← Back to playbooks

NORLA ACADEMY / PLAYBOOK

Decide a creative test before spending

A scored list of ad assets, a test plan with budget split and a written approval request for the media buyer.

Paid media review loopAny plan (3 agents)About 30 minutes

What you need

  • Template: Paid media review loopBuilt-in template · included with every plan. Pick it under “Start from” when you create a workspace.See the template
  • An active Orchestrator planAny plan (3 agents). Templates run only inside Orchestrator; the trial works for a first run.Compare plans
  • Course: Data-Driven Campaigns with AITeaches the method behind this playbook: 3.8 hours · On demand. Optional, but the checks below come from it.View the course

The brief to paste

Replace every bracket. The agents use nothing you do not put here.

Score these [n] assets for the [campaign] test: [one line per asset: name, format, hook, offer shown]. Audiences available: [list]. Total test budget: [amount]. Metric that decides: [metric]. Decision due: [date].

Steps

  1. 1
    Describe each asset in one line

    The scorer works from your descriptions, not from the files. Say what the hook is, whether the offer is visible and what the brand cue is; it cannot see the creative.

  2. 2
    Run and read the scores

    Each criterion gets a score and the lowest one gets a one-line reason. If a reason is wrong because your description was thin, fix the description and run again rather than arguing with the score.

  3. 3
    Check the plan’s assumptions

    The Test planner lists what it assumed (audience sizes, timing). Anything you know to be false goes back as a returned step with a note.

  4. 4
    Approve the spend request

    The Buyer’s desk writes the exact question being asked and the spend it commits. Approve it as the buyer, or return it with the change.

Where a person decides

Before anything is proposed as spend: the plan waits for the buyer’s approval with the commitment stated as a number.

Before you trust the output

  • Budget split adds up to the total in the brief.
  • The deciding metric is the one you named, not a proxy.
  • No asset was dropped without a reason in the scored table.

What this does not do: The system does not launch, pause or change campaigns; it prepares the decision.