Measured Marketing Performance reports confirmed provider observations and tenant POS evidence. It does not display predicted reach, predicted revenue, or fabricated ROAS. Missing evidence remains Unavailable instead of being converted into a measured zero.
Read the headline metrics
Open Admin → Marketing → Measured Marketing Performance. Measured reach and Measured clicks use confirmed observation snapshots from Instagram or signed tracking events. Attributed orders and Net attributed revenue use the available POS attribution evidence and recorded refunds.
The snapshot count matters. A low or empty sample does not support the same conclusion as a mature observation history. If attributed orders or revenue shows Unavailable, no adequate attribution evidence exists yet; it does not mean that the campaign produced zero orders.
Distinguish observed and inferred attribution
Under POS attribution evidence, UTM and coupon matches are Observed. Customer, product, and time-window matches are Inferred and remain labeled as such. Review gross, refunded, and net values together so a refund is not mistaken for retained revenue.
Observed and inferred records answer different questions. Do not present an inferred match as proof that a campaign caused the sale. Use the underlying record, timing, product, and customer evidence when investigating a disputed attribution.
Interpret Instagram Insights ingestion
Pending means the workflow is awaiting its first measured snapshot. Synced means a measured snapshot was stored. Unavailable means Meta omitted required metrics. Retrying represents a transient provider failure. Action required means the account must be reconnected or granted Insights access. Closed means the polling window completed.
If reach looks like zero, check the ingestion state before reporting zero performance. A missing provider metric is stored as unavailable, not as a measured zero.
Use recommendations and the audit trail
Evidence-backed recommendations expose observation IDs, sample size, and measured period. Insufficient evidence produces Continue learning, not a prediction. Under Observation audit, review the publication, creative, recipe version, hook, format, product, channel, measured period, and stored engagement fields.
Before making a budget or content decision, confirm that the publication has a real remote media ID, observation ingestion is synced, the sample size is adequate, and attribution type is stated. Ellich's measured report supports operational review; it does not replace channel billing records or establish causal lift by itself.