Skip to main content
Use Workspace settings → Conversion events to choose what the Conversions metric counts. The default is Purchase. You can select purchases, subscriptions, trial starts and leads; the count is not a count of unique customers or proof of advertising delivery.

​
Before you begin

You need access to the workspace and permission to update its settings. Events must already reach Signals through the browser, Conversion API or supported Stripe webhook flow. Selecting an event does not create it or make an otherwise ineligible event count.

​
How do I choose conversion events?

  1. Select the intended workspace and open Workspace settings.
  2. Find Conversion events and select at least one supported event.
  3. Choose Save conversion events.
  4. Open Dashboard and check Conversions, its reporting period and the effective choices. The selection applies to retained history; a new selection can change historical totals.
Use the choices that match your business outcome. Selecting both a trial and a subscription can count two different stages for the same customer. Subscription payments can add revenue without adding another Purchase conversion when Subscribed is selected.

​
Why does Conversions differ from Purchases or Meta?

The workspace total uses the selected event definition and applicable provenance, consent, market, traffic and retention rules. Synthetic diagnostics do not establish real conversion evidence. Revenue verification, counted conversions and destination acceptance are separate observations. The arrival comparison can return a conversions block with count, effective_choices, selection_version, period and availability. An unavailable or null count is unknown, not zero. Paid performance can show Conversions beside Purchases; ad-level values require eligible attribution context and can differ from the workspace total. Meta reports conversions under its own attribution window and rules. Keep returned basis_note explanations with comparisons. Do not reconcile a Meta claim, a counted conversion and an earned journey credit as though they were the same metric.

​
Do changes rewrite existing Tests or proposals?

No. A new judging Test can use conversion_mode: "workspace_default" to freeze the current selection. With conversion_mode: "explicit", supply the supported event_name instead. Later workspace edits do not change that Test’s captured definition. A judging Test records a verdict; it does not itself create an advertising proposal or write to Meta. Supported proposal evidence also freezes its conversion_definition, including the version and event choices, beside the counted metrics. Inspect that captured definition when reviewing older proposals rather than applying today’s settings to their evidence. See frozen counted evidence.

​
Troubleshoot