50 Google Ads Prompts for Claude (2026)

Use Google Ads account [CUSTOMER_ID] for this chat.

Run campaign_performance_report for the last 7 complete
days, then for the 7 days before. Show each campaign
side by side: cost, clicks, conversions and cost per
conversion, with the % change. Flag any cost per
conversion that rose over [20]%.

Use gaql on campaign for the last 30 days with
metrics.cost_micros and metrics.conversions_value.
Rank campaigns by ROAS and mark any below [3.0].

Use gaql on campaign with segments.device for the last
30 days. Show cost, conversions and cost per conversion
by device. Which device converts worst?

Use gaql on campaign with segments.day_of_week and
segments.hour for the last 30 days. Show cost per
conversion by day and hour, and the three worst slots.

Use gaql on campaign with
campaign.advertising_channel_type for the last 30 days.
Total cost, conversions and conversion value by type,
with each type's share of spend.

Use gaql on customer with segments.date for the last 30
days. Flag days whose cost or conversions sit over two
standard deviations from the mean, and name the campaign
that moved most.

Run account_performance_report and
campaign_performance_report for last month and the month
before. Write a client summary under [150] words: spend,
conversions, top 3 campaigns and one next step.

Run search_terms_report. List every term that cost more
than [$50] with zero conversions, sorted by cost, with
its campaign and ad group. Group them by theme.

Run search_terms_report for the last 60 days. List terms
with [2]+ conversions whose status is NONE, meaning not
yet a keyword. Suggest a match type and ad group for each.

Run keyword_performance_report. List keywords whose cost
per conversion is above [$TARGET_CPA], and keywords that
spent more than that with no conversion.

Use gaql on keyword_view for the last 30 days with the
ad_group_criterion.quality_info fields. List keywords
scoring below 5 by cost, and the component pulling each
one down.

Run keyword_performance_report and total cost, clicks and
conversions by match type. Which converts cheapest, and
what share of spend sits in broad?

Run keyword_performance_report. For my [20] highest-spend
keywords, show the first-page, top-of-page and
first-position CPC estimates beside the average CPC paid.

Run generate_keyword_ideas with page_url [URL], seeds
[SEED 1] and [SEED 2], locations [US], limit 50. Group
the ideas by intent and drop any I already bid on.

Run search_terms_report for the last 90 days. Find terms
containing [COMPETITOR 1] or [COMPETITOR 2]. Show cost,
clicks and conversions per competitor, and the campaigns.

Use gaql on campaign for THIS_MONTH with
campaign_budget.amount_micros and metrics.cost_micros.
Compare each enabled campaign's cost with daily budget
times days elapsed, and project month-end spend. Flag
over [110]% or under [90]%.

Use gaql on campaign for the last 30 days with
metrics.search_impression_share and
metrics.search_budget_lost_impression_share. List Search
campaigns losing over [10]% to budget, with their budget.

Use gaql on campaign for the last 30 days with
metrics.search_budget_lost_impression_share and
metrics.search_rank_lost_impression_share. Say whether
budget or rank costs each Search campaign more.

Use gaql on campaign with
campaign_budget.explicitly_shared and
campaign_budget.reference_count. List each shared budget,
its daily amount and every campaign using it.

Run campaign_performance_report and read each daily
budget with gaql. Plan moving [$100] a day from campaigns
above [$TARGET_CPA] to campaigns losing impression share
to budget. Plan only.

Run list_bidding_strategies, then bidding_strategy_report.
For each strategy with a target CPA, compare the target
with its actual cost per conversion. Flag gaps over [20]%.

Run ad_performance_report and pick my [3] ad groups with
the most conversions. Read their headlines with gaql on
ad_group_ad. Draft 15 headlines (30 characters max) and
4 descriptions (90 max) for each. Wait for my approval,
then add each with create_responsive_search_ad, using
the current ad's final URL.

Use gaql on ad_group_ad with ad_group_ad.ad_strength for
responsive search ads. List ads rated POOR or AVERAGE,
with impressions and conversions for the last 30 days.

Run ad_performance_report. In each ad group with [2]+
ads and [500]+ impressions, name the ads with the best
and worst conversion rate.

Use gaql on ad_group_ad with
ad_group_ad.policy_summary.approval_status and
policy_topic_entries. List every DISAPPROVED or
APPROVED_LIMITED ad with its policy topic.

Use gaql on ad_group_ad_asset_view for the last 30 days
with asset.text_asset.text and field_type. Rank the
headlines in my top [5] ad groups by impressions and
conversions, and name the weakest.

