Pipelines overview

A pipeline is the core runtime workflow in Detect that evaluates streaming feed data, applies business rules, and sends qualified events to configured destinations.

Use pipelines to convert real-time feed input into actionable events. A pipeline combines audience qualification, event conditions, destination mapping, schedule controls, and contact policy limits.

All Pipelines dashboard

The All Pipelines dashboard provides operational visibility and control for pipeline management. You can review overall status counts, narrow the list by filters, search for a specific pipeline, and open tile-level actions.

  • Summary cards show All Pipelines and status totals for Running, Drafts, Completed, and Stopped.
  • Filter controls are available for folders, feeds, and status.
  • The status filter supports Draft, PendingApproval, Approved, Declined, Ready, Running, and Stopped.
  • + New Pipeline to create a new pipeline.
  • Search is available to locate pipelines by name.
  • Grid and list display toggles are available for pipeline results.
  • Action menu is available to delete multiple pipelines simultaneously.
Each pipeline tile shows the pipeline name, current status, and key metadata such as updated date and user. A per-tile action menu is available. In the shown running pipeline state, the menu includes Edit, and Delete.
Note:
For Pipeline in the Declined status, the delete option is disabled.

Pipeline building blocks

A pipeline is configured from the following functional blocks:

  • Audience: Defines the eligible audience by using attribute-based matching conditions.
  • Trigger: Selects the source feed and transactional conditions that determine when an event is matched.
  • Destination: Specifies the target system and attributes to forward for each matched event.
  • Schedule: Controls active days and time windows for pipeline execution.
  • Contact Policy: Applies global and per-user limits to control event frequency.

Message-only and action-based configurations

By default, a data destination sends a message each time a customer meets the audience and trigger conditions. You can add an action to the message. An action tracks whether the customer completes a follow-up transaction goal within a time period. When the customer meets the goal, Detect sends a fulfillment message. For more information, see Action-based pipelines.

Multiple data destinations

A pipeline can have more than one data destination. Each destination has its own event ID and configuration, and it can have its own actions. The mindmap shows the total number of data destinations.

Pipeline lifecycle

Pipeline execution follows a governed lifecycle:

  1. Create and configure the pipeline on the visual canvas.
  2. Save as draft.
  3. Submit for approval or review based on your access model.
  4. Approve where administrator approval is required.
  5. Review and deploy.
  6. Start immediately or configure schedule start and end criteria.
  7. Monitor runtime status and throughput from reports.

Lifecycle states visible from the dashboard filter are Draft, PendingApproval, Approved, Declined, Ready, Running, and Stopped.

Pipeline reporting

After deployment, open the pipeline report to validate performance and delivery quality. Key metrics include:

  • Total events matched
  • Total events sent
  • Matched-versus-sent percentage