Identity is the layer that makes POAS honest. It stitches ad exposure to downstream orders without relying on a single third-party pixel as the source of truth, so the dashboard does not shrink every time Apple or Google ships an update.
How resolution works
Identity resolution combines three signal sources in priority order:
- Server-side conversion events from your storefront — deterministic, durable, owned by you.
- First-party identifiers (email, order ID, customer ID) joined to ad-click events where the click context is available.
- Probabilistic matches (IP + user-agent + timing) used only as a tiebreaker when deterministic signal is missing, with a confidence score attached.
What it produces
Every order has a most-likely attribution chain with a confidence score. POAS is computed against the deterministic and high-confidence chains by default; low-confidence chains are surfaced but tagged so the operator can choose whether to include them.
What you control
You choose the confidence floor for what counts toward POAS, the retention window for raw signals, and the export format if you want to ship the resolved identity graph into your own warehouse. Identity data is yours; we do not resell it or model it across customers.
Takeaways
- Server-side conversion events are the durable foundation; pixels become a redundancy, not the source of truth.
- Every order has a confidence-scored attribution chain.
- You own the confidence floor, the retention window, and the export.