This plan links each creative decision to its source, approval, version, and observation. It is designed to preserve what the team learned without turning a limited run into a broad promise.
Stage 1: Assemble the evidence pack
- Approved product, offer, audience, and destination facts
- Rights and consent records for every source, likeness, voice, and audio asset
- Current channel and disclosure checks owned by the responsible reviewers
- Known measurement limits and the decision the work is intended to inform
Stage 2: Create a hypothesis register
| Field | What to record |
|---|---|
| Question | The narrow uncertainty this creative version is intended to explore |
| Change | The specific difference from the reference version |
| Supported facts | The evidence and approved wording used in the creative |
| Decision rule | Who will decide what happens next and which evidence they require |
| Stop condition | The safety, budget, policy, or quality event that stops the run |
Stage 3: Keep a production record
- Assign a stable version identifier to the script, inputs, settings, and output.
- Link every asset to its permission record and every claim to its approved source.
- Record manual edits, review findings, rejected versions, and unresolved limitations.
- Retain the approved export and the exact handoff accepted by the launch owner.
Stage 4: Pre-register the observation
- State the audience, placement, destination, observation window, and measurement owner.
- Change one defined element where the campaign design makes that comparison valid.
- Choose measures that answer the stated question and document known gaps.
- Let the accountable analyst define adequacy for the decision instead of using an arbitrary universal threshold.
- Set safety and budget stop conditions before launch.
Stage 5: Write the learning record
| Record | Prompt |
|---|---|
| Observation | What happened in the documented window? |
| Limitations | What prevents a broader conclusion? |
| Interpretation | Which explanations remain plausible? |
| Decision | Keep, revise, stop, or collect more evidence - and who approved it? |
| Next question | What is the smallest useful uncertainty to investigate next? |
Avoid false causal conclusions
Audience, auction conditions, offer, destination, timing, measurement loss, and other changes can affect an observation. Keep those limits beside the decision and do not describe the creative as the cause unless the method supports that claim.
Start the plan with script hypotheses that use approved evidence and explicit review gates.
Open the script worksheet