The weekly review is a meeting with a fixed agenda and a non-negotiable rule: no claim without evidence, no evidence without a source, no decision without a reason. Run it the same way every week, because the power is in the repetition. The fifth review is where patterns start appearing that the first four couldn't show.
Open the Marketing Plan and work the checklist in order. The outcome: baseline, current, target, or the honest admission that success is not defined yet. The one priority move, with the evidence and capacity reasoning behind it. What changed since last week, read from "What changed and why" rather than from memory. What shipped, task by task. What was measured: each completed task either has a recorded result or it doesn't, and "doesn't" gets a plan to fix it. What was learned: one belief that changed because of evidence, stated plainly. What decision is needed: named, owned, and decided before you leave. The Decision Inbox does not roll forward silently.
Then set next week. For each active task: keep, simplify, skip, or change, with a reason for each. Skipping with a reason is a planning decision; the system learns from it. Small changes happen on the tasks (reorder, skip with a reason, add a task); rework happens through "Edit plan," which builds a new draft plan you approve. Either way it happens in the open, never as a silent change. If the business is a client, run the Client review agenda: the outcome, what was observed, what remains Unknown, what shipped, what the result shows, what it does not show, and the decision for next week. Clients can handle bad news delivered this way; what they can't handle is discovering later that Unknown was presented as a conclusion.
Close the review by writing down the one sentence the whole week hangs on: "We now believe [X], because [evidence], so next week we will [decision]." File it where future-you can find it: Decision Memory keeps the record, and Marketing Memory keeps the learning. Then start the week. The loop only compounds if you close it.