Meta Ads dashboard: metrics, template and AI prompts

Meta Ads dashboard at a glance

5 panels, read top to bottom once a week.

The next section maps each metric to its Insights API field.

The 14 Meta Ads dashboard metrics and their API fields

14 metrics answer 3 weekly questions: what did we spend, what came back, and which ads are wearing out.

API field names let a query or an AI agent pull the exact column. Definitions follow the Ad Account Insights reference in Meta's Marketing API docs, which use v27.0 in their example requests.

3 things that trip up a first dashboard

• Attribution window. Per Meta's docs, action_attribution_windows defaults to 7d_click and 1d_view. A number in Ads Manager with a different window will not match an API pull with the default. Fix one window and use it everywhere.
• Action lists. actions, action_values and cost_per_action_type return lists keyed by action type, so a dashboard must choose which type counts as a result, for example purchase or lead.
• Breakdown limits. Meta states that only some permutations of breakdowns are available. Age with gender works; pa_creator_ig_handle cannot be combined with age, gender or publisher_platform. Metrics are also unavailable across several attribution settings at once unless you pass the windows you want.

Insights data is available at 4 levels: ad account, campaign, ad set and ad, each with its own /insights edge (Insights overview). Per Meta, a request with no date preset returns roughly the last 30 days.

The Meta Ads dashboard template: 5 panels

Put the 14 metrics on 5 panels, ordered from the money to the creative.

Read panel 1 first. If cost per action rose, panel 2 shows whether CPM or CTR moved, panel 3 names the campaign, and panels 4 and 5 show whether an audience or an ad is behind it.

These are request shapes built from field names in Meta's reference; run them in Meta's Graph API Explorer or through an agent and check the account currency and window before trusting the totals.

Option 1: Ads Reporting inside Meta Ads Manager

Meta Ads Manager has its own report builder, called Ads Reporting. It is free and covers one business at a time.

Meta's help page for Ads Reporting did not load on the build day, so this section follows third-party guides (Metricool, Madgicx); check the steps against your own Ads Manager.

1. Open Ads Manager and choose Create custom report from the Reports menu.
2. Start from a template or a blank report and check the ad account and date range.
3. Add the metrics from the table above and the breakdowns for panels 4 and 5.
4. Optionally add a custom metric, which applies a formula to existing metrics, for example purchases divided by link clicks.
5. Save the report, then export it as .xlsx or .csv, share it by link or schedule it by email.

Strength: native, free and uses the same numbers as Ads Manager. Limit: one fixed layout you maintain by hand, and it holds Meta data only.

Option 2: the prompts that pull it through Ryze AI

Ryze AI answers questions about Meta spend, ROAS, CTR, creatives and ad sets from the Meta Ads accounts you connect, so each panel is one prompt.

Connect Meta Ads under Integrations with your Meta sign-in. Per the Ryze AI Meta Ads guide, Ryze AI then sees every ad account that login can reach, and you need at least Advertiser access on the account. Then open a new chat and paste a prompt.

Behind these prompts sit two read tools: meta_ads__getAccountSummary returns aggregate metrics on Meta's default attribution window, and meta_ads__runRawInsights runs an Insights query at account, campaign, ad set or ad level with breakdowns, filters and custom attribution windows. Ad creatives come from meta_ads__listCreatives.

For a quick look with no prompt, the Analytics page, then Ads Performance, shows the last 30 days of Meta results, including creatives.

The same pulls in Claude or ChatGPT

The same tools load in Claude and ChatGPT through the Ryze AI connector (Claude setup, ChatGPT setup); the full tool list is on the Meta Ads MCP page. For one Claude view across Google and Meta ads, see the live ads dashboard guide.

A weekly review routine for the dashboard

Spend 20 minutes on the same day each week and read the panels in order.

Re-read last week too: Meta can update attributed results after the fact, so recent days may change.

Scheduled version

The Ryze AI help center describes scheduled tasks; they run without an approval step, so keep the prompt read only.

Ads Reporting vs Ryze AI for a Meta Ads dashboard

Ads Reporting gives a saved layout in Meta. Ryze AI gives answers to new questions across accounts.

Meta also launched its own Meta Ads MCP server and CLI in open beta on April 29, 2026 (Meta announcement). Third-party guides give its URL as https://mcp.facebook.com/ads; Meta's page did not show it on the build day, so confirm it in Meta's help center. The Ryze AI connector at https://connector.get-ryze.ai/mcp is a separate server that also covers the other platforms Ryze AI connects.

Which Meta Ads dashboard to use

Pick by how many accounts you read and how often the questions change.

You can use both: keep an Ads Reporting report for the fixed numbers and ask Ryze AI for the why.

Sources

The verdict

A Meta Ads dashboard is 14 metrics on 5 panels read in the same order each week. Build the fixed version in Ads Reporting if one account is enough.

Use Ryze AI when you read several accounts or want to ask new questions of the same data. Connect Meta Ads, paste the prompts above and keep them read only.