Use gaql on campaign_asset where field_type is SITELINK
to find Search campaigns with none. Draft [4] sitelinks
each (25 characters max) from [PAGE URLS]. Wait for my
approval, then add them with add_sitelinks.

Use gaql on age_range_view and gender_view for the last
30 days. Show cost per conversion for each age range and
gender, and flag any above [1.5] times the average.

Use gaql on geographic_view with segments.geo_target_city
for the last 30 days. List the [10] cities with the most
cost and no conversions, named via geo_target_constant.

Use gaql on ad_group_audience_view for the last 30 days
with ad_group_criterion.display_name. Rank segments by
cost per conversion, and flag spend with no conversions.

From prompt 29, take cities with over [$100] in cost and
no conversions, with the campaign for each. Wait for my
approval, then exclude them with add_location_targeting
and negative set to true.

Use gaql on ad_group with segments.device for the last
30 days, in MANUAL_CPC campaigns. Where [MOBILE] costs
[1.5]x the average per conversion, propose a modifier
such as 0.8. Wait for my approval, then apply it with
set_device_bid_modifier.

Run suggest_geo_targets for [CITY] in [COUNTRY CODE] and
show the matches with their IDs. Wait for my approval,
then add the one I pick to [CAMPAIGN] with
add_location_targeting.

Run list_conversion_actions, then gaql on
conversion_action with primary_for_goal, counting_type
and attribution_model_settings.attribution_model. Flag
duplicates and enabled actions that are not primary.

Use gaql on campaign with segments.conversion_action_name
for the last 30 days. Show conversions and all
conversions by action, and flag any action with none.

Use gaql on customer to check auto_tagging_enabled. Then
run list_conversion_actions with conversion_action_id
[ID] and show its tag snippets for my developer.

Draft a conversion action named [NAME]: type WEBPAGE,
category SUBMIT_LEAD_FORM, counting ONE_PER_CLICK. Wait
for my approval, then run create_conversion_action and
give me its tag snippets.

Here are [N] closed deals with gclid, time and value:
[ROWS]. Check that conversion action [ID] is type
UPLOAD_CLICKS. Wait for my approval, then run
upload_click_conversions.

Open prompts_menu for google-ads and run the
Comprehensive Account Audit Report for the last 30 days.
Give me the top [10] issues by money at stake.

Use gaql on change_event for the last 7 days with
change_date_time, user_email, change_resource_type and
changed_fields, limit 100. Summarize who changed what,
and match changes to any jump in cost.

Use gaql on campaign for campaign.optimization_score,
then on recommendation for open recommendation types.
List campaigns below [70]%. Do not apply anything.

Use gaql on landing_page_view for the last 30 days with
unexpanded_final_url, clicks, cost and conversions. Flag
pages with [100]+ clicks and no conversions.

List my connected Google Ads accounts with
query_contexts. Run account_performance_report on each.
One table: account, cost, conversions, cost per conversion.

Use gaql on campaign for the last 30 days with
campaign.bidding_strategy_type and metrics.conversions.
List MANUAL_CPC or TARGET_SPEND campaigns with [30]+
conversions as Target CPA candidates. Plan only.

Take the terms from prompt 8. Propose each as an EXACT
negative at the campaign level, grouped by campaign.
Wait for my approval, then add them with
add_negative_keywords.

Take the keywords from prompt 10, with ad group, cost
and conversions. Wait for my approval, then pause them
with update_keywords, status PAUSED.

Raise [CAMPAIGN A] to [$X] a day and lower [CAMPAIGN B]
to [$Y]. First tell me if either budget is shared. Wait
for my approval, then run update_campaign with
budget_amount.

Set the target CPA of [CAMPAIGN] to [$X]. First show the
current target and its cost per conversion. Wait for my
approval, then run update_campaign with target_cpa.

Plan a Search campaign for [PRODUCT] in [US] at [$50] a
day, sending clicks to [URL]: create_search_campaign,
2 ad groups with create_ad_group, 10 phrase keywords
each with add_keywords, one ad each with
create_responsive_search_ad. Wait for my approval.
Leave everything paused.

Show me [CAMPAIGN]: budget, bid strategy, ad groups, ads
and keywords. Wait for my approval, then set it live
with update_campaign status ENABLED, resume_ad_group and
resume_ad.

claude mcp add --transport http google-ads-mcp https://google-ads.insightfulmcp.com/