# Test, launch, and manage a Campaign


Once the draft is saved, open **2 Test & Launch** at the top of Builder. Before launching, give it a name, brief, and cover image, and save the human-work form and accepted path. Work through the checks in order:

| Step | Enter or inspect | When it fails |
| --- | --- | --- |
| **Pricing & Budget** | Human-step rewards, expected spend, total budget, and available funds | Return to the relevant Supply reward or budget amount |
| **Eligibility** | Agreement, course/version, profile, and invitation conditions for the intended audience | Return to Entry Gate or Qualification requirements |
| **Dry-run** | Simulated test items, each node's result, and the run record | Open the failing node in Builder, fix it, save node and canvas, then run the test again |
| **Settings** | Public identity and remaining launch settings | Return to the named setting and save it |
| **Launch** | Current organization, Campaign, budget commitment, and final confirmation | Resolve the named blocker before attempting launch |

<a id="launch-blockers"></a>
When launch is refused, the message names the blocker:

| Message | What to do |
| --- | --- |
| This campaign has no task or data source for work to enter through. | Add a task or data source on the canvas, save, and launch again |
| This campaign has no saved Campaign File version. | Save the canvas, then launch |
| The saved campaign graph is invalid. | Return to Builder and repair the draft |
| The saved campaign graph could not be read. | Try again; nothing was changed |
| The campaign request failed. | Try again |
| Counsel hasn't approved a template version yet — you can save the campaign, but launch will be blocked until counsel review completes. | Save your work; launch opens once the agreement template is approved. Ask support for its status |

Use **Save & continue** after each part. A dry run is a simulation: it moves a set of synthetic items through the workflow's structure. In the illustrative labeling workflow, for example, it checks that items reach Supply and follow the connected branches to Sink, and a human step reports that it *would collect* a number of submissions. It does not generate labels or run the configured acceptance decision, so a passing dry run is not evidence that the acceptance logic works as intended; review that logic on its own. Inspect each node's actual result rather than assuming the path passed because it is connected.

Changing workflow settings that affect execution can invalidate an earlier check. Repeat the affected checks; merely moving a node's position on the canvas does not invalidate them.

The **Dry-run** tests the path without creating real contributions, payments, or production deliveries. A successful dry run is not a published Campaign. Select **Launch**, confirm the action, then return to the Campaign and check that its state is **Live**. Launch locks the applicable budget; [billing and budget](/collaboration/billing) are tracked separately from results.

## Understand running states

In an open Campaign, **Pause**, **Resume**, and **End** sit beside the Campaign name. Pause is offered for Live work and opens a confirmation showing work still in flight; Resume is offered when Paused. **End** opens a separate confirmation for a running Campaign, which you complete with **End campaign**. Read that confirmation and the pending-work counts before changing the state. After an action, check the new state and any **Closing** banner; a button click alone is not the result.

| State or action | Effect on work |
| --- | --- |
| Paused | No new work; existing work cannot be submitted |
| Resume | Restores running work; for a closing Campaign, cancels closing |
| Closing | Stops new claims, while still-valid existing work can finish |
| Finished | Submissions can no longer continue |

A closing Campaign can also show an underlying Paused state. Read the closing information together with that state. Unused locked budget is released after the system finishes closing, not as soon as someone selects the close action.

Before ending work, check open claims, submissions awaiting review, unsettled rewards, and data still to be delivered. Closing keeps those obligations in progress; **Finished** confirms that collection has ended, not that every result and payment has been reconciled. Use [results](/tasks/results) and [billing](/collaboration/billing) for the separate records.

After launch, the total budget can only be increased, through **Overview → Increase budget** or **Setup → Budget & costs**.

**Setup → Workflow & rewards** is a read-only comparison of the current version, and the running Campaign's **Test & Launch** is not a second launch button; to revise the workflow, use **Open builder**. Saving the canvas of a live or paused Campaign publishes the change at once: after you select **Save and apply**, queued work moves to the new version, and work a Contributor has already started finishes on the version it began with. There is no test step between that save and Contributors receiving the new version, so read [what the confirmation changes](/tasks/define#save-and-revise) before you save.