Google Ads guide

How to Set Up Google Ads Search Terms Report for Agencies

To set up a Google Ads search terms report for an agency, establish a separate source and decision log for each client, define that client’s acceptable demand, and give analysts a consistent sequence for review. Pull the report for the correct account and period, label query intent, compare platform activity with client outcomes, route changes for approval, and record the next check. Standardize the process; never assume that a keyword rule or negative list means the same thing for every advertiser.

Google’s search terms report shows searches that resulted in ads being shown. The agency setup must add the context that Google cannot know: the client’s offer, margin, geography, sales qualification, inventory, service capacity, and commercial priorities.

Create a client brief before opening the rows

Write a short account brief that the analyst can use while reviewing the report. Include the products or services being promoted, the audiences and markets the client wants, the customers it cannot serve, the key competitors or brands, price or margin constraints, and the business outcome that matters. Note the sales cycle and who can confirm whether a lead, order, or appointment was actually useful.

The brief should answer questions such as:

  • What does this client sell today, and what is only planned?
  • Which locations, industries, company types, or customer profiles are in scope?
  • Which queries are clearly irrelevant, risky, or outside capacity?
  • Is the goal a purchase, qualified lead, sales opportunity, booked appointment, store visit, or something else?
  • Which changes can the agency make directly, and which need approval?

Keep the brief account-specific. An education term, a competitor name, or a free-intent phrase can be a negative for one client and an important product theme for another. The brief prevents a shared agency rule from quietly overriding the client’s strategy.

Pull the correct report and protect account boundaries

Start each review from the intended customer account rather than from a combined agency view. Verify the account name or identifier, campaign scope, date range, time zone, currency, and conversion actions. If the agency uses manager-account exports, retain the customer identifier on every row and make it impossible to merge two clients into the same analytical key.

Use fields that connect the search with a decision: search term, campaign, ad group or asset group where available, keyword, match type, impressions, clicks, cost, conversions, and conversion value. Add landing page, device, location, product, call outcome, or lead status when the account has reliable data for them. Do not expose another client’s raw query or cost in a report, screenshot, dashboard, or working copy.

Keep a dated source snapshot or a clear report reference. Record any filters, exclusions, or late data. If the account has a long sales cycle, do not declare a query unsuccessful simply because the platform conversion has not yet matured. If the client has incomplete CRM data, state that limitation in the action log.

Define the review labels and owners

Use a small set of labels that describe both what the query means and what happens next. One workable setup is:

  1. relevant and retain;
  2. relevant opportunity;
  3. landing-page or offer mismatch;
  4. outcome needs client validation;
  5. negative candidate; and
  6. not enough evidence.

Add fields for proposed action, scope, owner, approval status, evidence link, and review date. “Relevant opportunity” might lead to a keyword or landing-page test. “Outcome needs client validation” might wait for sales disposition. “Negative candidate” should not become an applied negative until the agency has checked scope and known good queries.

Separate analyst classification from client decision. The analyst can describe a query as “employment intent,” but the client may still have a recruiting offer. The account strategist can propose a campaign-level negative, while the client owner decides whether that segment is strategically excluded. Clear ownership prevents a report from presenting an internal assumption as an approved business rule.

Review rows in a deliberate order

Do not start with the first row in an alphabetical export. Sort or filter to find expensive mismatches, high-value outcomes, repeated themes, and new language. Read the exact query and the matching context. Then ask whether the ad and destination made an accurate promise.

For each important row, note:

  • the customer problem or product named;
  • the intended market or audience;
  • the matching keyword or targeting context;
  • the cost and resulting platform action;
  • the client outcome known so far; and
  • the next action or missing evidence.

A semantically related query is not automatically a good query. A software company may not want implementation jobs; an ecommerce client may not stock the searched model; a local provider may not cover the city; and a lead form may attract students rather than buyers. Write the mismatch precisely so the next person can reproduce the decision.

For uncertain rows, preserve the uncertainty. Assign a client or sales-team question instead of forcing a negative. For repeated relevant terms, consider whether the account needs better coverage, a dedicated page, a clearer ad, or a feed correction. A search terms review can identify an account-structure or messaging problem, not only a keyword problem.

