Skip to Content
ActivationCanvasBuilding Your First Canvas

Building Your First Canvas

This page walks you through creating an Audience Canvas from scratch — from clicking Create Canvas to publishing your first set of audiences.

Prerequisites

  • At least one entity type defined in your data model (for example, a users or customers entity).
  • A connected warehouse with populated tables.
  • Optionally, computed attributes — they unlock the most powerful filtering and ranking options.

Step 1: Create the canvas

  1. Navigate to Activation > Audience Canvases in the left sidebar.
  2. Click Create Canvas.
  3. Enter a Name that describes the family of audiences this canvas will produce (for example, "Customer Value Tiers" or "Winback Campaign Audiences").
  4. Select the Entity model the canvas operates on. All audiences produced by this canvas will be membership lists of entities of this type.
  5. Optionally, attach an Identity graph to enable cross-device resolution for every audience this canvas produces. If the graph is probabilistic, set the Minimum confidence threshold (0–100%) you want applied at query time.
  6. Click Create.

The canvas editor opens with an empty surface.

Step 2: Add your first entry node

A canvas needs at least one entry node before you can add any processing or output nodes.

  1. Open the Node palette on the left side of the editor.
  2. Choose an entry node type:
    • Criteria — if you want to define the root conditions directly on the canvas using the filter builder.
    • Audience Reference — if you want to start from an existing standalone audience.
  3. Drag the node onto the canvas surface, or click it to place it at the default position.
  4. Select the node to open its Inspector panel on the right.
    • For a Criteria node: build your filter tree — add conditions on attributes, events, time windows, and computed attributes, then group them with AND/OR logic. This works exactly like the standard audience builder.
    • For an Audience Reference node: select the standalone audience you want to use as the starting set from the dropdown. Only standalone (non-canvas) audiences on the same entity model appear in this list.

Each node displays a live estimated count below its title showing how many entities flow through it. Counts update as you configure the node.

Step 3: Add processing nodes

Add nodes to narrow, combine, partition, rank, or cap the set flowing through the canvas.

  1. In the Node palette, select a processing node type (Criteria, And, Or, Not, Split, Prioritize, or Limit).
  2. Place it on the canvas.
  3. Connect the upstream node’s output handle to the new node’s input handle by dragging between handles. A blue line confirms the connection.
  4. Select the node and configure it in the Inspector:
NodeWhat to configure
Criteria (mid-pipeline)Add filter conditions using the rule editor — same operators as the standard audience builder. When wired downstream, the criteria is AND-applied to the incoming set.
AndNo configuration needed — connects multiple upstream nodes; emits their intersection.
OrNo configuration needed — connects multiple upstream nodes; emits their union.
NotNo configuration needed — emits the complement of its single upstream input.
SplitChoose a Split mode (Predicate, Percentage, or Random), define branches, and optionally add a Remainder branch. See Node reference for details.
PrioritizeChoose the attribute to Sort by, the Direction (ascending or descending), and the number of rows to Keep (Top N).
LimitChoose a Limit mode (Random sample or Percentage) and the Value. For top-N by a specific column, place a Prioritize node upstream.

Repeat for each processing node you need.

Step 4: Add Output nodes

Every branch you want to activate must terminate with an Output node.

  1. From the Node palette, drag an Output node onto the canvas.
  2. Connect an upstream node’s output handle to the Output node’s input handle.
  3. In the Inspector, enter the Audience name — this becomes the name of the audience row on the Audiences list.
  4. Repeat for each output branch.

You can have as many Output nodes as you need. A canvas must have at least one Output before you can publish.

Step 5: Validate and review counts

Before publishing, review the live count badges on each node:

  • Counts update automatically as you edit nodes (with a short debounce delay).
  • A validation banner at the top of the editor lists any issues — for example, disconnected nodes, missing Output names, or unsupported cycles.
  • Resolve all validation errors before proceeding.

Step 6: Publish

  1. Click Publish in the top-right corner of the editor.
  2. A Publish diff panel appears, summarising what will change:
    • Added — new Output nodes that will create new audiences.
    • Renamed — Output nodes whose audience name changed.
    • Archived — Output nodes removed since the last publish (audiences are archived, not deleted).
    • Conflicts — removed outputs whose audiences are still referenced by an active sync or journey. Publishing is blocked until you resolve these conflicts.
  3. Review the diff and click Confirm & Publish.

Your audiences are now live. Each Output node’s audience appears on Activation > Audiences with a Canvas type badge and a link back to this canvas.

Next Steps

Last updated on
OverviewNode Reference