Review a product workspace
Read one product in depth: its current standing, the weeks behind it, and the decisions that shaped it. Run this when the portfolio flags a product worth a closer look.
Goal
A grounded read of one product, with its evidence history and decision history in view, so the next call about it is not made from this week alone.
Steps
Open the product from the portfolio
The workspace reads one product and its latest state. An agent can read the same data over MCP with get_product. Like the portfolio, this path is read-only.
Read the current standing
Stage, posture, and monthly budget are what the company has decided about this product so far. Each of these was set by a decision, so if one looks wrong, the decision log tells you when it was set and why.
Read this week across the three areas
Product, Sales, and Marketing each carry one signal for the week, with a value against a target, a trend, and a status. An area marked missing is a gap in your evidence, not a result.
Walk back through the weeks
One week is a point; several weeks are a direction. Read the snapshot history to separate a real move from a single noisy week. An agent reads the same history with list_product_snapshots.
Read the decisions behind it
The decision log for the product shows what was decided, when it took effect, and what each one changed. Where a decision was superseded, the replacement links back to it, so you can read the sequence rather than guess it.
Return to the portfolio before deciding
Depth on one product tells you what is happening; it does not tell you whether this product deserves the focus or the money over the others. Take the read back to the portfolio and make the call there.
Worth knowing
History is what a single week cannot give you
Because snapshots are immutable and one per week, the history is a real record rather than a rolling figure that was quietly restated.
A target that moved is not a result that moved
Each signal stores the target as it stood that week. If the numbers look better, check whether the value rose or the target was lowered by a metric decision.