Skip to main content
A reasonable first question is whether you can try Concord on a real account without risking a live campaign. You can, and this page is the honest version of how.

​
There is no disconnected sandbox

Concord does not provide a separate test environment disconnected from the advertising platforms. You connect the platforms and accounts you are individually authorized to access, and you work on your real structure. What protects you is not a sandbox. It is the execution model.

​
The three guarantees

1. Nothing is written before you approve. Reporting, analysis, and campaign building can all be exercised end to end without submitting anything. A prepared platform action shows exactly what would be written, and stops there. 2. Everything created lands in a non-active state. Newly created campaign resources are created paused, draft, or disabled depending on the platform. This is enforced below the agent — a create action requesting an active status is overridden, on every platform and every tier. 3. Activation is a separate step. Turning something live is its own operation, with its own approval. A campaign Concord builds does not begin delivering because it was built.

​
What that makes possible

  • Build a full campaign tree from a real brief, review the diff, and never submit it
  • Submit it, inspect what landed in the platform interface, and delete it — it never spent
  • Run reporting and pacing analysis on live campaigns with no write path involved at all
  • Let a new team member work on a real account before you trust them with activation

​
What still needs care

  • Updates to existing live objects are real changes on delivering campaigns. The non-active-on-create rule protects creations, not edits — review budget, status, and date changes on live campaigns as you would any manual change.
  • Deletes are not always reversible. On some platforms a delete cannot be undone through the API. Concord defaults to pausing when you ask to stop or archive something, and asks for explicit confirmation before a hard delete.
The safest way to evaluate Concord is the least imaginative one: take a brief you have already executed by hand, build it in Concord, and compare the diff to what you actually did.