Run a weekly Google Ads audit with your AI agent

SKILL.md---
name: google-ads-weekly-audit
description: "Every Monday: last week's Google Ads spend and CPA, wasted spend on search terms, disapproved ads and campaigns held back by budget, in one card. Use when the user asks for this, or when it runs on a schedule."
---

# Google Ads weekly audit

Every Monday: last week's Google Ads spend and CPA, wasted spend on search terms, disapproved ads and campaigns held back by budget, in one card.

Most Google Ads accounts do not fail loudly. An ad is disapproved on Thursday and the ad group quietly serves its weaker variant, a campaign runs out of budget by mid-afternoon every day, a broad match keyword starts buying searches for jobs. Each problem surfaces in a different corner of the interface, if you go looking.

The agent goes looking every Monday. It runs the same checks through the Google Ads API, compares spend and CPA with the week before, and lists the problems in the order they cost money, so the account manager starts with the fix instead of the hunt.

## Two ways to get the data

Use the API or MCP route when you have an MCP server, a CLI or an API token for the app. Otherwise, if you can drive a browser the user is signed in to (Claude in Chrome, a Playwright or Chrome DevTools MCP server), use the browser route. Say which one you used.

## Route A: API or MCP server

### What you need

- Read access to the Google Ads account, or to the manager (MCC) account above it, and the customer ID.
- Google Ads API access: Google's own read-only Google Ads MCP server (`googleads/google-ads-mcp`) or a connector; or OAuth credentials for a user on the account, a developer token where your access level still needs one, and `login-customer-id` for a manager account.
- Your target CPA, so the agent knows what counts as waste.

If something is missing, say what and stop; do not invent numbers.

### Steps

1. **Pull last week against the week before** (Google Ads): Run GAQL on `customer` with `metrics.cost_micros`, `metrics.conversions`, `metrics.conversions_value` and `metrics.clicks` for `segments.date DURING LAST_7_DAYS`, then for the 7 days before with `segments.date BETWEEN`. Divide micros by 1,000,000; CPA is cost over conversions.
2. **Find budget-limited campaigns** (Google Ads): Query enabled `campaign` rows with `campaign.primary_status_reasons`, `campaign_budget.amount_micros` and `metrics.search_budget_lost_impression_share`. Flag those whose reasons include `BUDGET_CONSTRAINED` or that lost more than 10% of impression share to budget.
3. **List disapproved and limited ads** (Google Ads): Query `ad_group_ad` where `ad_group_ad.status = ENABLED` and `ad_group_ad.policy_summary.approval_status IN (DISAPPROVED, APPROVED_LIMITED)`, with `ad_group_ad.policy_summary.policy_topic_entries` for the reason and the campaign and ad group names.
4. **Measure wasted spend** (Google Ads): Query `search_term_view` for `LAST_7_DAYS` with `metrics.conversions = 0` and cost above your target CPA, ordered by `metrics.cost_micros`. Their total is the week's wasted spend; the top terms go in the card.
5. **Rank the fixes by money at stake**: Order the findings by what they cost: wasted spend in dollars, budget-limited campaigns by the impression share they lost, disapprovals by last week's spend in their ad group. Say what to do first in one line.

## Route B: the browser

### What you need

- Signed in to Google Ads with at least Read-only access to the account, or to the manager account above it with the client account selected.

If something is missing, say what and stop; do not invent numbers.

### Steps

Start at https://ads.google.com/aw/overview. The user is signed in already; never type passwords or codes, and ask if a sign-in page appears.

1. **Pick the account and the week**: If you sign in through a manager account, choose the client account in the account picker. In the date range at the top right choose `Last week` (Monday to Sunday), switch on Compare with `Previous period` and Apply.
2. **Read spend, conversions and CPA**: Open Campaigns in the left menu and read the Totals row of the table: Cost, Conversions and Cost / conv. with the comparison values. Use Columns > Modify columns if any of the three is hidden.
3. **Find campaigns held back by budget**: On the same Campaigns page, look for `Limited by budget` in the Status column, and add `Search lost IS (budget)` from Columns > Modify columns > Competitive metrics. Flag campaigns above 10%.
4. **List disapproved and limited ads**: Open Campaigns > Ads, add the `Policy details` column from Attributes, and filter the Status column for `Disapproved` and `Eligible (limited)`. Tools > Troubleshooting > Policy manager > Policy issues shows the same list with the policy and the appeal option.
5. **Measure wasted spend**: Open Campaigns > Insights and reports > Search terms with the same week, add a filter `Conversions = 0` and `Cost >` your target CPA, and sort by Cost. The total is the week's waste; note the top 5 terms with their campaign.
6. **Rank the fixes by money at stake**: Wasted spend in dollars, budget-limited campaigns by lost impression share, disapprovals by last week's spend in their ad group. Say what to do first in one line and change nothing.

