who made it or copying a neighboring client’s policy.

Add keywords in a controlled implementation

Open the correct client account and campaign, then verify the campaign type and ad group before adding anything. Google’s current keyword addition instructions describe going to Keywords, selecting the add control, choosing the campaign and ad group, entering or pasting keywords, and saving them. The labels can change, so use the account’s current interface rather than relying on an old agency video.

For a small build, add the reviewed keywords directly and inspect each row after saving. For a larger approved build, prepare a controlled upload or bulk edit with the same columns used in the worksheet. Keep a copy of the proposed state, the approver, the implementation date, and the rollback instruction.

Before saving, check the following:

the campaign has the intended location, language, budget, and network settings;
the ad group has one coherent promise and the selected landing page answers it;
each keyword has the intended punctuation and no accidental leading or trailing characters;
duplicates, misspellings, and unrelated synonyms have been removed; and
the conversion action and value settings match the client brief.

Do not use an agency template as an excuse to paste the same match-type mix into every account. A template can standardize fields and review gates; the client’s offer, page, audience, economics, and response capacity must determine the actual entries.

Add negative keywords with scope and review ownership

Negative keywords should block clear irrelevant intent while preserving the client’s eligible products and services. Start with categories such as jobs, training, support, free, used, wholesale, do-it-yourself, unrelated brands, wrong locations, or other business-model exclusions only when they are truly outside the client’s offer.

Google’s negative keyword documentation explains that negative broad, phrase, and exact matching differs from positive keyword matching and does not automatically cover close variants in the same way. Select the negative type intentionally, record its scope, and check whether it belongs at the campaign, ad-group, or shared-list level.

The person reviewing Search terms should be able to trace each exclusion to a query and a client decision. Before applying a shared list, compare it against the client’s product catalog, service area, locations, partner names, and campaign routing. A negative that protects one offer can block a legitimate offer in another campaign.

QA the setup before launch

Treat a match-type build as a routing change and run a client-ready checklist before enabling it. Inspect the keyword row, ad, page, conversion action, and campaign settings together. A correctly punctuated keyword still creates a bad experience if the ad promises a product the page does not sell.

Verify:

the campaign and ad group are enabled only when the client has approved them;
the keyword sends the intended theme to the intended page;
the page works on mobile, presents the offer and service area clearly, and exposes the intended form, call, checkout, or signup action;
primary and secondary conversions are distinguished in reporting;
location, language, schedule, budget, network, and audience settings match the brief; and
negative keywords do not block the client’s own brands, products, locations, or legitimate variants.

Preview representative searches from the eligible, review, and excluded groups. This is a reasoning check, not a promise that every future query can be predicted. Record what should happen and then use the Search terms report after traffic arrives to compare reality with the plan.

If the client’s sales or support team handles a lead, include a response test. Confirm where the lead appears, who owns it, what qualifies it, and how a sales outcome returns to the advertising account. Without that loop, the agency may choose match types based on cheap form fills while the client experiences poor pipeline quality.

Run Search terms review as a client process

The initial build is only the first decision. Assign an owner, review cadence, approval threshold, and escalation path before launch. A small local account may need a weekly review; a high-volume ecommerce account may need a more frequent queue; a low-volume B2B account may need a longer observation window so that a single query does not drive a premature policy change.

For each meaningful query, classify the intent and record:

the matched keyword, match type, campaign, and ad group;
the query’s product, service, location, audience, and buying stage;
spend, clicks, conversions, and the relevant qualified or revenue outcome;
whether it should remain eligible, become a deliberate keyword, or be excluded; and
the client decision, approver, implementation date, and follow-up owner.

Use Google’s Search terms and negative keyword guidance when reviewing exclusions. Add a negative for a clear mismatch, but inspect its scope and matching behavior first. Promote a useful query to a keyword only when it deserves a distinct bid, page, ad, budget, or reporting question; otherwise, keep the broader theme and document the reason.

Do not change match type, budget, landing page, bid strategy, conversion action, and ad copy in the same unlabelled edit. If several changes are necessary, describe the work as a combined intervention, preserve the prior state, and set a review date. That makes the client report honest about what caused a result.

Adapt the setup to the client’s business model

The same syntax can support different operating constraints. Make the account type a planning input, not a replacement for account evidence.

Ecommerce accounts

Group keywords around products or categories that have a working page, feed, and commercial measurement. Separate brand, product, comparison, replacement, and support intent when those searches have different economics. Review query quality alongside order value, margin, returns, stock, and product-level availability; click volume alone does not show whether discovery is useful.

Local and lead-generation accounts

Tie each service theme to the areas the client actually serves and to a page that explains availability, urgency, price context, and next steps. Separate calls, bookings, quote requests, and weak contact events in the outcome model. A broad keyword can be reasonable when service-area and response signals are strong; a more controlled type may be safer when one missed call is expensive or the client has limited capacity.

SaaS, B2B, and enterprise accounts

Map the search to the intended company, use case, product tier, and buying stage. Keep jobs, documentation, support, free tools, and competitor research separate from acquisition unless the client has explicitly approved them. Use longer qualification and sales-lag windows, and connect query decisions to accepted opportunities, pipeline, activation, or revenue rather than early forms alone.

For an enterprise client, document approval and routing more formally. One shared negative list or broad discovery policy can affect regions, product teams, resellers, and named accounts differently. Keep the owner and rollback visible to everyone who can change the account.

Avoid common agency setup mistakes

The most expensive errors are usually process errors around a reasonable-looking keyword list. Watch for these patterns:

Copying one client’s match-type mix. Similar words can represent very different economics, locations, pages, and qualification standards.
Choosing broad before the account can evaluate it. Without a specific page, reliable outcomes, and a staffed Search terms process, discovery can create work that nobody can classify or improve.
Treating phrase or exact as literal-only. Current matching can include related or same-intent searches, so query review remains necessary.
Putting every exclusion in a shared list. Scope and negative behavior matter; a list built for a recruitment exclusion can block a legitimate client product or service elsewhere.
Mixing implementation with measurement changes. If the agency changes the keyword type, page, conversion event, budget, and bidding at once, the client cannot tell which decision produced the result.
Reporting clicks instead of client value. CTR, CPC, and raw leads are useful diagnostics, but they do not replace orders, booked jobs, activation, accepted opportunities, revenue, or margin.
Leaving ownership implicit. A negative policy without a reviewer, due date, and approval rule will decay as the account and client offer change.

When one of these mistakes appears, do not hide it by adding more keywords or more negatives. Preserve the current state, describe the risk, agree on the smallest corrective change, and set a follow-up review.

Final agency match-type setup checklist

Before handing the build to the client or launch owner, confirm:

the client objective and mature outcome are written down;
each campaign, ad group, keyword, page, and match type has a clear relationship;
broad, phrase, or exact was selected for a stated reach-versus-control reason;
the account’s product, service area, audience, and buying-stage boundaries are represented in the structure;
negative categories have evidence, scope, and an owner;
conversion actions, values, lag, qualification, and response routing are understood;
Search terms review has a cadence, decision log, and escalation path;
client approval, implementation date, guardrails, and rollback are recorded; and
the first report will compare qualified outcomes with the prior state and any other changes made at the same time.

For a broader conceptual comparison, see the best Google Ads match types for agencies guide and the Google Ads match types best-practices hub. Start with one client theme, complete the QA and review loop, and only then expand the pattern to other campaigns or accounts.