Route negatives through approval and scope

Use the Google Ads negative-keyword guidance as a technical reference, then apply the client’s campaign structure. For every negative proposal, record the observed phrase, why it is wrong for this client, where it should apply, which useful queries were checked, who approves it, and when the result will be reviewed.

Keep universal exclusions separate from client-specific exclusions. A shared list can be appropriate for a pattern that is genuinely irrelevant to every receiving account, but it should be treated as a governed asset. A word such as “course,” “repair,” “near me,” or a place name can be a legitimate theme elsewhere.

When scope is uncertain, start narrower. A campaign-level exclusion may protect one offer while leaving another campaign available. Review match behavior and check whether the negative could remove a valuable longer query. After applying it, compare irrelevant spend and qualified coverage. If coverage falls, inspect the negative first rather than stacking another rule.

Do not use a negative to hide a broken conversion event, slow client follow-up, weak qualification, or a page that does not match the ad. Those are useful findings that should go to the correct owner.

Reconcile platform data with client outcomes

The report should show what happened after the click whenever the client can provide that evidence. For lead generation, request lead disposition, accepted lead rate, opportunity stage, revenue, and lost reason where available. For ecommerce, consider product availability, margin, returns, new-customer status, and repeat purchase. For local accounts, connect calls with qualified conversations, bookings, shows, and completed jobs.

Google’s measurement documentation gives platform context, but an agency should state which numbers are platform events and which come from the client’s system. A form completion is not necessarily a qualified lead. A purchase is not automatically profitable. If client-side evidence is missing, show the gap and make the next request explicit.

When a relevant query underperforms, inspect the handoff rather than changing targeting immediately:

  1. Is the query aligned with the offer?
  2. Does the ad set a truthful expectation?
  3. Does the page or product route answer the query?
  4. Did the client respond and qualify the interaction?
  5. Is the conversion or revenue data trustworthy?

This sequence turns the search terms report into a cross-functional diagnosis. It can reveal that media targeting is acceptable while the page, sales response, inventory, or tracking needs attention.

Build an approval-ready agency output

Give the client a concise summary supported by an internal detail view. The summary should state the account and date range, major query themes, spend or outcome context, recommended changes, unresolved questions, and approvals needed. Use representative examples, not a long list of unexplained rows.

For each proposed change, state:

  • what was observed;
  • why it matters to this client;
  • what will change and where;
  • who owns the decision or implementation;
  • what could be lost or changed by the action; and
  • what signal will determine whether it worked.

For example, “exclude free template searches from the nonbrand lead campaign and review qualified-lead coverage in two weeks” is clearer than “add negatives.” “Create a product-category page for the recurring high-margin theme and watch product-level revenue” is clearer than “add keywords.” Keep client approval and implementation status distinct.

The agency Performance Max guide can provide broader account context when the report includes PMax. adz.so can help inspect supported account data and structure the workflow, but it does not replace client authorization, sales qualification, product availability checks, or revenue reconciliation.

Check the setup before making it routine

Run a QA pass on the first report and whenever the process changes. Confirm:

  1. account, customer, campaign, date range, and time zone are correct;
  2. the client brief reflects the current offer and market;
  3. platform conversions and client outcomes are labeled separately;
  4. every important row has a meaningful classification and reason;
  5. proposed negatives have scope, approval, and a review condition;
  6. shared lists and templates have been checked for client vocabulary;
  7. landing pages, forms, calls, feeds, and follow-up were considered;
  8. the client-facing output contains no other customer’s data; and
  9. the next owner and review date are recorded.

Make changes in controlled groups when possible. If the agency changes negatives, landing pages, conversion actions, and budgets at once, the next report will not explain which intervention affected coverage. Keep a small change log and annotate unusual launches, outages, promotions, feed changes, or sales-policy changes.

An agency search terms report is set up well when an analyst can classify a query, a strategist can explain the tradeoff, a client can approve the right action, and the next review can test the result. Protect account boundaries, retain the exact evidence, and make local assumptions or missing outcome data visible. That discipline is more valuable than a universal list of “good” and “bad” words.