Note: Conversions keep arriving for days after the click, so Monday's CPA for last week reads higher than it will by Friday in the UI as well. The Search terms table hides terms below a privacy threshold, so the UI total can be a little lower than the API's. Policy manager and column settings are desktop-only pages.

## The report

- **Spend, conversions and CPA** last week, with the change on the week before.
- **Wasted spend**: search terms that cost more than your target CPA with no conversions, the total and the top 5.
- **Disapproved and limited ads**, with the policy reason and the ad group.
- **Budget-limited campaigns** and the impression share they lost to budget.
- **Fix first**: the one change with the most money behind it.

## Post it to Briefkin

When `BRIEFKIN_TOKEN` is set, or `~/.config/briefkin/env` exists (the installer from https://briefkin.dev/docs writes it), also send the result to the user's Briefkin board as one card. Keep the `item_key` the same on every run so the card updates in place, use a new `event_id` per run, and put today's real numbers in place of the example ones. If the request fails, say so once and carry on.

```bash
[ -n "${BRIEFKIN_TOKEN:-}" ] || . ~/.config/briefkin/env
curl -sS -X POST "${BRIEFKIN_URL:-https://briefkin.dev}/v1/events" \
  -H "authorization: Bearer $BRIEFKIN_TOKEN" \
  -H "content-type: application/json" \
  -d @- <<'JSON'
{
  "v": 1,
  "event_id": "google-ads-weekly-audit/<date of this run, e.g. 2026-10-16>",
  "item_key": "google-ads-weekly-audit",
  "project": "Acme",
  "source": "Google Ads weekly audit skill",
  "kind": "update",
  "title": "CPA up 9%, $1,140 spent on dead search terms",
  "summary": "Last week: **$12,480** spent (+6%), 296 conversions (-3%), CPA **$42.16** (+9%). Four things to fix, starting with $1,140 on search terms that never converted.",
  "widget": [
    {
      "type": "Grid",
      "props": {
        "columns": 4
      },
      "children": [
        {
          "type": "Stat",
          "props": {
            "label": "Spend",
            "value": 12480,
            "prefix": "$",
            "delta": 6
          }
        },
        {
          "type": "Stat",
          "props": {
            "label": "Conversions",
            "value": 296,
            "delta": -3
          }
        },
        {
          "type": "Stat",
          "props": {
            "label": "CPA",
            "value": 42.16,
            "prefix": "$",
            "delta": 9,
            "better": "lower"
          }
        },
        {
          "type": "Stat",
          "props": {
            "label": "Wasted spend",
            "value": 1140,
            "prefix": "$",
            "tone": "bad"
          }
        }
      ]
    },
    {
      "type": "Table",
      "props": {
        "columns": [
          "Check",
          "Finding",
          "Where"
        ],
        "rows": [
          [
            "Wasted spend",
            "$1,140 on 18 search terms with 0 conversions",
            "Non-brand Search, Competitors"
          ],
          [
            "Budget",
            "Lost 31% of impression share to budget",
            "Non-brand Search"
          ],
          [
            "Disapproved ad",
            "Destination not working",
            "Non-brand Search > Invoice app"
          ],
          [
            "Limited ad",
            "Trademarks in ad text",
            "Competitors > Alternatives"
          ]
        ]
      }
    },
    {
      "type": "Callout",
      "props": {
        "text": "Fix first: add negatives for the 18 dead search terms, then move that money to Non-brand Search, which lost 31% of its impression share to budget.",
        "tone": "warn"
      }
    }
  ]
}
JSON
```

Widgets come from the Briefkin catalog (https://briefkin.dev/v1/catalog). Without a token, just give the user the report.

## Run it on a schedule

Suggested: Mondays at 8:30 (`30 8 * * 1`). From cron, with the agent signed in on that machine:

```bash
30 8 * * 1 claude -p "Run the google-ads-weekly-audit skill"
# Codex: codex exec, Gemini CLI: gemini -p, Copilot CLI: copilot -p, OpenCode: opencode run
```

Or use the agent's own scheduler (Claude Code routines, Codex automations, Claude or ChatGPT scheduled tasks): https://briefkin.dev/skills/google-ads-weekly-audit#schedule

Source: https://briefkin.dev/skills/google-ads-weekly-audit