Keep delivery, attention and business value separate
Delivery asks whether the approved asset appeared on the intended account. Attention asks what people watched or did. Business value asks whether a defined customer outcome followed. A dashboard should preserve these layers rather than collapse them into one fleet score.
Use consistent time windows and platform definitions. Do not sum unlike view metrics and call the total unique people. Report deduplicated reach only when the data actually supports deduplication.
Compare complete costs
Organic distribution includes production, review, tools, people and exception handling. Paid media includes creative and operating costs as well as spend. Compare equivalent outcomes and cohorts. A low CPM alone does not establish that either channel scales profitably.
The calculator reports cost per thousand entered impressions and cost per entered qualified action. Revenue contribution uses your entered contribution per action. It is an accounting scenario, not proof that the program caused the action.
Investigate anomalies without inventing a bot score
Unexpected geography, repeated patterns, unusual engagement or a mismatch between views and downstream activity can justify investigation. None alone proves purchased traffic. Ask for platform-native reports, source breakdowns, post URLs and a clear measurement definition.
YouTube prohibits fake engagement. A supplier’s promise of guaranteed views should therefore prompt careful scrutiny of how the result is produced. Preserve evidence and separate confirmed facts from suspicions; do not accuse a provider based only on a cheap quote.
Treat attribution as a model
First-touch, last-touch and multi-touch systems assign credit under different rules. They do not automatically measure incrementality. Check identity matching, consent, missing events, attribution windows and CRM reconciliation when evaluating attribution tools.
Investors and operators need the underlying cohort, retention and contribution picture. A blended headline can conceal a small number of strong assets or a changing acquisition mix. Show distributions and denominators where available.
Use the result to choose a next test
A program with strong reach and weak qualified response may need a clearer audience or offer. A program with useful outcomes but high labor may need a better workflow. Inspect the actual bottleneck before adding accounts or budget. Keep the next experiment specific enough to learn from.