Skip to content
DocsAnalytics & reports
ANALYTICS & REPORTS

Read performance with context

Understand core metrics, compare like-for-like periods, and turn account observations into a considered next step.

3 min read5 sectionsUpdated September 14, 2026
Open AI Media Buyer
01

Start with the account and period

Confirm the selected ad account before interpreting a report. A different account, date range, or currency can change the meaning of every number.

Compare complete periods of the same length when possible. If a campaign only started recently, distinguish limited data from poor performance.

02

Know what the metrics measure

Use the metric definition shown in the report. Clicks and conversions can have different definitions depending on the view.

Know what the metrics measure
MetricMeaningCheck before acting
SpendAdvertising cost in the selected periodAccount currency and date range
CPMCost per 1,000 impressionsAudience, placements, and delivery context
CTRClicks as a share of impressionsWhether it uses all clicks or link clicks
CPCSpend divided by reported clicksThe click definition and destination quality
CPASpend divided by reported actionsWhich action or conversion is counted
ROASReported purchase value divided by spendTracking, attribution window, and purchase value
03

Read combinations, not isolated numbers

A higher CTR does not by itself prove that an ad is producing better customers. Read it alongside conversion results and spend.

A low reported CPA based on one conversion is different from the same CPA across many conversions. Check the amount of evidence before deciding to scale or stop.

Missing conversions may reflect tracking or attribution as well as actual performance. Check the underlying measurement before treating a zero as proof of no sales.

04

Use AI recommendations as a starting point

Ask a focused question, such as ‘Which creatives should I investigate in this date range, and what evidence supports that?’

AdRiseLab analyzes and recommends, and can apply supported budget or status changes after explicit approval. The user reviews campaign settings and confirms publishing; budget, bidding and campaign-structure decisions remain under user control.

Review the evidence behind a proposed change, the affected account and campaign, and the scope of the action before approving it.

05

Keep revenue sources distinct

Revenue connects website visits to sales and the ads that produced them. A tracker script measures visits on the sites you register, and orders are reported by your store's order-confirmation code, by a signed server event from your payment backend, or on Shopify by an Order payment webhook. Shopify and server-signed orders are marked verified; browser-reported orders are marked unverified. Meta's own reported conversions are shown beside the website ledger as a separate measurement rather than merged into it, and an order that never passed the storefront counts as revenue credited to no campaign.

When two views disagree, first check the date range, source, and attribution method. Do not add Meta-reported purchase value to the website ledger as though they represent different orders.