What is the difference between a creative decision and an optimization decision?
In the creative judgment panel, Shannon Ford separated creativity from data and optimization. A team can use both, as long as it doesn't ask them to answer the same question. Decide the concept first, then let optimization improve versions built around it.
Write the concept decision first. Describe the audience, the problem and the idea the work should communicate, and explain why that direction deserves to exist before anyone produces many versions of it.
Then define what optimization may change, such as wording, placement or another execution detail. The core promise and the verified facts stay consistent unless the test is explicitly examining them. Afterward, check whether the winning version still serves the concept. A local measure may favor a treatment that doesn't support the broader assignment. Investigate that tradeoff before adopting the file with the largest number.
Keep a separate record of each decision. The creative choice explains what the team is trying to say. The optimization choice explains how it's improving one expression of that idea. With both visible, the work is easier to defend and easier to revise when the evidence changes.