Why don’t Apple Ads keyword totals match campaign totals?

Keywords, search terms and campaign totals answer different questions

• Keyword report: performance attributed to the keywords you bid on. A broad-match keyword can match multiple related searches; it is not the exact query a person typed.
• Search-term report: actual queries Apple discloses, their matching source and associated metrics. Some query text may be withheld.
• Campaign and ad-group reports: independently reported totals for the selected scope. Use these to assess overall delivery and spend.

Broad match does not by itself prove that delivery is missing from a keyword report. Check Search Match, filters and completeness before explaining a gap. Never add campaign totals to ad-group or keyword totals: these are overlapping views of activity.

Before comparing numbers, align the report settings

1. Select the same app, campaign, ad group and country.
2. Match the exact start and end dates and account reporting timezone. Today’s results may still change.
3. Compare the same metric: total installs and tap-through installs are different measures. Keep currencies separate.
4. Check filters, pagination and provider errors. An unavailable report is not zero activity.
5. Read the search-term report, including rows with undisclosed query text. Complete pagination does not mean Apple disclosed every query.

If a difference remains, record it as unexplained coverage until the evidence identifies its source. A reporting gap alone does not tell you which searches wasted money.

What we observed in Respira’s US Discovery campaign

On September 15, 2026, we queried the founder’s app, Respira, using Storewake’s Apple Ads integration. The report covered September 9–15 in the account reporting timezone (ORTZ), for the US market.

The search-term report returned 11 rows, including 5 with undisclosed query text. Those rows retained their metrics. We did not invent search terms for them or treat them as zero activity.

Spend is rounded from EUR 7.0478. This is one recorded observation, not a live dashboard or a profitability result. The window included the current day, so a later query can return different numbers. The equal campaign and ad-group values describe this example, not a rule for every campaign.

The lesson is practical: an agent needs official totals and disclosure limits alongside the visible searches. Read the Respira ranking case study for a separate example of investigating organic movement.

Ask your AI agent to reconcile the reports

Connect Apple Ads in Storewake Settings, then connect a compatible AI agent through Storewake’s remote MCP. Live integrations and MCP access require Grow. Use list_apps to find your own app ID.

For my app in [country], investigate why Apple Ads keyword
metrics differ from campaign totals over the last 7 days.
Use campaign_performance and search_terms.
Keep the returned dates, timezone and filters identical.
Show official campaign and ad-group totals separately.
Then show disclosed queries, matching source, spend, taps
and installs. Preserve undisclosed queries and missing values.
Check completeness and explain any remaining uncertainty.
Suggest one investigation before increasing the budget.
Do not change bids, budgets or targeting.

For a focused search-term query, supply your app ID, country and optional campaign ID:

search_terms({
  appId: YOUR_APP_ID,
  country: "us",
  days: 7,
  campaignId: YOUR_CAMPAIGN_ID
})

These Apple Ads MCP tools read reports; they do not execute targeting changes. Provider requests use your workspace allowances.

Turn a disclosed query into a decision to review

Start with relevance and evidence. A relevant query with conversions may be worth evaluating as an exact-match keyword. Check whether it already exists and choose an explicit bid before creating it.

A clearly irrelevant query may be worth excluding. Spend without installs warrants investigation, but a small sample or attribution delay is not enough to label it wasted spend. Compare against your own acquisition target and allow time for results to mature.

Storewake’s search-term screen provides reviewed actions to add an exact keyword or a negative in the source ad group. An owner or admin must preview and confirm the change. Undisclosed queries cannot be targeted by guessing their text.

Sources and measurement limits

Example: a recorded Storewake integration validation on September 15, 2026, for Respira’s US Discovery campaign. We verified the Apple API response; this example does not claim a separate manual dashboard reconciliation, incremental installs or revenue lift.

• Apple: understand keyword match types
• Apple: reporting options and definitions
• Apple: search-term reports and disclosure limits

For changes in organic visibility, use the ranking-drop investigation guide. Paid and organic observations provide context; neither automatically establishes the cause of the other.